Compare commits

..
107 Commits
Author SHA1 Message Date
fawney19 4e96b9d870 Merge pull request #418 from Avilianb/feat/codex-responses-websocket
feat: support Codex Responses WebSocket transport
2026-05-15 12:49:45 +08:00
Avilianb d3c0b1aa7f feat: support Codex Responses WebSocket transport 2026-05-10 01:59:48 +08:00
phenetr0andphenetr0 66bfd3e592 fix: avoid double conversion for forced cli sync streams (#367)
Co-authored-by: phenetr0 <[email protected]>
2026-05-03 02:11:16 +08:00
DaoandYour Name 392c557831 优化 OAuth Token 导入解析与号池管理直刷空列表问题 (#330)
* docs: add aether-proxy C++ rewrite design

* chore: ignore local worktrees directory

* 修复 OAuth Token 导入识别与账号信息解析

* 修复号池管理直刷账号列表为空

---------

Co-authored-by: Your Name <[email protected]>
2026-04-30 18:07:00 +08:00
49 c34565b02b fix(provider-test): 修复测试供应商配置时 DetachedInstanceError (#270)
两处修复:

1. endpoint_checker.py: 移除 asyncio.to_thread 包装同步 DB 查询,
   避免跨线程导致 SQLAlchemy Session 失效

2. provider_query.py: 并发测试预加载 ProviderEndpoint 和 ProviderAPIKey 时
   添加 joinedload(provider),确保 expunge 后不会触发懒加载
2026-04-17 10:37:19 +08:00
fawney19 f57fe6e13e fix(migration): 用 UPDATE...FROM 子查询修复 total_tokens 自引用更新问题,增加批次上限防止死循环 2026-03-24 02:07:24 +08:00
fawney19 7c678b715f fix(migration): 修复 usage token 语义迁移脚本
- 将内联 SQL 提取为模块级常量 _UPGRADE_BACKFILL_SQL / _DOWNGRADE_BACKFILL_SQL
- 抽取 run_backfill_in_batches() 统一批量更新逻辑,通过 autocommit_block 释放 ALTER TABLE 锁
- 修正 upgrade 中 total_tokens 计算逻辑:用 COALESCE(input_tokens,0)+COALESCE(output_tokens,0) 补全 input_output_total_tokens
- WHERE 条件改为按实际字段差异筛选待回填行,替换原先仅过滤 input_context_tokens=0 AND total_tokens=0 的不完整条件
- 批量大小由 5000 调整为 500,减少单批锁持有时间
2026-03-24 01:56:36 +08:00
fawney19andNyaDoo bd4e5f3a5d feat(analytics): 重构统计分析模块,统一 API 与前端视图
close #260

- 新增 `src/services/analytics/query_service.py`,集中实现排行榜、性能、时间序列等查询逻辑
- 新增 `src/api/analytics/routes.py`,替代原 `stats/` 和 `dashboard/` 的分散路由
- 删除旧 `src/api/admin/stats/`、`src/api/dashboard/` 模块
- 重构 `src/api/user_me/routes.py` 与 `src/api/admin/usage/routes.py`,精简用量查询接口
- 新增 Alembic 迁移,修正 token 语义字段
- 前端新增 Analytics.vue、LeaderboardTab、PerformanceTab、ReportsTab 及 Reports 用户视图
- 新增 composables(useAnalyticsFilters、useReportsData、useLeaderboardData、usePerformanceData)
- 新增工具函数:analyticsGranularity、analyticsTimeseries、chartTheme、csvExport、usageBreakdown
- 删除旧 CostAnalysis、PerformanceAnalysis、UserStats 页面及相关组件
- 前端 API 层重组:新增 analytics.ts、request-details.ts,删除 dashboard.ts 和 usage.ts

Co-authored-by: NyaDoo <[email protected]>
2026-03-24 01:51:53 +08:00
York Zang 165d9eab8f fix(proxy): 修正代理连通性测试地址,避免 1.1.1.1 证书校验失败 (#257)
代理连通性测试此前使用 https://1.1.1.1/cdn-cgi/trace 作为探测地址,在标准 TLS 校验下会因为证书与 IP 不匹配而失败。
改为使用基于域名的 Cloudflare trace 地址,避免触发 CERTIFICATE_VERIFY_FAILED,并恢复代理测试结果的准确性。
2026-03-24 00:04:01 +08:00
fawney19andAAEE86 dfb95f09e1 feat(health): 增强健康监控面板,支持按 Key 查看详情与摘要统计
- 后端 health monitor 新增 summary 接口,按 api_format 聚合健康摘要
- 新增 GroupedFormatKey 类型与 key 分组查询接口
- HealthMonitorCard 重构为卡片+详情对话框,展示成功率、响应时间、Key 状态
- 抽取 HealthMonitorDetailDialog 独立组件
- 新增 useRouteQuery composable 用于 URL query 参数双向绑定
- PoolManagement/ProviderManagement 集成健康监控入口
- 新增 health monitor summary 单元测试

Closes #256

Co-authored-by: AAEE86 <[email protected]>
2026-03-24 00:00:08 +08:00
fawney19andhemo94931 4d5c591654 fix: 覆写规则条件可使用映射前请求体
新增 rules_original_body 参数贯穿请求构建链路,确保 body_rules/header_rules
条件评估使用模型映射前的原始请求体;附带将 handlers __init__ 改为延迟导入。

Closes #255

Co-authored-by: hemo94931 <[email protected]>
2026-03-23 17:31:29 +08:00
fawney19 46737d32f8 feat: 引入 status_snapshot 统一 provider key 状态管理
- 新增 StatusSnapshot 模型,聚合 OAuth / 账号 / 配额三维状态
- 新增 StatusSnapshotStore 负责快照的持久化与查询
- 重构 response_builder / endpoint_models,基于 snapshot 输出状态字段
- 前端抽取 providerKeyStatus / oauthRefreshFeedback 工具函数,
  统一 PoolManagement、ProviderDetailDrawer、BatchDialog 的状态展示
- errorParser 增加已知 OAuth 错误的友好提示
- refresher 适配 snapshot 写入,account_state 扩展状态分类
- 新增 alembic 迁移及存量数据回填脚本
- 补充前后端单元测试
2026-03-20 19:16:52 +08:00
fawney19 25d38ae632 feat(oauth): 账号封禁前置 OAuth 验证、抽取 provider_context、完善账号状态分类
- 新增 verify_oauth_before_account_block:在标记账号封禁前先尝试刷新 token,
  区分 OAuth 过期与真正的账号级封禁,避免误标
- 抽取 provider_context.py 统一解析 provider_type,解决 ORM detached 访问问题
- account_state 新增 workspace_deactivated 分类和 auto-removable 状态集合,
  补充中文验证关键词匹配
- OAuth refresh 成功后仅清除可恢复的 token 错误,不再自动清除账号级 block
- deploy.sh 依赖指纹改用纯 shell 实现,移除对 Python tomllib 的依赖
- 前端 Pool 管理页面新增筛选和批量操作优化
- 补充对应测试用例
2026-03-20 16:50:59 +08:00
fawney19 aa83b4a7a7 fix: 加固续租失败处理、verify_auth 异常捕获及调度器注册追踪
- task_coordinator: 续租连续失败 5 次后主动触发 lock_lost 回调,失败间加指数退避
- proxy_nodes: lock_lost 回调由 lambda 改为具名 async 函数,确保异步停止逻辑正确执行
- provider_ops: 将 prepare_verify_config 纳入外层 try,捕获 ValueError 并返回失败响应
- maintenance_scheduler: 用 _registered_job_ids 动态追踪已注册任务,stop 时按列表清理
- stats_aggregator: 内联 _do_aggregate 为 for/range(2) 循环,消除内嵌函数
2026-03-20 01:22:07 +08:00
fawney19 913ce2dbcb Merge pull request #252 from AAEE86/nn
fix(startup): 收口 leader 失锁后的后台任务
2026-03-20 01:13:01 +08:00
fawney19 cae5e520ac fix(frontend): 修复 restoreOriginalPlaceholder 递归调用、优化日志参数格式,补全 tsconfig lib 配置 2026-03-20 01:06:15 +08:00
fawney19 772f2ea601 fix(vertex): SA 认证注入代理配置,细化 token 获取异常处理
- _auth_service_account 接收 endpoint 参数,通过 _get_proxy_config 解析代理
- vertex_auth 区分 TimeoutException/RequestError/通用异常,提供可读错误信息
- 新增测试覆盖代理传递和超时场景
2026-03-20 00:49:44 +08:00
fawney19 28fa03451c feat(provider): 模型测试支持自定义请求头,优化对话框布局与并发策略
- 前后端新增 request_headers 字段,测试时可自定义额外请求头
- ModelTestDialog 拆分为请求头/请求体并排双栏布局,增加格式化与重置按钮
- 区分 Pool 托管(并发5)和单 Key Provider(并发1)的测试并发数
- JsonImportInput 新增 multiple prop 支持单文件模式
- KeyFormDialog Service Account 输入改用 JsonImportInput,支持拖拽导入
2026-03-20 00:24:56 +08:00
fawney19 6984984c22 feat(provider): 重构模型测试对话框,加固 Vertex AI 传输层
模型测试:
- 将消息输入替换为完整 JSON 请求体编辑器,支持格式化和校验
- 新增端点选择面板,测试前可选择目标端点
- 新增调试检查器,可查看每次尝试的请求/响应头和体
- 结果视图改用 HorizontalRequestTimeline 组件展示请求追踪
- endpoint_checker 返回完整调试数据,通过 candidate extra_data 持久化

Vertex AI:
- 改进上下文检测逻辑,不再仅依赖 provider_type,支持从 base_url 推断
- Service Account 密钥现支持自动拉取模型(使用 auth_config 而非 api_key)
- 移除 Gemini Developer API 回退,API Key 仅走 Express 模式
- 端点表单为 Vertex AI 显示格式特定的默认路径模板
- 密钥格式校验仅在 auth_type/api_formats 变更时执行

其他:
- 禁用 ClaudeCode 提供商类型创建入口
- Dialog 组件新增 closeOnBackdrop 属性
2026-03-19 23:52:17 +08:00
AAEE86 1209c835c7 fix(startup): 收口 leader 失锁后的后台任务
- 为后台调度器注册失锁回调并只在 stop 成功后清空生命周期引用
- 停止调度器时移除定时 job,补充启动与任务协调器回归测试
- 降低多 worker 下重复调度风险,保持停机收口与统计聚合回归一致
2026-03-19 22:26:51 +08:00
fawney19 e4ebd5cca1 refactor(oauth): LinuxDo 备用端点回退、Basic Auth 认证,修复 session 外访问 ORM 对象
- LinuxDo provider: token/userinfo 请求增加 backup 端点自动回退
- LinuxDo provider: token 请求改用 HTTP Basic Auth 认证
- 授权 URL 构建: scope 为空时不再发送该参数
- OAuthService: 引入 OAuthAuthenticatedUser 快照,避免 DB session 关闭后访问 ORM 对象
- OAuthService: _handle_login_sync 设置 expire_on_commit=False 防止属性过期
- 新增 LinuxDo provider 单元测试(Basic Auth、端点回退)
- 新增 _handle_login_sync 返回快照的集成测试
2026-03-19 20:32:33 +08:00
fawney19 f573110725 fix(frontend): 用 CSS text-security 替代 password 输入框,简化配额进度条 UI
- Input 组件 masked 模式改用 WebkitTextSecurity: disc 替代 type=password,
  避免浏览器密码管理器自动填充干扰
- LoginDialog/UserFormDialog/Settings/ProxyNodes 密码字段统一迁移到 masked 属性
- PoolManagement 配额进度条布局从 grid 改为 flex,移除未使用的
  getQuotaProgressDisplayClass/getQuotaProgressTooltip 函数

Close #249
Co-authored-by: AAEE86 <[email protected]>
2026-03-19 20:04:56 +08:00
fawney19andAAEE86 a8620e133a fix(kiro): 加固 Kiro adapter 错误处理与请求构建逻辑
- 提取 request.py 统一 URL/headers/payload 构建,消除 handler_adapter_base 与 envelope 的重复逻辑
- 新增 error_enhancer 模块,分类 HTTP 状态码与连接错误,增强上游错误诊断信息
- envelope 实现 extract_error_text / on_http_status / on_connection_error,透传错误上下文到 eventstream rewriter
- KiroRequestContext 扩展错误状态字段,支持网络诊断信息传递
- provider_oauth_utils 改用 importlib 动态加载,避免 core 层对 services 的静态依赖
- idc auth_method 下跳过 profileArn,修复 usage 查询参数

Closes #247

Co-authored-by: AAEE86 <[email protected]>
2026-03-19 13:54:40 +08:00
RWDai ddd6adbcf7 feat(provider): support custom prompts for model tests (#242) 2026-03-19 13:11:54 +08:00
fawney19 b570aaac48 fix: 补全 OpenAI 工具参数 schema 中缺失的 properties 字段
OpenAI API 要求 type=object 的 schema 节点必须声明 properties,
否则会拒绝请求。在 request_from_internal 出口处对工具参数 schema
进行深拷贝并递归补全缺失的空 properties,不影响内部表示。
2026-03-19 02:36:22 +08:00
fawney19 086efe6efe refactor(usage-queue): ACK 后立即删除消息,清理 consumer 元数据
- 消费成功/移入 DLQ 后立即 XDEL,避免主流 Redis 保留已入库历史
- 缩小 usage_queue_stream_maxlen 默认值:200000 -> 2000(仅作短暂缓冲)
- 启动时清理 pending=0 且长期闲置的旧 consumer(防 consumer group 元数据累积)
- 停机时主动 XGROUP DELCONSUMER 移除自身
- 移除 cache_fingerprint 模块及其对 telemetry/recording_helpers 的引用
- 同步更新相关测试,验证 xdel 调用及 consumer 生命周期行为
2026-03-19 02:09:09 +08:00
fawney19 56f3c95763 fix: 统一 input_context_expr 计算口径,移除按 api_format 分支的 CASE 逻辑
input_context_expr() 原先按 OpenAI/Gemini 和 Claude 分支计算输入上下文,
现统一为 input_tokens + cache_read_input_tokens,与 usage 表展示口径一致。
同步更新 admin 和 user_me 路由中的缓存命中率注释,并新增单元测试。
2026-03-19 01:35:00 +08:00
fawney19 8a6a961900 feat: 增强 OAuth 标识展示与 Codex 调试日志
- 优化 OAuth badge 支持 account_id/account_user_id,tooltip 展示完整身份信息
- 重构池管理额度进度条布局,倒计时独立展示并增加样式区分
- Codex 插件增加 token 解析快照日志,便于排查 OAuth 透传问题

Closes #245
Co-Authored-By: kayphoon <[email protected]>
2026-03-19 00:50:18 +08:00
kayphoon 6e55968487 feat: Codex account_name 透传并优化池列表 OAuth 标识与额度重置倒计时展示
(cherry picked from commit 1b9176fc4a)
2026-03-18 23:58:42 +08:00
fawney19 8d8cddcef6 fix(replay): rerun model mapping on replay target 2026-03-18 23:54:19 +08:00
RWDai b90d5095f1 Fix replay fallback when model mapping is missing 2026-03-18 23:54:19 +08:00
RWDai a4505b1281 Fix usage replay model remapping 2026-03-18 23:54:19 +08:00
fawney19 1d72a8f9c1 feat: 流式空闲超时、健康监控查询优化、限流桶内存上限与维护清理修复
Close #233

Co-authored-by: AAEE86 <[email protected]>

- cli_monitor_mixin: 引入 STREAM_IDLE_TIMEOUT_SECONDS(可通过环境变量配置),
  流传输开始后若超出空闲窗口无新 chunk 则提前取消并返回 504,避免长时间挂起
- stream_context: 新增 managed_recorded_bodies 上下文管理器,确保 chunks 在
  telemetry 完成后及时释放;stream_telemetry 使用该接口统一管理 response body 构建
- health endpoint: 将状态聚合改为 GROUP BY 直接统计,事件列表按 api_format
  单独查询,避免单次 limit 拉取大量记录导致的遗漏与性能问题;同时过滤不活跃
  provider/endpoint,与公开健康接口保持一致
- endpoint health service: 修正时间线数据按 endpoint_id 而非 key_id 聚合
- token_bucket: 引入 max_buckets/bucket_expiry 上限与定时清理,防止内存无限增长;
  修复 refill_rate=0 时 get_reset_time 除零异常;新增 _is_unlimited_rate_limit 判断
- maintenance_scheduler: 调整清理顺序(先删整行再按窗口清理),新增 newer_than
  边界参数,避免同一行在同一轮中被重复改写
- sync_execute: 新增 create_pending_usage 开关,允许已预创建记录的调用方跳过重复创建
- quota_reader / provider_ops balance: 小幅修复与健壮性提升
- Dockerfile: 添加 MALLOC_ARENA_MAX=2 环境变量以降低 gunicorn worker RSS
- 补充相关测试覆盖
2026-03-18 23:38:26 +08:00
fawney19 3d5b6141a5 feat(codex): 引入 upstream_headers hook 机制,为 Codex 注入 session/conversation/account headers
- 新增 upstream_headers.py:可注册 provider+endpoint 维度的 extra headers 构建 hook
- Codex openai:cli 注入 session_id + conversation_id(由 prompt_cache_key sha256 派生)
- Codex openai:compact 注入 chatgpt-account-id(来自 auth_config)+ session_id,不注入 conversation_id
- 修复 prompt_cache:compact 格式现统一为 codex 策略,不再跳过注入
- chat_handler_base / cli_request_mixin 均在 extra_headers 阶段调用 build_upstream_extra_headers
2026-03-18 20:43:14 +08:00
fawney19 203cd5a9d5 fix(redis): 按事件循环隔离 Redis 连接,防止子线程 asyncio.run 导致连接泄漏
RedisClientManager 新增 _redis_by_loop 字典,按 event loop id 维护独立连接,
避免 usage consumer 在 asyncio.to_thread + asyncio.run 场景下复用主循环连接。

同步重构 consumer_streams 的写库路径:_apply_record_event 统一走
record_usage_batch,_apply_streaming_event 改为 to_thread 执行同步 DB 操作,
移除冗余的 session 传递和手动 rollback 逻辑。
2026-03-18 17:56:48 +08:00
fawney19 696ec65175 refactor(normalizer): 移除 request key reorder 机制,保持自然插入顺序
移除 OpenAI/OpenAI CLI normalizer 中的 _reorder_request_prefix_keys 及
基类 _reorder_request_keys 死代码,request_from_internal 直接返回构建
顺序的 dict,测试同步更新为验证自然插入顺序。
2026-03-18 14:01:56 +08:00
fawney19 cbb66a5667 refactor(task): 引入 MutableRequestBodyState 替代 request_body_ref 字典容器
将请求体可变状态从 {"body": dict} 字典容器重构为独立的
MutableRequestBodyState 类,统一管理 original_body / current_body /
build_attempt_body / rectify 等语义,消除各层通过 ref["body"] 间接
读写的隐式约定。

- 新增 src/services/task/request_state.py 定义 Protocol 与实现
- handler/executor/mixin 层改用 request_state 参数传递
- error_handler/state_transition 通过 request_state 判断整流状态
- 新增 request_state 单元测试与 chat/cli 请求体隔离测试
2026-03-18 13:43:39 +08:00
fawney19 53ef35ec80 refactor(cli): 提取 _build_upstream_request 统一流式/非流式的上游请求构建逻辑
将 cli_stream_mixin 和 cli_sync_mixin 中重复的上游请求构建代码
(provider behavior / stream policy / envelope / auth / RequestBuilder / URL 构建)
提取到 cli_request_mixin._build_upstream_request,返回 CliUpstreamRequestResult dataclass。
2026-03-18 13:13:47 +08:00
fawney19 1af3067303 refactor(codex): 移除 envelope/request_patching 层,用 context var 统一 compact 状态判断
- 删除 CodexOAuthEnvelope 和 request_patching 模块,Codex 不再需要 envelope 层
- 移除 _aether_compact 请求体内部标记,改用 is_codex_compact_request() 集中查询
- 简化 OpenAI CLI adapter,移除 Codex 专用的 get_cli_extra_headers/build_test_request_body 逻辑
- 移除 Codex behavior variant 注册(same_format/cross_format)
- normalizer patch_same_format_request 对 codex 变为 no-op
- 更新相关测试适配新的架构
2026-03-18 12:35:57 +08:00
fawney19 d026398bab refactor(body-rules): 移除 protected_body_keys 机制,允许 body_rules 自由修改所有请求体字段
删除 get_cache_sensitive_protected_body_keys 函数及相关常量、_is_protected_path
辅助函数,从 apply_body_rules、RequestBuilder、ProviderRequestResult 等处移除
protected_body_keys 参数,更新所有调用点和测试用例。
2026-03-18 10:52:07 +08:00
fawney19 684689a82b fix(usage): session touch 独立提交避免行锁阻塞 & 管理员页面顺序加载降低并发压力
后端: 将 session touch 的 commit 从请求事务中分离,防止管理员 usage
页面的长查询持有 user_sessions 行锁阻塞后续请求。touch_session 改为
返回 bool 以支持按需提交。

前端: 管理员 Usage 页面将并行 API 调用改为顺序加载,优先显示记录表格,
统计面板在后台异步刷新,避免瞬时并发打满后端 worker。loadRecords 支持
传入 dateRange 参数确保时间范围一致性。
2026-03-18 00:13:28 +08:00
github-actions[bot] eeb5f41bad chore(proxy): update download links for proxy-v0.2.5 2026-03-17 14:23:27 +00:00
fawney19 7180eaea88 chore: bump aether-proxy version to 0.2.5 2026-03-17 22:16:47 +08:00
fawney19 7cb204f18a chore: bump aether-hub version to 0.2.0 2026-03-17 22:15:07 +08:00
fawney19 0342f609d0 feat(tunnel): 请求体流式传输 & OpenAI CLI 请求 key 排序优化
Hub 端:
- open_local_stream 不再接收 body 参数,改为通过 push_local_request_body 分块推送
- 请求体按 32KB 分帧发送,避免大请求一次性压缩和传输
- local_relay 改为流式解析 envelope 和转发请求体

Proxy 端:
- stream_handler 改为流式传输请求体到上游,不再预先收集完整 body
- upstream_client 请求体类型从 Full<Bytes> 改为 UnsyncBoxBody 以支持流式传输
- dispatcher 将 StreamEnd/StreamError 事件转发给 stream handler

Python 端:
- hub_transport relay envelope 改为异步生成器流式发送
- 提取 reorder_openai_cli_request_prefix_keys 为公共函数
- Codex passthrough 路径也应用稳定的前缀 key 排序
2026-03-17 22:07:09 +08:00
fawney19 59840fa419 feat(docker): 支持通过 GITHUB_MIRROR 参数加速 hub 二进制下载
Dockerfile.app.local 新增 GITHUB_MIRROR 构建参数,deploy.sh
新增 --mirror 选项,国内服务器可指定镜像代理地址加速下载。
2026-03-17 20:43:14 +08:00
fawney19 d390d46ee8 feat(docker): Dockerfile.app.local 支持本地 hub 二进制文件
将 aether-hub tar.gz 放到项目根目录即可跳过 GitHub 下载,
解决国内服务器构建时 GitHub 访问慢的问题。
2026-03-17 20:37:09 +08:00
fawney19 37eada9682 chore: bump aether-hub version to 0.1.9 2026-03-17 20:26:29 +08:00
fawney19 73a5325a38 refactor(hub): 用本地 HTTP relay 替代 Worker WebSocket 长连接
Hub 数据面改为 /local/relay/{node_id} HTTP 端点,Worker 通过本机
HTTP 请求转发 tunnel 帧,不再维护 /worker WebSocket 长连接。

Hub 侧:
- 新增 control_plane.rs: Hub 通过 HTTP 回调 Aether app 处理心跳 ACK 和节点状态变更
- 新增 local_relay.rs: 接收本地 HTTP 请求,在 Hub 内部打开 LocalStream 并透传到 proxy
- 移除 worker_conn.rs 及 Worker WebSocket 处理逻辑
- 简化 protocol.rs: 移除 NODE_STATUS 帧类型,抽取通用 encode_frame/decode_payload

Python 侧:
- 删除 tunnel_manager.py 及其 WebSocket 连接管理器 (HubConnectionManager)
- 简化 hub_transport.py 为 HTTP relay 调用
- 新增 src/api/internal/hub.py 接收 Hub 控制面回调 (heartbeat/node-status)
- hub_config.py 移除 WebSocket 相关配置,改为 HTTP relay URL
- service.py 新增 update_tunnel_status 方法
- 删除 src/api/admin/proxy_tunnel.py (旧管理接口)
- proxy_node 缓存 TTL 从 15s 降至 3s 加速状态感知
2026-03-17 20:22:12 +08:00
fawney19 460eb5434d fix(proxy-resolver): 将 resolve_ops_proxy_config 改为异步调用避免阻塞事件循环
在 anyrouter/nekocode/sub2api/yescode 架构及 service.py 中,
将同步的 resolve_ops_proxy_config 替换为 resolve_ops_proxy_config_async,
通过 asyncio.to_thread 包装避免同步 DB 查询阻塞事件循环。
2026-03-17 18:06:15 +08:00
fawney19 8a21cb9a55 chore: bump aether-hub version to 0.1.8 2026-03-17 17:32:01 +08:00
fawney19 d0df52ce35 fix(hub): worker 断连时向 proxy 端发送 STREAM_ERROR 防止流阻塞
将 unregister_worker 中的 stream 清理逻辑提取为 cancel_streams_for_worker
方法,与 cancel_streams_for_proxy 形成对称设计。worker 断连时主动通知
proxy 端终止相关 stream,防止 proxy 的 stream handler 永久阻塞等待请求体。
添加单元测试覆盖该场景。
2026-03-17 17:29:12 +08:00
fawney19 c1ed42fd3a feat(body-rules): 支持按 provider_type 覆盖 cache-sensitive 保护字段集合
Codex 通过 OpenAI CLI/Compact 格式转发时,endpoint body_rules 需要能
修改 instructions/input/tools 等 prompt 字段。新增 provider_type 维度的
保护集合映射,Codex 场景仅保护 prompt_cache_key,其余 prompt 字段交由
body_rules 自由调整。
2026-03-17 17:09:39 +08:00
fawney19andLewisPen c4bb6b8161 feat(auth): 重构认证系统,引入 session 会话管理
- 新增 user_sessions 数据库表及 Alembic 迁移
- 实现 SessionService 会话生命周期管理(创建/刷新/撤销/清理)
- 认证流程改用 refresh token cookie + access token 双令牌模式
- 前端实现自动静默刷新、跨标签页同步及设备指纹
- 用户设置页新增会话管理和密码修改功能
- 管理员用户管理新增强制登出和会话查看
- 密码策略增强,支持强度校验和泄露检测
- OAuth 登录流程适配新会话机制
- 新增完整的单元测试和 API 测试覆盖

Closes #232

Co-authored-by: LewisPen <[email protected]>
2026-03-17 16:34:09 +08:00
fawney19 d480aa11f3 fix(request-body): 使用 deepcopy 防止请求体在处理流程中被意外修改
handler 基类和格式转换 registry 中,原始请求体通过浅拷贝或直接引用传递,
导致下游处理(模型映射、格式转换、重试整流)可能修改原始数据,
影响后续重试或并发请求的正确性。统一改用 copy.deepcopy 隔离副本。
2026-03-17 13:23:02 +08:00
fawney19 d63d5eff85 fix(nginx): 统一所有 location 块的 CF 头剥离,补充 proxy_hide_header 防止响应泄露 2026-03-17 11:33:43 +08:00
fawney19 5dae2a4792 fix(normalizer): 移除 Claude system 中的 billing header 2026-03-17 11:10:04 +08:00
fawney19 dcbd7dc219 feat(normalizer): 稳定请求体字段顺序以提升 prompt cache 命中率
在 FormatNormalizer 基类新增 _reorder_request_keys 方法,OpenAI/OpenAI CLI
normalizer 各自定义前缀字段顺序(model, tools, messages / model, instructions,
tools, input),确保静态字段前置、动态内容后置。同时调整 OpenAI CLI 中
instructions 字段的构建位置使其在 input 之前。
2026-03-17 10:26:57 +08:00
fawney19 2dcf8b8414 refactor(prompt-cache): 移除 client_family 命名空间拆分,统一缓存 key 提升复用率
不再按 User-Agent 客户端类型拆分 prompt cache namespace,
所有客户端共享同一缓存 key,版本升级至 v3。
2026-03-17 09:49:35 +08:00
fawney19 438f16094f refactor(stream): 将不完整流 token 估算逻辑收敛到 StreamContext
- 新增 has_partial_response / ensure_estimated_output_tokens / should_estimate_incomplete_tokens 方法
- CancelledError 路径在归因前即补充 output_tokens,确保日志包含估算值
- CLI Handler 和 Chat Handler 的兜底估算统一使用 should_estimate_incomplete_tokens
- 移除 cli_monitor_mixin 和 stream_telemetry 中重复的条件判断
- 新增对应单元测试
2026-03-17 03:37:06 +08:00
fawney19 40e0b82fa0 fix(chat-handler): 修复 _prepare_provider_request 缺少 original_headers 参数导致的 NameError 2026-03-17 03:21:19 +08:00
fawney19 b9b0a75fe4 feat(cache-fingerprint): 新增字段级指纹,支持逐字段 sha256 和字节数追踪
在请求缓存指纹中为每个 cache_relevant 字段单独计算 sha256 和字节数,
便于精确定位哪些字段发生了变化。fingerprint version 升级为 2。
2026-03-17 02:57:36 +08:00
fawney19 b6cc0bc3a7 refactor(usage): 统一 cache token 提取逻辑,新增请求缓存指纹记录
- 新增 extract_cache_read_tokens() 兼容 OpenAI/Claude/Gemini 多种字段命名
- parsers/stream_processor/cli_event_mixin 统一使用提取函数替换内联逻辑
- 同时兼容 prompt_tokens/completion_tokens (OpenAI) 和 input_tokens/output_tokens (Claude)
- 新增 cache_fingerprint 模块,在 telemetry 记录时自动计算并附带请求缓存指纹
- 新增对应单元测试
2026-03-17 02:34:32 +08:00
fawney19 d2f1431269 refactor(prompt-cache): 将 prompt_cache_key 生成从 Codex 专用模块提取为通用服务,支持 OpenAI 官方 API 和 Codex 端点
- 新增 prompt_cache.py 统一管理 prompt cache key 的生成逻辑
- 基于 User-Agent 识别客户端家族(openai_python/openai_node/codex_desktop 等),不同客户端生成不同 cache key
- 在 chat_handler_base/cli_stream_mixin/cli_sync_mixin 统一调用 maybe_patch_request_with_prompt_cache_key
- Codex request_patching 不再负责 prompt cache key 注入,仅保留内部标记清理
- 新增 is_official_openai_api_url 工具函数区分 OpenAI 官方 API 与兼容端点
2026-03-17 01:55:24 +08:00
fawney19 4ecaefbade refactor(conversion): body_rules 保护 cache-sensitive 字段,normalizer 保真优化与诊断日志
- RequestBuilder 新增 protected_body_keys 机制,按 provider API 格式阻止 body_rules
  改写 prompt cache 相关的顶层请求字段(messages/tools/system 等)
- Claude/Gemini normalizer 优先复用原始 raw tool_choice,避免 round-trip 丢失信息
- Claude normalizer 修复 content blocks 输出顺序(flush_text_parts),
  _coerce_claude_message_sequence 返回结构化诊断
- OpenAI normalizer 保留 raw tool call arguments 字符串与原始 tool 定义 extra 字段
- schema_utils allOf 合并保持 required 字段插入顺序
- 各转换环节增加结构化 debug 日志用于调试
2026-03-17 01:25:29 +08:00
fawney19 c97c9332eb feat(codex): 基于用户 API key 生成稳定的 prompt_cache_key,实现跨 provider key 的 prompt 缓存复用
- prepare_context 传递 user_api_key_id 到 CodexRequestContext
- wrap_request 中调用 patch_openai_cli_request_for_codex 注入 prompt_cache_key
- 客户端已提供 prompt_cache_key 时不覆盖
- 补充对应单元测试
2026-03-16 17:25:52 +08:00
fawney19 c070e5a9f6 fix(conversion): OpenAI CLI 流式转换 tool 调用与文本输出 block index 不再复用
引入统一的 block index 分配器(_allocate_block_index / _ensure_text_block_index),
避免工具调用后紧跟文本输出时 Claude block index 冲突。
2026-03-16 16:19:27 +08:00
fawney19 8cd3a69803 fix(oauth): ACCOUNT_BLOCK 标记不再禁用 key,token 刷新成功自动解除所有 invalid 标记
- error_handler 标记 ACCOUNT_BLOCK 时保持 is_active=True,oauth_invalid 标记已
  足够阻止调度,配额刷新仍可覆盖该 key
- 移除 is_account_level_block 守卫:token 刷新成功即证明账号可用,清除所有
  oauth_invalid 标记(含 ACCOUNT_BLOCK)
- health_policy 401 瞬时失败不再设置 60s cooldown,仅清除 token 缓存后立即重试
- key_quota_service 全量刷新时纳入 ACCOUNT_BLOCK 的 key,账号恢复后可自动解除
- 各 refresher 刷新成功时显式恢复 is_active=True
2026-03-16 15:50:41 +08:00
fawney19 791c9c98dc refactor(conversion): OpenAI Chat/Responses API 跨格式字段双向转换统一化
- 将工具/tool_choice/web_search/custom tool 的双向转换函数提取到 constants.py,
  openai.py 和 openai_cli.py 共享,消除两端逻辑不一致
- 新增 Chat <-> Responses 的 passthrough 字段白名单,支持 metadata/user/
  service_tier/prompt_cache_key 等字段跨格式透传
- 支持 text config (response_format + verbosity) 在 Chat/Responses 间互转
- 修复 Gemini json_schema 解包:OpenAI 的 {name, schema, strict} 包装层
  不再被整体传入 Gemini responseSchema
- 修复 reasoning_effort 优先级:显式 effort 优先于 budget_tokens 反推,
  避免 Claude output_config.effort 被覆盖
- Gemini 格式输出增加 web_search_options -> googleSearch 工具映射
- Claude schema validator 增加 web_search 类型工具的宽松校验
- 新增覆盖测试:custom tool/tool_choice、allowed_tools、web_search 双向转换、
  text config 映射、passthrough 字段保留、跨格式 schema 校验
2026-03-16 14:21:14 +08:00
fawney19 025e979935 fix(codex): 配额刷新 401/403 不再自动禁用 key,区分软性请求失败与账户封禁
- 新增 OAUTH_REQUEST_FAILED_PREFIX 标记非 token 失效的 403 请求失败
- 引入 _merge_invalid_reason 合并逻辑,避免低优先级原因覆盖高优先级状态
- account_state 中 REFRESH_FAILED/REQUEST_FAILED 前缀不再触发封禁判定
- 401/403 返回 auto_disabled=False,不再直接设置 is_active=False
2026-03-16 01:12:35 +08:00
fawney19 131471a13f refactor(maintenance): body 压缩改为逐条独立事务,降低内存占用与锁粒度
将 _cleanup_body_fields 从批量加载完整记录改为先查询 ID 列表,
再逐条独立会话处理压缩,避免大批量事务导致的内存和锁问题。
批次大小上限从 100 降至 25,新增排序保证处理顺序确定性。
2026-03-15 23:46:01 +08:00
fawney19 7ff63077c3 feat(providers): provider 摘要列表排序增加启用状态优先
后端查询和前端展示均按 is_active 降序、priority 升序、created_at 升序排列,
确保已启用的 provider 始终排在前面。
2026-03-15 23:13:11 +08:00
fawney19 faba0cbd07 fix(openai-cli): item_id 与 call_id 不一致时 tool delta 映射错误
Responses API 中 function_call 的 item.id 和 call_id 可能不同,
增加别名注册和解析机制,确保后续 delta/done 事件统一使用 call_id。
2026-03-15 22:50:50 +08:00
fawney19 75f17935f9 fix(openai-cli): 补全 function_call 的 arguments 快照同步,修复无 delta 时参数丢失
在 output_item.done 和 function_call_arguments.done 事件中,通过 args snapshot
对比已发送的增量,补发缺失的 tool arguments delta,确保客户端收到完整参数。
2026-03-15 22:20:54 +08:00
fawney19 60842fbbb5 fix(openai): tool_call delta 重复携带 function.name,增强严格客户端兼容性
在 ToolCallStart 时记录 block_index 到 tool_name 的映射,
ToolCallDelta 时同步输出 function.name,避免严格客户端丢弃无 name 的 delta。
提取 _ss_dict 辅助方法统一 stream state 中 dict 字段的初始化逻辑。
2026-03-15 22:04:21 +08:00
fawney19 d58c27d22d fix(openai): tool_call delta 重复携带 id/type,修复严格客户端兼容性
部分 OpenAI 兼容客户端要求每个 tool_call delta chunk 都包含 id 和 type
字段,否则会将后续 delta 视为无效。通过 block_to_tool_id 映射在
ContentBlockStart 时记录 tool_id,确保后续 ToolCallDelta 能正确回填。
2026-03-15 21:51:44 +08:00
fawney19 65550159bb refactor(frontend): 优化条件编辑器 UI,限制嵌套深度为两层
- 顶层组合条件增加 ListFilter 图标按钮,可转回单条件
- 子组隐藏"+ 子组"按钮,防止无意义的深层嵌套
- 嵌套叶子节点隐藏转组合按钮,保持最多两层结构
2026-03-15 21:02:05 +08:00
fawney19 3c7ad81d62 feat(rules): 条件系统增强,支持 all/any 组合条件和 original/current 数据源切换
- evaluate_condition 支持递归 all/any 组合节点和 source 字段
- header_rules 支持 condition 条件触发,HeaderBuilder.apply_rules 透传 body/original_body
- 提取 EndpointConditionEditor 组件统一请求头/请求体规则的条件编辑 UI
- header_rules 新增服务端结构校验(action/key/from/to/condition)
- 新增组合条件、source 切换、fail-closed 等测试用例
2026-03-15 20:27:53 +08:00
fawney19 6b23c9b3ce fix(frontend): 使用记录表格"命中率"列名改为"缓存命中率" 2026-03-15 18:32:42 +08:00
fawney19 900e54d740 feat(conversion): 支持 Claude output_config.effort 跨格式转换,新增 xhigh 档位
- 新增 Claude output_config.effort 与标准化 reasoning_effort 的双向映射
- 新增 xhigh 档位(budget_tokens=8192),对应 Claude effort=max
- reasoning_effort 独立于 thinking 存入 extra,支持无 thinking 场景的跨格式传递
- OpenAI/Responses API 输出时 xhigh 自动降级为 high
2026-03-15 17:57:12 +08:00
fawney19 8cc70934da fix(openai,oauth): 修复 resp_ ID 前缀转换和 token 失效误判为账号级 block
- OpenAI normalizer: resp_ 前缀 ID 规范化为 chatcmpl- 前缀,流式 tool_call index 增加 block_index 回落
- OAuth token: 提取 token 失效关键词为共享常量,token invalidated 不再被判定为账号级 block
- Codex refresher: 403 + token invalidated 标记为 OAUTH_EXPIRED,允许 refresh_token 恢复
2026-03-15 16:56:58 +08:00
fawney19andLewisPen f92b0943b5 feat(rate-limit): 实现分层 RPM 限速,支持系统默认/用户/独立Key三级配置
- 新增用户级 rate_limit 字段,支持系统默认/用户自定义/不限制三种模式
- 独立 Key 的 rate_limit 语义调整:null=跟随系统默认,0=不限制,>0=自定义
- 实现 UserRpmLimiter 基于 Redis sliding window 的 RPM 限速引擎
- Pipeline 请求流程集成用户级 RPM 检查
- 管理后台和用户面板新增 RPM 限速配置与实时状态查看
- 系统设置新增全局默认 RPM 配置项
- 迁移脚本回填现有 API Key 的 rate_limit 默认值
- 新增用户/Key RPM 状态监控 API 和前端展示

Closes #231

Co-authored-by: LewisPen <[email protected]>
2026-03-15 14:22:59 +08:00
fawney19 920a383136 refactor(claude-code): 移除 TLS 指纹模拟配置,简化上下文构建逻辑
移除 enable_tls_fingerprint 配置项、TLS_PROFILE_CLAUDE_CODE 常量及
resolve_claude_code_tls_profile 函数,TLS profile 改为仅由 fingerprint
的 impersonate 字段决定,简化 build_and_set_claude_code_request_context
返回值为单一 context 对象。
2026-03-15 00:23:19 +08:00
fawney19 693e37d2df fix(frontend): 统计表格缓存列仅显示缓存读取token,与命中率计算口径一致
缓存列移除cache_creation_tokens,只保留cache_read_tokens,
避免缓存值大于输入值的反直觉展示。
2026-03-14 16:39:48 +08:00
fawney19 9c6036a103 fix(frontend): 统计表格token列防止换行,缓存token合并为总值显示 2026-03-14 16:26:07 +08:00
fawney19 751a4d9111 feat(usage): 统计表格新增输入/输出与缓存创建token细分展示
- 后端接口(admin/user_me/query)新增 output_tokens、cache_creation_tokens、total_input_context 字段返回
- 前端统计表格(模型/提供商/API格式)合并 Tokens 列为"输入/输出"+"缓存读取/创建"两行紧凑布局
- 合并"缓存Token"和"缓存命中率"为单一"命中率"列,减少表格宽度
- 修正 UsageRecordsTable 缩进及模板格式
2026-03-14 16:11:56 +08:00
fawney19 aafd332198 revert(frontend): 使用记录表格恢复格式与类型为独立列 2026-03-14 15:01:38 +08:00
fawney19 337cd0c505 fix(frontend): 修复模型缓存价格设为0时不生效的问题
缓存价格更新函数使用 numValue > 0 判断是否手动设置,
导致输入 0 时被错误地当作清空处理。改为检查原始输入
是否为空字符串/null/undefined 来区分清空与有效输入。
2026-03-14 14:32:58 +08:00
fawney19 b15ce9977a feat(frontend): 用量页面 UI 优化 - 合并格式/类型列、分析面板可折叠
- UsageRecordsTable: 合并 API 格式与类型为单列,缩减列宽使表格更紧凑
- MainLayout: 添加 header 级 Teleport 插入点供子页面注入操作按钮
- Usage: 用量分析面板改为可折叠,折叠状态持久化到 localStorage
2026-03-14 13:28:56 +08:00
fawney19andAAEE86 e0286aebe3 refactor: 共享请求管道、按需懒加载、流式内存护栏与连接池治理
- 抽取 ApiRequestPipeline 单例,44 个路由文件共享同一实例
- Handler/Adapter 模块级 __getattr__ 延迟导入,减少启动时间
- 新增 ensure_stream_buffer_limit() 流式内存护栏(16MB 单行 / 32MB 总量)
- HTTP 空闲连接清理与 curl_cffi LRU 会话池
- ensure_providers_bootstrapped 按需引导指定 provider_types
- Usage 事件序列化迁移至 msgpack,Redis codec 隔离
- 启动预热任务(/readyz 就绪门控)与优雅关闭
- 通知邮件模块独立开关与 SMTP 配置校验
- CryptoService DCL 线程安全修复
- 通知模块开关 DB 查询 30s 内存缓存
- /readyz 对 unknown 状态返回 503
- 预热关闭 5s 超时保护
- 预热适配器逐个 try-except 容错
- FormatConversionRegistry 哨兵模式防并发重复物化
- 流式缓冲检查无条件执行

Closes #230

Co-authored-by: AAEE86 <[email protected]>
2026-03-14 11:59:07 +08:00
fawney19 45985f1c04 Merge pull request #228 from NyaDoo/fix/config-timezone-and-key-dialog
fix: APP_TIMEZONE 循环导入 & 独立密钥额度编辑空白
2026-03-14 01:42:28 +08:00
fawney19andEntropy-Xu 776dd2f8ea feat(cleanup): 解耦 request_candidates 与 provider_api_keys 生命周期
- 移除 request_candidates.key_id 对 provider_api_keys 的外键约束(含迁移脚本)
- 删除 Key 时不再级联删除候选记录,改为独立按保留天数定时清理
- 新增 request_candidates_retention_days / request_candidates_cleanup_batch_size 配置项
- batch_delete_task 增加 lock_timeout 及超时自动降批重试机制
- cleanup_key_references 提取阶段化清理流程,移除 RequestCandidate 联动删除
- 前端 CleanupPolicySection 新增候选记录保留天数和清理批次配置

Closes #227

Co-authored-by: Entropy-Xu <[email protected]>
2026-03-14 01:33:08 +08:00
fawney19andAAEE86 bdfe4adc98 feat(usage): 修复缓存命中率计算并新增用户端 API 格式统计
- 新增 input_context_expr() 按 api_format 区分 input_tokens 语义
  (OpenAI/Gemini input_tokens 已含 cache_read,Claude 需额外加上)
- 缓存命中率统一改为基于归一化后的 total_input_context 计算
- 用户 /me/usage 接口新增 summary_by_api_format 后端聚合字段
- 前端 API 格式统计改用后端聚合数据,移除前端逐条记录手动统计
- 提取 formatHitRate 到 utils/format.ts 消除三处重复定义
- 移除 PoolManager 中未使用的 select_key 方法

Co-Authored-By: AAEE86 <[email protected]>
2026-03-14 00:33:19 +08:00
LewisPen 00a0371997 fix(frontend): 独立密钥编辑模式额度区域空白问题
v-else-if 条件遗漏导致编辑模式下非 unlimited 的密钥
额度区域既不显示 Input 也不显示提示文本,对齐 UserFormDialog
的 v-else 逻辑。
2026-03-13 10:58:34 +08:00
LewisPen ded6b5b081 fix(config): 统一 APP_TIMEZONE 至 Config 类,修复 wallet 循环导入
将散落在 scheduler / stats_aggregator / daily_usage_ledger / routes
中的 os.getenv("APP_TIMEZONE") 收归 Config.app_timezone,消除
wallet → system → maintenance_scheduler → user.preference → wallet
的循环导入链。
2026-03-13 10:54:21 +08:00
fawney19 e9678ea899 fix(admin): 优先级脏检查、base_url 校验、usage detail 延迟加载及导入数据验证
- 前端优先级管理: 保存时对比原始快照,仅提交实际变更的 provider/key 优先级,
  并限制并发请求数(SAVE_CONCURRENCY=6),避免无效 API 调用
- handler_adapter_base: _normalize_test_base_url 改为 _validate_test_base_url,
  移除对 dict 类型 base_url 的兼容,严格要求字符串输入
- provider_query: 新增 _require_test_endpoint_base_url,在测试链路提前校验
  endpoint.base_url 类型和非空
- system.py: 导入 endpoint 时通过 ProviderEndpointCreate 模型校验数据,
  拒绝非法 base_url 类型(如 dict)
- usage detail: 使用 defer() 延迟加载 body 列,通过 SQL CASE 表达式在
  数据库端计算 has_*_body 标记,减少不必要的大字段传输
- provider routes: 新建 provider 时 priority=0 边界处理,clamp 并 shift
2026-03-12 17:11:14 +08:00
fawney19 8d69f72e2a fix(stability): 健康监控 DB 操作异步化,防止 worker 超时崩溃
- executor: record_success 通过 asyncio.to_thread offload 到线程池
- error_handler: 3 处 record_failure 同样 offload 到线程池
- sync_execute: 设置 expire_on_commit=False 防止 commit 后 ORM 懒加载
- handler_adapter_base: 归一化 check_endpoint 的 base_url 输入
2026-03-12 16:17:22 +08:00
fawney19 127b4e11de ci(hub): 移除 build-hub workflow 中的 Docker 构建和推送步骤 2026-03-12 15:34:09 +08:00
fawney19 22093bed4d chore: bump aether-hub version to 0.1.7 2026-03-12 15:21:47 +08:00
fawney19 ebd53ad679 fix(hub): 修复 worker 连接清理逻辑,按完成顺序有序取消对端任务 2026-03-12 15:17:00 +08:00
fawney19 280c604327 移除deplpy.sh中每次自动拉取最新代码, 以便于回退版本 2026-03-12 14:57:15 +08:00
fawney19 ad31cdbf85 perf(stability): batch committer 异步化及降级冷却参数调优
- 将 batch_committer 的 DB commit 操作通过 asyncio.to_thread 移至线程池,避免阻塞事件循环
- 将 f-string 日志替换为 loguru 惰性格式化
- 降低事件循环延迟降级的冷却时间和乘数,加快降级恢复
2026-03-12 14:22:55 +08:00
fawney19 0112ab752b refactor(proxy): 将 proxy resolver 同步阻塞操作异步化,避免阻塞事件循环
- 为 resolve_proxy_info、resolve_delegate_config、build_proxy_url、
  get_system_proxy_config、build_post_kwargs、build_stream_kwargs 新增
  _async 异步版本,通过 asyncio.to_thread 在工作线程中执行同步 DB 查询
- 为 _proxy_node_cache 和 _system_proxy_cache 添加 threading.Lock 保护
  多线程并发读写安全
- 大 payload 的 gzip 压缩超过 64KB 阈值时走线程池,小 payload 仍在事件
  循环中同步执行以避免不必要的线程调度开销
- hub_transport 的 frame 压缩同样增加异步版本
- 删除已无调用者的同步方法 create_client_with_proxy,将其逻辑内联至
  get_upstream_client 并改为异步
- 更新所有 handler/executor/failover 调用点使用新的异步 API
- 补充 async 版本的单元测试
2026-03-12 13:59:21 +08:00
fawney19 66fec80e79 fix(frontend): 修复用户表单密码框无法输入及编辑模式额度框不显示
- 密码框: masked 属性改为始终开启,避免聚焦时 DOM 重建导致焦点丢失
- 额度框: 编辑模式下 unlimited=false 时显示"按钱包余额限制"而非空白
2026-03-12 13:31:43 +08:00
fawney19 fddfaecf5e chore: bump aether-hub version to 0.1.6 2026-03-12 12:21:52 +08:00
fawney19 71ae1a2307 feat(hub,stability): bounded outbound queue、worker liveness 检测、事件循环 watchdog 及 DB 操作异步化
- aether-hub: unbounded channel 改为 bounded channel (BoundedOutbound),队列满时标记拥塞并主动关闭连接,防止内存无限增长
- aether-hub: worker idle timeout 从命令行参数改为基于心跳的 liveness 检测,默认 60 秒
- aether-hub: 新增 ConnConfig 统一管理连接配置,新增 outbound_queue_capacity 参数
- hub_transport: 新增事件循环 watchdog,检测 lag 超过阈值时临时降级暂停新流
- gunicorn_conf: 启用 faulthandler,worker abort 时自动 dump 全部线程栈用于诊断
- health/endpoint_checker/recording: 同步 DB 操作移至 asyncio.to_thread,避免阻塞事件循环
- Dockerfile: 移除 --worker-idle-timeout 0 命令行参数,改由环境变量和默认值控制
2026-03-12 12:19:04 +08:00
fawney19 4d338ebd3d fix(priority): 允许号池聚合项在全局 Key 优先级管理中拖拽和编辑排序
之前号池(Pool)类型的 Key 在优先级管理对话框中被禁止拖拽和编辑优先级,
用户只能去号池高级设置中单独修改 global_priority,体验不一致。

现在号池聚合项可以像普通 Key 一样参与拖拽排序和点击编辑优先级,
变更会同步到 provider.pool_advanced.global_priority 并在保存时持久化。
2026-03-12 11:44:19 +08:00
535 changed files with 53857 additions and 23431 deletions
+35
View File
@@ -49,6 +49,10 @@ ADMIN_PASSWORD=admin123456
# max-requests-jitter 会自动设置为 MAX_REQUESTS/20 (5%)
# MAX_REQUESTS=4000
# glibc malloc arena 上限(默认 2)
# 降低 malloc 内存碎片,减少 gunicorn worker RSS
# MALLOC_ARENA_MAX=2
# HTTP 连接池上限(默认总预算约 200,按 worker 平分)
# 如果容器内存偏高,可继续下调;例如 2 worker 时设为 80-100
# HTTP_MAX_CONNECTIONS=100
@@ -56,6 +60,29 @@ ADMIN_PASSWORD=admin123456
# HTTP 保活连接数(默认约为 max_connections 的 30%)
# HTTP_KEEPALIVE_CONNECTIONS=30
# HTTP 代理/Tunnel 客户端空闲清理(默认每 5 分钟扫描,空闲 600 秒即关闭)
# HTTP_CLIENT_IDLE_CLEANUP_INTERVAL_MINUTES=5
# HTTP_CLIENT_IDLE_CLEANUP_MAX_SECONDS=600
# curl_cffi session 池上限(默认 20,按 impersonate + proxy 组合缓存)
# CURL_CFFI_MAX_SESSIONS=20
# 流式响应块缓存上限(单位 MB,默认 2)
# 说明:
# - 这是单个流式请求可保留的“解析后响应块”内存上限,不是全局上限
# - 粗略峰值内存 ≈ 并发流数量 × RESPONSE_CHUNKS_MAX_SIZE_MB
# 例如:100 并发、2MB 上限,理论峰值约 200MB
# - 建议:
# - 内存敏感环境:1
# - 通用生产环境:2(默认)
# - 需要更多调试上下文:4
# RESPONSE_CHUNKS_MAX_SIZE_MB=2
# 流式空闲超时(单位秒,默认 30)
# 当流已经开始但连续一段时间没有任何新 chunk 时,提前中断并返回 504,
# 避免一直等到 worker 超时(如 300s)
# STREAM_IDLE_TIMEOUT_SECONDS=30
# API Key 前缀(默认 sk)
# API_KEY_PREFIX=sk
@@ -67,6 +94,14 @@ ADMIN_PASSWORD=admin123456
# 默认: * (允许所有源)
# CORS_ORIGINS=*
# 启动预热配置(默认启用,降低首请求冷启动延迟)
# 是否启用启动期预热任务(默认 true)
# STARTUP_WARMUP_ENABLED=true
# /readyz 是否等待预热完成(默认 true)
# STARTUP_WARMUP_GATE_READINESS=true
# 预热时优先 bootstrap 的 provider_type 列表(逗号分隔;留空表示自动探测)
# STARTUP_WARMUP_PROVIDER_TYPES=codex,kiro
# ==================== 计费系统(可选) ====================
# Video/Image/Audio 缺失 billing_rule 时是否拒绝请求(默认 false:允许请求但 cost=0 并告警)
# BILLING_REQUIRE_RULE=false
-81
View File
@@ -7,12 +7,6 @@ on:
permissions:
contents: write
packages: write
env:
REGISTRY: ghcr.io
GHCR_IMAGE: fawney19/aether-hub
DOCKERHUB_IMAGE: fawney19/aether-hub
jobs:
build:
@@ -95,78 +89,3 @@ jobs:
artifacts/aether-hub-*
artifacts/SHA256SUMS.txt
fail_on_unmatched_files: true
docker:
needs: build
runs-on: ubuntu-latest
if: startsWith(github.ref, 'refs/tags/')
steps:
- uses: actions/checkout@v5
- name: Download Linux artifacts
uses: actions/download-artifact@v5
with:
pattern: aether-hub-linux-*
merge-multiple: true
path: artifacts
- name: Prepare binaries
run: |
mkdir -p aether-hub/build/linux-amd64 aether-hub/build/linux-arm64
tar xzf artifacts/aether-hub-linux-amd64.tar.gz -C aether-hub/build/linux-amd64
tar xzf artifacts/aether-hub-linux-arm64.tar.gz -C aether-hub/build/linux-arm64
- name: Generate CI Dockerfile
run: |
cat > aether-hub/Dockerfile.ci << 'EOF'
FROM debian:bookworm-slim
ARG TARGETARCH
RUN apt-get update && apt-get install -y --no-install-recommends ca-certificates && rm -rf /var/lib/apt/lists/*
COPY build/linux-${TARGETARCH}/aether-hub /usr/local/bin/aether-hub
EXPOSE 8085
ENTRYPOINT ["/usr/local/bin/aether-hub"]
CMD ["--bind", "0.0.0.0:8085"]
EOF
- name: Set up QEMU
uses: docker/setup-qemu-action@v3
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Log in to GHCR
uses: docker/login-action@v3
with:
registry: ${{ env.REGISTRY }}
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Log in to Docker Hub
uses: docker/login-action@v3
with:
username: ${{ secrets.DOCKERHUB_USERNAME }}
password: ${{ secrets.DOCKERHUB_TOKEN }}
- name: Extract metadata
id: meta
uses: docker/metadata-action@v5
with:
images: |
${{ env.REGISTRY }}/${{ env.GHCR_IMAGE }}
docker.io/${{ env.DOCKERHUB_IMAGE }}
tags: |
type=match,pattern=hub-v(.*),group=1
type=match,pattern=hub-v(\d+\.\d+),group=1
type=sha,prefix=
flavor: |
latest=auto
- name: Build and push
uses: docker/build-push-action@v6
with:
context: ./aether-hub
file: ./aether-hub/Dockerfile.ci
push: true
tags: ${{ steps.meta.outputs.tags }}
labels: ${{ steps.meta.outputs.labels }}
platforms: linux/amd64,linux/arm64
+2
View File
@@ -2,6 +2,7 @@
# Edit at https://www.toptal.com/developers/gitignore?templates=python
# AI Assistant Configuration
.codex/
.claude/
.serena/
.gemini*/
@@ -203,6 +204,7 @@ logs/
# Git backup
.git.backup/
.worktrees/
# Database backups
backups/
+66 -5
View File
@@ -25,7 +25,12 @@ RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
nginx \
supervisor \
libpq5 \
curl
curl \
libjemalloc2
RUN set -eux; \
jemalloc_path="$(find /usr/lib -type f -name 'libjemalloc.so.2' | head -n1)"; \
[ -n "$jemalloc_path" ]; \
ln -sf "$jemalloc_path" /usr/local/lib/libjemalloc.so.2
# 从 base 镜像复制 Python 包
COPY --from=builder /usr/local/lib/python3.13/site-packages /usr/local/lib/python3.13/site-packages
# 只复制需要的 Python 可执行文件
@@ -85,6 +90,11 @@ RUN printf '%s\n' \
' "" $remote_addr;' \
'}' \
'' \
'map $http_upgrade $connection_upgrade {' \
' default upgrade;' \
' "" "";' \
'}' \
'' \
'server {' \
' listen 80;' \
' server_name _;' \
@@ -124,6 +134,25 @@ RUN printf '%s\n' \
' proxy_set_header X-Forwarded-Proto $scheme;' \
' proxy_set_header Upgrade $http_upgrade;' \
' proxy_set_header Connection "upgrade";' \
' # 剥离 CF 头,防止泄露给上游或返回给客户端' \
' proxy_hide_header CF-Connecting-IP;' \
' proxy_hide_header CF-IPCountry;' \
' proxy_hide_header CF-Ray;' \
' proxy_hide_header CF-Visitor;' \
' proxy_hide_header CDN-Loop;' \
' proxy_hide_header True-Client-IP;' \
' proxy_hide_header CF-Worker;' \
' proxy_hide_header CF-EW-Via;' \
' proxy_hide_header CF-Warp-Tag-ID;' \
' proxy_set_header CF-Connecting-IP "";' \
' proxy_set_header CF-IPCountry "";' \
' proxy_set_header CF-Ray "";' \
' proxy_set_header CF-Visitor "";' \
' proxy_set_header CDN-Loop "";' \
' proxy_set_header True-Client-IP "";' \
' proxy_set_header CF-Worker "";' \
' proxy_set_header CF-EW-Via "";' \
' proxy_set_header CF-Warp-Tag-ID "";' \
' proxy_read_timeout 86400s;' \
' proxy_send_timeout 86400s;' \
' }' \
@@ -136,12 +165,22 @@ RUN printf '%s\n' \
' proxy_set_header X-Real-IP $real_ip;' \
' proxy_set_header X-Forwarded-For $forwarded_for;' \
' proxy_set_header X-Forwarded-Proto $scheme;' \
' proxy_set_header Connection "";' \
' proxy_set_header Upgrade $http_upgrade;' \
' proxy_set_header Connection $connection_upgrade;' \
' proxy_set_header Accept $http_accept;' \
' proxy_set_header Content-Type $content_type;' \
' proxy_set_header Authorization $http_authorization;' \
' proxy_set_header X-Api-Key $http_x_api_key;' \
' # 剥离 CF 头,防止泄露给上游 AI 提供商' \
' # 剥离 CF 头,防止泄露给上游或返回给客户端' \
' proxy_hide_header CF-Connecting-IP;' \
' proxy_hide_header CF-IPCountry;' \
' proxy_hide_header CF-Ray;' \
' proxy_hide_header CF-Visitor;' \
' proxy_hide_header CDN-Loop;' \
' proxy_hide_header True-Client-IP;' \
' proxy_hide_header CF-Worker;' \
' proxy_hide_header CF-EW-Via;' \
' proxy_hide_header CF-Warp-Tag-ID;' \
' proxy_set_header CF-Connecting-IP "";' \
' proxy_set_header CF-IPCountry "";' \
' proxy_set_header CF-Ray "";' \
@@ -150,6 +189,7 @@ RUN printf '%s\n' \
' proxy_set_header True-Client-IP "";' \
' proxy_set_header CF-Worker "";' \
' proxy_set_header CF-EW-Via "";' \
' proxy_set_header CF-Warp-Tag-ID "";' \
' proxy_buffering off;' \
' proxy_cache off;' \
' proxy_request_buffering off;' \
@@ -169,6 +209,25 @@ RUN printf '%s\n' \
' proxy_set_header X-Real-IP $real_ip;' \
' proxy_set_header X-Forwarded-For $forwarded_for;' \
' proxy_set_header X-Forwarded-Proto $scheme;' \
' # 剥离 CF 头,防止泄露给上游或返回给客户端' \
' proxy_hide_header CF-Connecting-IP;' \
' proxy_hide_header CF-IPCountry;' \
' proxy_hide_header CF-Ray;' \
' proxy_hide_header CF-Visitor;' \
' proxy_hide_header CDN-Loop;' \
' proxy_hide_header True-Client-IP;' \
' proxy_hide_header CF-Worker;' \
' proxy_hide_header CF-EW-Via;' \
' proxy_hide_header CF-Warp-Tag-ID;' \
' proxy_set_header CF-Connecting-IP "";' \
' proxy_set_header CF-IPCountry "";' \
' proxy_set_header CF-Ray "";' \
' proxy_set_header CF-Visitor "";' \
' proxy_set_header CDN-Loop "";' \
' proxy_set_header True-Client-IP "";' \
' proxy_set_header CF-Worker "";' \
' proxy_set_header CF-EW-Via "";' \
' proxy_set_header CF-Warp-Tag-ID "";' \
' }' \
'' \
' # 所有其他路由 → 前端 SPA(先尝试静态文件,再回退到 index.html)' \
@@ -199,10 +258,10 @@ RUN printf '%s\n' \
'stdout_logfile_maxbytes=0' \
'stderr_logfile=/dev/stderr' \
'stderr_logfile_maxbytes=0' \
'environment=PYTHONUNBUFFERED=1,PYTHONIOENCODING=utf-8,LANG=C.UTF-8,LC_ALL=C.UTF-8,DOCKER_CONTAINER=true' \
'environment=PYTHONUNBUFFERED=1,PYTHONIOENCODING=utf-8,LANG=C.UTF-8,LC_ALL=C.UTF-8,DOCKER_CONTAINER=true,LD_PRELOAD=/usr/local/lib/libjemalloc.so.2,MALLOC_CONF="background_thread:true,dirty_decay_ms:5000,muzzy_decay_ms:5000"' \
'' \
'[program:tunnel-hub]' \
'command=/usr/local/bin/aether-hub --bind 0.0.0.0:8085 --worker-idle-timeout 0' \
'command=/usr/local/bin/aether-hub --bind 0.0.0.0:8085' \
'autostart=true' \
'autorestart=true' \
'stdout_logfile=/dev/stdout' \
@@ -220,6 +279,8 @@ ENV PYTHONUNBUFFERED=1 \
PYTHONIOENCODING=utf-8 \
LANG=C.UTF-8 \
LC_ALL=C.UTF-8 \
LD_PRELOAD=/usr/local/lib/libjemalloc.so.2 \
MALLOC_CONF=background_thread:true,dirty_decay_ms:5000,muzzy_decay_ms:5000 \
PORT=8084 \
GUNICORN_WORKERS=2 \
MAX_REQUESTS=4000
+90 -17
View File
@@ -20,6 +20,11 @@ ARG HUB_RELEASE_REPO=fawney19/Aether
ARG HUB_TAG
ARG TARGETARCH
ARG GITHUB_TOKEN
# GitHub 下载镜像前缀,国内构建时传入可用的镜像加速地址
# 用法: --build-arg GITHUB_MIRROR=https://ghfast.top
# 或: --build-arg GITHUB_MIRROR=https://gh-proxy.com
# 或: --build-arg GITHUB_MIRROR=https://mirror.ghproxy.com
ARG GITHUB_MIRROR
# 运行时依赖(使用清华镜像源 + BuildKit 缓存加速)
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
@@ -29,7 +34,12 @@ RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
nginx \
supervisor \
libpq5 \
curl
curl \
libjemalloc2
RUN set -eux; \
jemalloc_path="$(find /usr/lib -type f -name 'libjemalloc.so.2' | head -n1)"; \
[ -n "$jemalloc_path" ]; \
ln -sf "$jemalloc_path" /usr/local/lib/libjemalloc.so.2
# 从 base 镜像复制 Python 包
COPY --from=builder /usr/local/lib/python3.13/site-packages /usr/local/lib/python3.13/site-packages
@@ -39,9 +49,20 @@ COPY --from=builder /usr/local/bin/gunicorn /usr/local/bin/
COPY --from=builder /usr/local/bin/uvicorn /usr/local/bin/
COPY --from=builder /usr/local/bin/alembic /usr/local/bin/
# Hub 预编译二进制(构建时从 GitHub Release 下载)
# Hub 预编译二进制
# 国内构建: --build-arg GITHUB_MIRROR=https://ghfast.top 即可走镜像下载
# GITHUB_TOKEN 可选:未认证 API 限流 60 次/小时,认证后 5000 次/小时
RUN set -eux; \
arch="${TARGETARCH:-}"; \
if [ -z "$arch" ]; then \
arch="$(dpkg --print-architecture)"; \
fi; \
case "$arch" in \
amd64|arm64) ;; \
x86_64) arch="amd64" ;; \
aarch64) arch="arm64" ;; \
*) echo "Unsupported architecture: $arch"; exit 1 ;; \
esac; \
auth_header=""; \
if [ -n "${GITHUB_TOKEN:-}" ]; then \
auth_header="Authorization: token ${GITHUB_TOKEN}"; \
@@ -54,18 +75,14 @@ RUN set -eux; \
echo "Failed to resolve hub release tag"; \
exit 1; \
fi; \
arch="${TARGETARCH:-}"; \
if [ -z "$arch" ]; then \
arch="$(dpkg --print-architecture)"; \
fi; \
case "$arch" in \
amd64|arm64) ;; \
x86_64) arch="amd64" ;; \
aarch64) arch="arm64" ;; \
*) echo "Unsupported architecture: $arch"; exit 1 ;; \
esac; \
echo "Using Hub release tag: $tag"; \
url="https://github.com/${HUB_RELEASE_REPO}/releases/download/${tag}/aether-hub-linux-${arch}.tar.gz"; \
origin_url="https://github.com/${HUB_RELEASE_REPO}/releases/download/${tag}/aether-hub-linux-${arch}.tar.gz"; \
if [ -n "${GITHUB_MIRROR:-}" ]; then \
url="${GITHUB_MIRROR}/https://github.com/${HUB_RELEASE_REPO}/releases/download/${tag}/aether-hub-linux-${arch}.tar.gz"; \
echo "Using mirror: ${GITHUB_MIRROR}"; \
else \
url="$origin_url"; \
fi; \
curl -L --fail -o /tmp/aether-hub.tar.gz "$url"; \
tar xzf /tmp/aether-hub.tar.gz -C /usr/local/bin; \
chmod +x /usr/local/bin/aether-hub; \
@@ -95,6 +112,11 @@ RUN printf '%s\n' \
' "" $remote_addr;' \
'}' \
'' \
'map $http_upgrade $connection_upgrade {' \
' default upgrade;' \
' "" "";' \
'}' \
'' \
'server {' \
' listen 80;' \
' server_name _;' \
@@ -134,6 +156,25 @@ RUN printf '%s\n' \
' proxy_set_header X-Forwarded-Proto $scheme;' \
' proxy_set_header Upgrade $http_upgrade;' \
' proxy_set_header Connection "upgrade";' \
' # 剥离 CF 头,防止泄露给上游或返回给客户端' \
' proxy_hide_header CF-Connecting-IP;' \
' proxy_hide_header CF-IPCountry;' \
' proxy_hide_header CF-Ray;' \
' proxy_hide_header CF-Visitor;' \
' proxy_hide_header CDN-Loop;' \
' proxy_hide_header True-Client-IP;' \
' proxy_hide_header CF-Worker;' \
' proxy_hide_header CF-EW-Via;' \
' proxy_hide_header CF-Warp-Tag-ID;' \
' proxy_set_header CF-Connecting-IP "";' \
' proxy_set_header CF-IPCountry "";' \
' proxy_set_header CF-Ray "";' \
' proxy_set_header CF-Visitor "";' \
' proxy_set_header CDN-Loop "";' \
' proxy_set_header True-Client-IP "";' \
' proxy_set_header CF-Worker "";' \
' proxy_set_header CF-EW-Via "";' \
' proxy_set_header CF-Warp-Tag-ID "";' \
' proxy_read_timeout 86400s;' \
' proxy_send_timeout 86400s;' \
' }' \
@@ -146,12 +187,22 @@ RUN printf '%s\n' \
' proxy_set_header X-Real-IP $real_ip;' \
' proxy_set_header X-Forwarded-For $forwarded_for;' \
' proxy_set_header X-Forwarded-Proto $scheme;' \
' proxy_set_header Connection "";' \
' proxy_set_header Upgrade $http_upgrade;' \
' proxy_set_header Connection $connection_upgrade;' \
' proxy_set_header Accept $http_accept;' \
' proxy_set_header Content-Type $content_type;' \
' proxy_set_header Authorization $http_authorization;' \
' proxy_set_header X-Api-Key $http_x_api_key;' \
' # 剥离 CF 头,防止泄露给上游 AI 提供商' \
' # 剥离 CF 头,防止泄露给上游或返回给客户端' \
' proxy_hide_header CF-Connecting-IP;' \
' proxy_hide_header CF-IPCountry;' \
' proxy_hide_header CF-Ray;' \
' proxy_hide_header CF-Visitor;' \
' proxy_hide_header CDN-Loop;' \
' proxy_hide_header True-Client-IP;' \
' proxy_hide_header CF-Worker;' \
' proxy_hide_header CF-EW-Via;' \
' proxy_hide_header CF-Warp-Tag-ID;' \
' proxy_set_header CF-Connecting-IP "";' \
' proxy_set_header CF-IPCountry "";' \
' proxy_set_header CF-Ray "";' \
@@ -160,6 +211,7 @@ RUN printf '%s\n' \
' proxy_set_header True-Client-IP "";' \
' proxy_set_header CF-Worker "";' \
' proxy_set_header CF-EW-Via "";' \
' proxy_set_header CF-Warp-Tag-ID "";' \
' proxy_buffering off;' \
' proxy_cache off;' \
' proxy_request_buffering off;' \
@@ -179,6 +231,25 @@ RUN printf '%s\n' \
' proxy_set_header X-Real-IP $real_ip;' \
' proxy_set_header X-Forwarded-For $forwarded_for;' \
' proxy_set_header X-Forwarded-Proto $scheme;' \
' # 剥离 CF 头,防止泄露给上游或返回给客户端' \
' proxy_hide_header CF-Connecting-IP;' \
' proxy_hide_header CF-IPCountry;' \
' proxy_hide_header CF-Ray;' \
' proxy_hide_header CF-Visitor;' \
' proxy_hide_header CDN-Loop;' \
' proxy_hide_header True-Client-IP;' \
' proxy_hide_header CF-Worker;' \
' proxy_hide_header CF-EW-Via;' \
' proxy_hide_header CF-Warp-Tag-ID;' \
' proxy_set_header CF-Connecting-IP "";' \
' proxy_set_header CF-IPCountry "";' \
' proxy_set_header CF-Ray "";' \
' proxy_set_header CF-Visitor "";' \
' proxy_set_header CDN-Loop "";' \
' proxy_set_header True-Client-IP "";' \
' proxy_set_header CF-Worker "";' \
' proxy_set_header CF-EW-Via "";' \
' proxy_set_header CF-Warp-Tag-ID "";' \
' }' \
'' \
' # 所有其他路由 → 前端 SPA(先尝试静态文件,再回退到 index.html)' \
@@ -210,10 +281,10 @@ RUN printf '%s\n' \
'stdout_logfile_maxbytes=0' \
'stderr_logfile=/dev/stderr' \
'stderr_logfile_maxbytes=0' \
'environment=PYTHONUNBUFFERED=1,PYTHONIOENCODING=utf-8,LANG=C.UTF-8,LC_ALL=C.UTF-8,DOCKER_CONTAINER=true' \
'environment=PYTHONUNBUFFERED=1,PYTHONIOENCODING=utf-8,LANG=C.UTF-8,LC_ALL=C.UTF-8,DOCKER_CONTAINER=true,LD_PRELOAD=/usr/local/lib/libjemalloc.so.2,MALLOC_CONF="background_thread:true,dirty_decay_ms:5000,muzzy_decay_ms:5000"' \
'' \
'[program:tunnel-hub]' \
'command=/usr/local/bin/aether-hub --bind 0.0.0.0:8085 --worker-idle-timeout 0' \
'command=/usr/local/bin/aether-hub --bind 0.0.0.0:8085' \
'autostart=true' \
'autorestart=true' \
'stdout_logfile=/dev/stdout' \
@@ -234,6 +305,8 @@ ENV PYTHONUNBUFFERED=1 \
PYTHONIOENCODING=utf-8 \
LANG=C.UTF-8 \
LC_ALL=C.UTF-8 \
LD_PRELOAD=/usr/local/lib/libjemalloc.so.2 \
MALLOC_CONF=background_thread:true,dirty_decay_ms:5000,muzzy_decay_ms:5000 \
PORT=8084 \
GUNICORN_WORKERS=2 \
MAX_REQUESTS=4000
+1
View File
@@ -64,6 +64,7 @@ cp .env.example .env
python generate_keys.py # 生成密钥, 并将生成的密钥填入 .env
# 3. 部署 / 更新(自动构建、启动、迁移)
git pull
./deploy.sh
```
+794 -11
View File
File diff suppressed because it is too large Load Diff
+5 -1
View File
@@ -1,6 +1,6 @@
[package]
name = "aether-hub"
version = "0.1.5"
version = "0.2.0"
edition = "2021"
description = "Tunnel Hub for Aether - frame router between workers and proxies"
@@ -16,6 +16,10 @@ dashmap = "6"
parking_lot = "0.12"
flate2 = "1"
futures-util = "0.3"
bytes = "1"
async-stream = "0.3"
http-body-util = "0.1"
reqwest = { version = "0.12", default-features = false, features = ["json", "rustls-tls"] }
[profile.release]
lto = true
+5
View File
@@ -28,6 +28,11 @@ cd /path/to/Aether
- `--load`: 加载到本地 Docker(单平台)
- `--latest`: 额外打 `latest` tag
## 运行时参数
- `TUNNEL_HUB_WORKER_IDLE_TIMEOUT`:worker 心跳空闲超时,默认 `60` 秒
- `TUNNEL_HUB_OUTBOUND_QUEUE_CAPACITY`:单连接出站队列容量,默认 `128`;队列打满时会把连接视为拥塞并主动关闭,避免 Hub 内存无限增长
## 与部署脚本关系
- `./deploy.sh`: 本地构建部署(会本地构建 app/base,并在构建 app 时从 GitHub Release 下载 Hub,可用 `--hub-tag` 固定版本)。
+85
View File
@@ -0,0 +1,85 @@
use reqwest::Client;
#[derive(Clone)]
pub struct ControlPlaneClient {
client: Option<Client>,
base_url: String,
}
impl ControlPlaneClient {
pub fn new(base_url: String) -> Self {
let client = Client::builder()
.timeout(std::time::Duration::from_secs(10))
.build()
.ok();
Self { client, base_url }
}
pub fn disabled() -> Self {
Self {
client: None,
base_url: String::new(),
}
}
pub async fn heartbeat_ack(&self, payload: &[u8]) -> Result<Vec<u8>, String> {
let Some(client) = &self.client else {
return Ok(b"{}".to_vec());
};
let url = format!(
"{}/api/internal/hub/heartbeat",
self.base_url.trim_end_matches('/')
);
let response = client
.post(&url)
.header("content-type", "application/json")
.body(payload.to_vec())
.send()
.await
.map_err(|e| format!("heartbeat callback request failed: {e}"))?;
if !response.status().is_success() {
return Err(format!(
"heartbeat callback failed with status {}",
response.status()
));
}
response
.bytes()
.await
.map(|bytes| bytes.to_vec())
.map_err(|e| format!("heartbeat callback body read failed: {e}"))
}
pub async fn push_node_status(
&self,
node_id: &str,
connected: bool,
conn_count: usize,
) -> Result<(), String> {
let Some(client) = &self.client else {
return Ok(());
};
let url = format!(
"{}/api/internal/hub/node-status",
self.base_url.trim_end_matches('/')
);
let response = client
.post(&url)
.json(&serde_json::json!({
"node_id": node_id,
"connected": connected,
"conn_count": conn_count,
}))
.send()
.await
.map_err(|e| format!("node-status callback request failed: {e}"))?;
if response.status().is_success() {
Ok(())
} else {
Err(format!(
"node-status callback failed with status {}",
response.status()
))
}
}
}
+637 -477
View File
File diff suppressed because it is too large Load Diff
+262
View File
@@ -0,0 +1,262 @@
use std::io;
use std::net::SocketAddr;
use std::time::Duration;
use async_stream::stream;
use axum::body::{Body, Bytes};
use axum::extract::{ConnectInfo, Path, Request, State};
use axum::http::{HeaderMap, HeaderName, HeaderValue, Response, StatusCode};
use axum::response::IntoResponse;
use bytes::BytesMut;
use futures_util::StreamExt;
use tracing::warn;
use crate::hub::{LocalBodyEvent, LocalStream};
use crate::protocol;
use crate::AppState;
pub const TUNNEL_ERROR_HEADER: &str = "x-aether-tunnel-error";
const MAX_RELAY_META_LEN: usize = 256 * 1024;
struct StreamGuard {
hub: std::sync::Arc<crate::hub::HubRouter>,
stream_id: u64,
finished: bool,
}
impl Drop for StreamGuard {
fn drop(&mut self) {
if !self.finished {
self.hub
.cancel_local_stream(self.stream_id, "local relay client dropped");
}
}
}
pub async fn relay_request(
Path(node_id): Path<String>,
State(state): State<AppState>,
ConnectInfo(addr): ConnectInfo<SocketAddr>,
request: Request,
) -> impl IntoResponse {
if !addr.ip().is_loopback() {
return tunnel_error_response(
StatusCode::FORBIDDEN,
"forbidden",
"local relay only accepts loopback requests",
);
}
let mut body_stream = request.into_body().into_data_stream();
let mut envelope_buf = BytesMut::new();
let mut meta: Option<protocol::RequestMeta> = None;
let mut stream: Option<std::sync::Arc<LocalStream>> = None;
while let Some(chunk_result) = body_stream.next().await {
let chunk = match chunk_result {
Ok(chunk) => chunk,
Err(error) => {
if let Some(active_stream) = &stream {
state
.hub
.cancel_local_stream(active_stream.id, "failed to read relay request body");
}
warn!(error = %error, "failed to read local relay request body");
return tunnel_error_response(
StatusCode::BAD_GATEWAY,
"relay",
"failed to read relay request body",
);
}
};
if stream.is_none() {
envelope_buf.extend_from_slice(&chunk);
let Some((parsed_meta, body_offset)) = (match try_decode_envelope_meta(&envelope_buf) {
Ok(result) => result,
Err(error) => {
return tunnel_error_response(StatusCode::BAD_REQUEST, "bad_request", &error);
}
}) else {
continue;
};
let opened_stream = match state.hub.open_local_stream(&node_id, &parsed_meta) {
Ok(stream) => stream,
Err(error) => {
return tunnel_error_response(
StatusCode::SERVICE_UNAVAILABLE,
"connect",
&error,
);
}
};
if envelope_buf.len() > body_offset {
let first_body_chunk = Bytes::copy_from_slice(&envelope_buf[body_offset..]);
if let Err(error) =
state
.hub
.push_local_request_body(opened_stream.id, first_body_chunk, false)
{
state.hub.cancel_local_stream(opened_stream.id, &error);
return tunnel_error_response(
StatusCode::SERVICE_UNAVAILABLE,
"connect",
&error,
);
}
}
envelope_buf.clear();
meta = Some(parsed_meta);
stream = Some(opened_stream);
continue;
}
let Some(active_stream) = &stream else {
continue;
};
if let Err(error) = state
.hub
.push_local_request_body(active_stream.id, chunk, false)
{
state.hub.cancel_local_stream(active_stream.id, &error);
return tunnel_error_response(StatusCode::SERVICE_UNAVAILABLE, "connect", &error);
}
}
let (meta, stream) = match (meta, stream) {
(Some(meta), Some(stream)) => (meta, stream),
_ => {
return tunnel_error_response(
StatusCode::BAD_REQUEST,
"bad_request",
"relay envelope metadata truncated",
);
}
};
if let Err(error) = state
.hub
.push_local_request_body(stream.id, Bytes::new(), true)
{
state.hub.cancel_local_stream(stream.id, &error);
return tunnel_error_response(StatusCode::SERVICE_UNAVAILABLE, "connect", &error);
}
let request_guard = StreamGuard {
hub: state.hub.clone(),
stream_id: stream.id,
finished: false,
};
let wait_timeout = Duration::from_secs(meta.timeout.clamp(5, 300));
let response_head = match stream.wait_headers(wait_timeout).await {
Ok(response) => response,
Err(error) => {
state.hub.cancel_local_stream(stream.id, &error);
return tunnel_error_response(StatusCode::GATEWAY_TIMEOUT, "timeout", &error);
}
};
let Some(mut body_rx) = stream.take_body_receiver() else {
state
.hub
.cancel_local_stream(stream.id, "missing relay response body receiver");
return tunnel_error_response(
StatusCode::BAD_GATEWAY,
"relay",
"missing relay response body receiver",
);
};
let hub = state.hub.clone();
let stream_id = stream.id;
let body_stream = stream! {
let mut guard = request_guard;
guard.hub = hub;
guard.stream_id = stream_id;
while let Some(event) = body_rx.recv().await {
match event {
LocalBodyEvent::Chunk(chunk) => yield Ok::<Bytes, io::Error>(chunk),
LocalBodyEvent::End => {
guard.finished = true;
break;
}
LocalBodyEvent::Error(error) => {
guard.finished = true;
yield Err(io::Error::other(error));
break;
}
}
}
guard.finished = true;
};
let mut builder = Response::builder().status(response_head.status);
if let Some(headers) = builder.headers_mut() {
append_headers(headers, &response_head.headers);
}
match builder.body(Body::from_stream(body_stream)) {
Ok(response) => response,
Err(error) => {
warn!(error = %error, "failed to build relay response");
tunnel_error_response(
StatusCode::BAD_GATEWAY,
"relay",
"failed to build relay response",
)
}
}
}
fn try_decode_envelope_meta(
buffer: &BytesMut,
) -> Result<Option<(protocol::RequestMeta, usize)>, String> {
if buffer.len() < 4 {
return Ok(None);
}
let meta_len = u32::from_be_bytes([buffer[0], buffer[1], buffer[2], buffer[3]]) as usize;
if meta_len > MAX_RELAY_META_LEN {
return Err("relay metadata too large".to_string());
}
let meta_end = 4usize
.checked_add(meta_len)
.ok_or_else(|| "relay envelope length overflow".to_string())?;
if buffer.len() < meta_end {
return Ok(None);
}
let meta = serde_json::from_slice::<protocol::RequestMeta>(&buffer[4..meta_end])
.map_err(|e| format!("invalid relay metadata: {e}"))?;
Ok(Some((meta, meta_end)))
}
fn append_headers(target: &mut HeaderMap, headers: &[(String, String)]) {
for (name, value) in headers {
let Ok(name) = HeaderName::from_bytes(name.as_bytes()) else {
continue;
};
let Ok(value) = HeaderValue::from_str(value) else {
continue;
};
target.append(name, value);
}
}
fn tunnel_error_response(status: StatusCode, kind: &str, message: &str) -> Response<Body> {
let mut builder = Response::builder().status(status);
if let Some(headers) = builder.headers_mut() {
headers.insert(
HeaderName::from_static(TUNNEL_ERROR_HEADER),
HeaderValue::from_str(kind).unwrap_or_else(|_| HeaderValue::from_static("relay")),
);
headers.insert(
axum::http::header::CONTENT_TYPE,
HeaderValue::from_static("text/plain; charset=utf-8"),
);
}
builder
.body(Body::from(message.to_string()))
.unwrap_or_else(|_| Response::new(Body::from("relay error")))
}
+42 -34
View File
@@ -1,20 +1,23 @@
mod control_plane;
mod hub;
mod local_relay;
mod protocol;
mod proxy_conn;
mod worker_conn;
use std::sync::Arc;
use std::net::SocketAddr;
use std::time::Duration;
use axum::extract::ws::WebSocketUpgrade;
use axum::extract::State;
use axum::response::{IntoResponse, Json};
use axum::routing::get;
use axum::routing::{get, post};
use axum::Router;
use clap::Parser;
use tracing::{info, warn};
use crate::hub::HubRouter;
use crate::control_plane::ControlPlaneClient;
use crate::hub::{ConnConfig, HubRouter};
use crate::local_relay::relay_request;
#[derive(Parser, Debug)]
#[command(name = "aether-hub", about = "Tunnel Hub for Aether")]
@@ -27,10 +30,6 @@ struct Args {
#[arg(long, default_value_t = 0, env = "TUNNEL_HUB_PROXY_IDLE_TIMEOUT")]
proxy_idle_timeout: u64,
/// Worker-side idle timeout in seconds (0 to disable)
#[arg(long, default_value_t = 0, env = "TUNNEL_HUB_WORKER_IDLE_TIMEOUT")]
worker_idle_timeout: u64,
/// Ping interval in seconds (for both sides)
#[arg(long, default_value_t = 15, env = "TUNNEL_HUB_PING_INTERVAL")]
ping_interval: u64,
@@ -38,15 +37,29 @@ struct Args {
/// Max concurrent streams per proxy connection
#[arg(long, default_value_t = 2048, env = "TUNNEL_HUB_MAX_STREAMS")]
max_streams: usize,
/// Per-connection outbound queue capacity before treating the socket as congested
#[arg(
long,
default_value_t = 128,
env = "TUNNEL_HUB_OUTBOUND_QUEUE_CAPACITY"
)]
outbound_queue_capacity: usize,
/// Local Aether app base URL for control-plane callbacks
#[arg(
long,
default_value = "http://127.0.0.1:8084",
env = "TUNNEL_HUB_APP_BASE_URL"
)]
app_base_url: String,
}
#[derive(Clone)]
struct AppState {
hub: Arc<HubRouter>,
proxy_idle_timeout: Duration,
worker_idle_timeout: Duration,
ping_interval: Duration,
max_streams: usize,
pub struct AppState {
pub hub: std::sync::Arc<HubRouter>,
pub proxy_conn_cfg: ConnConfig,
pub max_streams: usize,
}
#[tokio::main]
@@ -61,12 +74,16 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
let args = Args::parse();
let hub = HubRouter::new();
let hub = HubRouter::new(ControlPlaneClient::new(args.app_base_url));
let outbound_queue_capacity = args.outbound_queue_capacity.clamp(8, 4096);
let ping_interval = Duration::from_secs(args.ping_interval);
let state = AppState {
hub,
proxy_idle_timeout: Duration::from_secs(args.proxy_idle_timeout),
worker_idle_timeout: Duration::from_secs(args.worker_idle_timeout),
ping_interval: Duration::from_secs(args.ping_interval),
proxy_conn_cfg: ConnConfig {
ping_interval,
idle_timeout: Duration::from_secs(args.proxy_idle_timeout),
outbound_queue_capacity,
},
max_streams: args.max_streams,
};
@@ -74,13 +91,17 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
.route("/health", get(health))
.route("/stats", get(stats))
.route("/proxy", get(ws_proxy))
.route("/worker", get(ws_worker))
.route("/local/relay/{node_id}", post(relay_request))
.with_state(state);
let listener = tokio::net::TcpListener::bind(&args.bind).await?;
info!(bind = %args.bind, "aether-hub started");
axum::serve(listener, app).await?;
axum::serve(
listener,
app.into_make_service_with_connect_info::<SocketAddr>(),
)
.await?;
Ok(())
}
@@ -139,21 +160,8 @@ async fn ws_proxy(
node_id,
node_name,
max_streams,
state.ping_interval,
state.proxy_idle_timeout,
state.proxy_conn_cfg,
)
})
.into_response()
}
async fn ws_worker(ws: WebSocketUpgrade, State(state): State<AppState>) -> impl IntoResponse {
ws.max_frame_size(64 * 1024 * 1024)
.on_upgrade(move |socket| {
worker_conn::handle_worker_connection(
socket,
state.hub,
state.ping_interval,
state.worker_idle_timeout,
)
})
}
+79 -135
View File
@@ -22,8 +22,6 @@ pub const PONG: u8 = 0x11;
pub const GOAWAY: u8 = 0x12;
pub const HEARTBEAT_DATA: u8 = 0x13;
pub const HEARTBEAT_ACK: u8 = 0x14;
pub const NODE_STATUS: u8 = 0x15;
// Flags
pub const FLAG_END_STREAM: u8 = 0x01;
pub const FLAG_GZIP_COMPRESSED: u8 = 0x02;
@@ -50,162 +48,89 @@ impl FrameHeader {
payload_len: u32::from_be_bytes([data[6], data[7], data[8], data[9]]),
})
}
}
/// Check if this is a stream-terminating frame
#[inline]
pub fn is_stream_terminal(&self) -> bool {
self.msg_type == STREAM_END || self.msg_type == STREAM_ERROR
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
pub struct RequestMeta {
pub method: String,
pub url: String,
pub headers: std::collections::HashMap<String, String>,
#[serde(default = "default_timeout", deserialize_with = "deserialize_timeout")]
pub timeout: u64,
}
fn default_timeout() -> u64 {
60
}
fn deserialize_timeout<'de, D>(deserializer: D) -> Result<u64, D::Error>
where
D: serde::Deserializer<'de>,
{
#[derive(serde::Deserialize)]
#[serde(untagged)]
enum TimeoutValue {
Int(u64),
Float(f64),
}
match <TimeoutValue as serde::Deserialize>::deserialize(deserializer)? {
TimeoutValue::Int(v) => Ok(v),
TimeoutValue::Float(v) => {
if !v.is_finite() || v < 0.0 {
return Err(serde::de::Error::custom(
"timeout must be a non-negative finite number",
));
}
if v.fract() != 0.0 {
return Err(serde::de::Error::custom("timeout must be integer seconds"));
}
if v > (u64::MAX as f64) {
return Err(serde::de::Error::custom("timeout is too large"));
}
Ok(v as u64)
}
}
}
#[derive(Debug)]
pub struct RequestHeadersExtracted {
pub node_id: String,
pub rebuilt_frame: Vec<u8>,
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
pub struct ResponseMeta {
pub status: u16,
pub headers: Vec<(String, String)>,
}
pub fn encode_frame(stream_id: u32, msg_type: u8, flags: u8, payload: &[u8]) -> Vec<u8> {
let mut buf = Vec::with_capacity(HEADER_SIZE + payload.len());
buf.extend_from_slice(&stream_id.to_be_bytes());
buf.push(msg_type);
buf.push(flags);
buf.extend_from_slice(&(payload.len() as u32).to_be_bytes());
buf.extend_from_slice(payload);
buf
}
/// Encode a STREAM_ERROR frame for a given stream_id with an error message
pub fn encode_stream_error(stream_id: u32, msg: &str) -> Vec<u8> {
let payload = msg.as_bytes();
let mut buf = Vec::with_capacity(HEADER_SIZE + payload.len());
buf.extend_from_slice(&stream_id.to_be_bytes());
buf.push(STREAM_ERROR);
buf.push(0); // flags
buf.extend_from_slice(&(payload.len() as u32).to_be_bytes());
buf.extend_from_slice(payload);
buf
}
/// Encode a NODE_STATUS frame (stream_id=0, Hub-generated)
pub fn encode_node_status(node_id: &str, connected: bool, conn_count: usize) -> Vec<u8> {
let payload = serde_json::json!({
"node_id": node_id,
"connected": connected,
"conn_count": conn_count,
});
let payload_bytes = payload.to_string().into_bytes();
let mut buf = Vec::with_capacity(HEADER_SIZE + payload_bytes.len());
buf.extend_from_slice(&0u32.to_be_bytes()); // stream_id = 0
buf.push(NODE_STATUS);
buf.push(0); // flags
buf.extend_from_slice(&(payload_bytes.len() as u32).to_be_bytes());
buf.extend_from_slice(&payload_bytes);
buf
encode_frame(stream_id, STREAM_ERROR, 0, msg.as_bytes())
}
/// Encode a PING frame (stream_id=0)
pub fn encode_ping() -> Vec<u8> {
let mut buf = Vec::with_capacity(HEADER_SIZE);
buf.extend_from_slice(&0u32.to_be_bytes());
buf.push(PING);
buf.push(0);
buf.extend_from_slice(&0u32.to_be_bytes());
buf
encode_frame(0, PING, 0, &[])
}
/// Encode a PONG frame (stream_id=0, echo payload)
pub fn encode_pong(payload: &[u8]) -> Vec<u8> {
let mut buf = Vec::with_capacity(HEADER_SIZE + payload.len());
buf.extend_from_slice(&0u32.to_be_bytes());
buf.push(PONG);
buf.push(0);
buf.extend_from_slice(&(payload.len() as u32).to_be_bytes());
buf.extend_from_slice(payload);
buf
encode_frame(0, PONG, 0, payload)
}
/// Encode a GOAWAY frame (stream_id=0)
pub fn encode_goaway() -> Vec<u8> {
let mut buf = Vec::with_capacity(HEADER_SIZE);
buf.extend_from_slice(&0u32.to_be_bytes());
buf.push(GOAWAY);
buf.push(0);
buf.extend_from_slice(&0u32.to_be_bytes());
buf
}
/// Rewrite the stream_id in raw frame bytes (first 4 bytes) -- near zero-copy
#[inline]
pub fn rewrite_stream_id(data: &mut [u8], new_stream_id: u32) {
let bytes = new_stream_id.to_be_bytes();
data[0] = bytes[0];
data[1] = bytes[1];
data[2] = bytes[2];
data[3] = bytes[3];
}
/// Get the payload portion of a raw frame (after the 10-byte header)
#[inline]
pub fn frame_payload(data: &[u8]) -> &[u8] {
if data.len() > HEADER_SIZE {
&data[HEADER_SIZE..]
} else {
&[]
}
}
/// Parse REQUEST_HEADERS payload, extract `node_id`, strip it from JSON,
/// and rebuild a new REQUEST_HEADERS frame with `new_stream_id`.
///
/// If the source frame is gzip-compressed, this function will decode it first,
/// then try to re-encode with gzip (only keeps compression when payload shrinks).
pub fn rebuild_request_headers_without_node_id(
data: &[u8],
new_stream_id: u32,
) -> Result<RequestHeadersExtracted, String> {
let header = FrameHeader::parse(data).ok_or_else(|| "invalid frame header".to_string())?;
if header.msg_type != REQUEST_HEADERS {
return Err("frame is not REQUEST_HEADERS".to_string());
}
let payload = frame_payload_by_header(data, &header)
.ok_or_else(|| "incomplete REQUEST_HEADERS payload".to_string())?;
let decoded_payload = if header.flags & FLAG_GZIP_COMPRESSED != 0 {
let mut decoder = GzDecoder::new(payload);
let mut decoded = Vec::new();
decoder
.read_to_end(&mut decoded)
.map_err(|e| format!("failed to decompress REQUEST_HEADERS: {e}"))?;
decoded
} else {
payload.to_vec()
};
let mut meta: serde_json::Value = serde_json::from_slice(&decoded_payload)
.map_err(|e| format!("invalid REQUEST_HEADERS JSON: {e}"))?;
let obj = meta
.as_object_mut()
.ok_or_else(|| "REQUEST_HEADERS payload must be a JSON object".to_string())?;
let node_id = obj
.remove("node_id")
.and_then(|v| v.as_str().map(|s| s.to_string()))
.map(|s| s.trim().to_string())
.filter(|s| !s.is_empty())
.ok_or_else(|| "missing node_id in REQUEST_HEADERS".to_string())?;
let stripped_payload = serde_json::to_vec(&meta)
.map_err(|e| format!("failed to encode REQUEST_HEADERS payload: {e}"))?;
let (final_payload, flags) =
maybe_recompress_payload(&stripped_payload, header.flags & FLAG_GZIP_COMPRESSED != 0)
.map_err(|e| format!("failed to recompress REQUEST_HEADERS payload: {e}"))?;
let mut rebuilt = Vec::with_capacity(HEADER_SIZE + final_payload.len());
rebuilt.extend_from_slice(&new_stream_id.to_be_bytes());
rebuilt.push(REQUEST_HEADERS);
rebuilt.push(flags);
rebuilt.extend_from_slice(&(final_payload.len() as u32).to_be_bytes());
rebuilt.extend_from_slice(&final_payload);
Ok(RequestHeadersExtracted {
node_id,
rebuilt_frame: rebuilt,
})
encode_frame(0, GOAWAY, 0, &[])
}
#[inline]
fn frame_payload_by_header<'a>(data: &'a [u8], header: &FrameHeader) -> Option<&'a [u8]> {
pub fn frame_payload_by_header<'a>(data: &'a [u8], header: &FrameHeader) -> Option<&'a [u8]> {
let payload_len = header.payload_len as usize;
let end = HEADER_SIZE.checked_add(payload_len)?;
if data.len() < end {
@@ -214,6 +139,25 @@ fn frame_payload_by_header<'a>(data: &'a [u8], header: &FrameHeader) -> Option<&
Some(&data[HEADER_SIZE..end])
}
pub fn decode_payload(data: &[u8], header: &FrameHeader) -> Result<Vec<u8>, String> {
let payload = frame_payload_by_header(data, header)
.ok_or_else(|| "incomplete frame payload".to_string())?;
if header.flags & FLAG_GZIP_COMPRESSED != 0 {
let mut decoder = GzDecoder::new(payload);
let mut decoded = Vec::new();
decoder
.read_to_end(&mut decoded)
.map_err(|e| format!("failed to decompress payload: {e}"))?;
Ok(decoded)
} else {
Ok(payload.to_vec())
}
}
pub fn compress_payload(payload: &[u8]) -> Result<(Vec<u8>, u8), std::io::Error> {
maybe_recompress_payload(payload, true)
}
fn maybe_recompress_payload(
payload: &[u8],
prefer_gzip: bool,
+42 -40
View File
@@ -7,10 +7,10 @@ use std::time::Duration;
use axum::extract::ws::{Message, WebSocket};
use futures_util::{SinkExt, StreamExt};
use tokio::sync::mpsc;
use tokio::sync::{mpsc, watch};
use tracing::{debug, info, warn};
use crate::hub::{HubRouter, ProxyConn};
use crate::hub::{ConnConfig, HubRouter, ProxyConn, SendStatus};
use crate::protocol;
/// Maximum single frame size: 64 MB
@@ -22,69 +22,71 @@ pub async fn handle_proxy_connection(
node_id: String,
node_name: String,
max_streams: usize,
ping_interval: Duration,
idle_timeout: Duration,
cfg: ConnConfig,
) {
let conn_id = hub.alloc_conn_id();
let (mut ws_tx, ws_rx) = ws.split();
// Create channel for outbound messages
let (tx, mut rx) = mpsc::unbounded_channel::<Message>();
let (tx, mut rx) = mpsc::channel::<Message>(cfg.outbound_queue_capacity);
let (close_tx, mut close_rx) = watch::channel(false);
let conn = Arc::new(ProxyConn::new(
conn_id,
node_id.clone(),
node_name.clone(),
tx,
close_tx,
max_streams,
));
hub.register_proxy(conn.clone());
// Spawn writer task: drains channel -> WebSocket
let writer = tokio::spawn(async move {
while let Some(msg) = rx.recv().await {
if ws_tx.send(msg).await.is_err() {
break;
loop {
tokio::select! {
msg = rx.recv() => match msg {
Some(msg) => {
if ws_tx.send(msg).await.is_err() {
break;
}
}
None => break,
},
changed = close_rx.changed() => {
if changed.is_err() || *close_rx.borrow() {
break;
}
}
}
}
let _ = ws_tx.close().await;
});
// Spawn ping task
let ping_tx = conn.tx.clone();
let ping_conn = conn.clone();
let ping_interval = cfg.ping_interval;
let ping_task = tokio::spawn(async move {
loop {
tokio::time::sleep(ping_interval).await;
let ping = protocol::encode_ping();
if ping_tx.send(Message::Binary(ping.into())).is_err() {
if !matches!(
ping_conn.send(Message::Binary(ping.into())),
SendStatus::Queued
) {
break;
}
}
});
// Spawn reader task
let reader_hub = hub.clone();
let reader_node_id = node_id.clone();
let reader_tx = conn.tx.clone();
let reader_conn = conn.clone();
let reader = tokio::spawn(async move {
run_proxy_reader(
ws_rx,
reader_hub,
conn_id,
reader_node_id,
reader_tx,
idle_timeout,
)
.await;
run_proxy_reader(ws_rx, reader_hub, reader_conn, cfg.idle_timeout).await;
});
// Wait for reader to end, then cleanup.
let _ = reader.await;
ping_task.abort();
conn.request_close();
hub.unregister_proxy(conn_id, &node_id);
// conn still holds an Arc<ProxyConn> with a channel sender clone.
// Drop it so the writer can drain and exit.
drop(conn);
tokio::time::sleep(Duration::from_millis(100)).await;
writer.abort();
@@ -94,9 +96,7 @@ pub async fn handle_proxy_connection(
async fn run_proxy_reader(
mut ws_rx: futures_util::stream::SplitStream<WebSocket>,
hub: Arc<HubRouter>,
conn_id: u64,
node_id: String,
tx: mpsc::UnboundedSender<Message>,
conn: Arc<ProxyConn>,
idle_timeout: Duration,
) {
let idle_enabled = !idle_timeout.is_zero();
@@ -106,8 +106,9 @@ async fn run_proxy_reader(
tokio::select! {
msg = ws_rx.next() => msg,
_ = tokio::time::sleep(idle_timeout) => {
warn!(conn_id = conn_id, node_id = %node_id, "proxy idle timeout");
let _ = tx.send(Message::Binary(protocol::encode_goaway().into()));
warn!(conn_id = conn.id, node_id = %conn.node_id, "proxy idle timeout");
let _ = conn.send(Message::Binary(protocol::encode_goaway().into()));
conn.request_close();
break;
}
}
@@ -121,12 +122,13 @@ async fn run_proxy_reader(
if data.len() > MAX_FRAME_SIZE {
oversized_count += 1;
warn!(
conn_id = conn_id,
conn_id = conn.id,
size = data.len(),
"oversized frame from proxy"
);
if oversized_count >= 5 {
warn!(conn_id = conn_id, "too many oversized frames, closing");
warn!(conn_id = conn.id, "too many oversized frames, closing");
conn.request_close();
break;
}
continue;
@@ -134,21 +136,21 @@ async fn run_proxy_reader(
oversized_count = 0;
if data.len() < protocol::HEADER_SIZE {
debug!(conn_id = conn_id, "frame too small, skipping");
debug!(conn_id = conn.id, "frame too small, skipping");
continue;
}
hub.handle_proxy_frame(conn_id, &mut data);
hub.handle_proxy_frame(conn.id, &mut data).await;
}
Some(Ok(Message::Close(_))) | None => {
info!(conn_id = conn_id, node_id = %node_id, "proxy WebSocket closed");
info!(conn_id = conn.id, node_id = %conn.node_id, "proxy WebSocket closed");
break;
}
Some(Err(e)) => {
warn!(conn_id = conn_id, error = %e, "proxy WebSocket error");
warn!(conn_id = conn.id, error = %e, "proxy WebSocket error");
break;
}
_ => {} // Ignore text/ping/pong at WS level
_ => {}
}
}
}
-141
View File
@@ -1,141 +0,0 @@
/// Worker-side WebSocket connection handler
///
/// Handles the lifecycle of a single Gunicorn worker connection:
/// accept -> read loop (route frames via Hub) -> cleanup
use std::sync::Arc;
use std::time::Duration;
use axum::extract::ws::{Message, WebSocket};
use futures_util::{SinkExt, StreamExt};
use tokio::sync::mpsc;
use tracing::{debug, info, warn};
use crate::hub::{HubRouter, WorkerConn};
use crate::protocol;
pub async fn handle_worker_connection(
ws: WebSocket,
hub: Arc<HubRouter>,
ping_interval: Duration,
idle_timeout: Duration,
) {
let conn_id = hub.alloc_conn_id();
let (mut ws_tx, ws_rx) = ws.split();
// Create channel for outbound messages
let (tx, mut rx) = mpsc::unbounded_channel::<Message>();
let conn = Arc::new(WorkerConn::new(conn_id, tx));
hub.register_worker(conn.clone());
// Spawn writer task
let writer = tokio::spawn(async move {
while let Some(msg) = rx.recv().await {
if ws_tx.send(msg).await.is_err() {
break;
}
}
let _ = ws_tx.close().await;
});
// Spawn ping task
let ping_tx = conn.tx.clone();
let ping_task = tokio::spawn(async move {
loop {
tokio::time::sleep(ping_interval).await;
let ping = protocol::encode_ping();
if ping_tx.send(Message::Binary(ping.into())).is_err() {
break;
}
}
});
// Spawn reader task
let reader_hub = hub.clone();
let reader_tx = conn.tx.clone();
let reader = tokio::spawn(async move {
run_worker_reader(
ws_rx,
reader_hub,
conn_id,
conn.clone(),
reader_tx,
idle_timeout,
)
.await;
});
// Wait for reader to end, then cleanup.
let _ = reader.await;
ping_task.abort();
// Unregister first so all channel senders are dropped (reader_tx dropped
// when reader completes, ping_tx dropped by abort, conn.tx dropped when
// the last Arc<WorkerConn> is removed from hub). This lets the writer
// drain buffered messages (e.g. GOAWAY) before we force-abort it.
hub.unregister_worker(conn_id);
tokio::time::sleep(Duration::from_millis(100)).await;
writer.abort();
let _ = writer.await;
}
async fn run_worker_reader(
mut ws_rx: futures_util::stream::SplitStream<WebSocket>,
hub: Arc<HubRouter>,
conn_id: u64,
conn: Arc<WorkerConn>,
tx: mpsc::UnboundedSender<Message>,
idle_timeout: Duration,
) {
let idle_enabled = !idle_timeout.is_zero();
loop {
let msg = if idle_enabled {
tokio::select! {
msg = ws_rx.next() => msg,
_ = tokio::time::sleep(idle_timeout) => {
warn!(worker_id = conn_id, "worker idle timeout");
let _ = tx.send(Message::Binary(protocol::encode_goaway().into()));
break;
}
}
} else {
ws_rx.next().await
};
match msg {
Some(Ok(Message::Binary(data))) => {
let mut data = data.to_vec();
if data.len() < protocol::HEADER_SIZE {
debug!(worker_id = conn_id, "frame too small, skipping");
continue;
}
let header = match protocol::FrameHeader::parse(&data) {
Some(h) => h,
None => continue,
};
// HEARTBEAT_ACK from worker -> route back to proxy
if header.msg_type == protocol::HEARTBEAT_ACK {
hub.handle_worker_heartbeat_ack(&mut data);
continue;
}
// Regular frames: route via hub
if let Some(err_msg) = hub.handle_worker_frame(conn_id, &mut data) {
// Send STREAM_ERROR back to worker
let err_frame = protocol::encode_stream_error(header.stream_id, &err_msg);
let _ = conn.send(Message::Binary(err_frame.into()));
}
}
Some(Ok(Message::Close(_))) | None => {
info!(worker_id = conn_id, "worker WebSocket closed");
break;
}
Some(Err(e)) => {
warn!(worker_id = conn_id, error = %e, "worker WebSocket error");
break;
}
_ => {} // Ignore text/ping/pong at WS level
}
}
}
+1 -1
View File
@@ -10,7 +10,7 @@ checksum = "320119579fcad9c21884f5c4861d16174d0e06250625266f50fe6898340abefa"
[[package]]
name = "aether-proxy"
version = "0.2.3"
version = "0.2.4"
dependencies = [
"anyhow",
"arc-swap",
+1 -1
View File
@@ -1,6 +1,6 @@
[package]
name = "aether-proxy"
version = "0.2.4"
version = "0.2.5"
edition = "2021"
description = "Tunnel proxy for Aether"
+5 -5
View File
@@ -19,11 +19,11 @@ docker compose up -d
<!-- DOWNLOAD_TABLE_START -->
| Platform | Download |
|----------|----------|
| Linux x86_64 | [aether-proxy-linux-amd64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.4/aether-proxy-linux-amd64.tar.gz) |
| Linux ARM64 | [aether-proxy-linux-arm64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.4/aether-proxy-linux-arm64.tar.gz) |
| macOS x86_64 | [aether-proxy-macos-amd64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.4/aether-proxy-macos-amd64.tar.gz) |
| macOS ARM64 | [aether-proxy-macos-arm64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.4/aether-proxy-macos-arm64.tar.gz) |
| Windows x86_64 | [aether-proxy-windows-amd64.zip](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.4/aether-proxy-windows-amd64.zip) |
| Linux x86_64 | [aether-proxy-linux-amd64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.5/aether-proxy-linux-amd64.tar.gz) |
| Linux ARM64 | [aether-proxy-linux-arm64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.5/aether-proxy-linux-arm64.tar.gz) |
| macOS x86_64 | [aether-proxy-macos-amd64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.5/aether-proxy-macos-amd64.tar.gz) |
| macOS ARM64 | [aether-proxy-macos-arm64.tar.gz](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.5/aether-proxy-macos-arm64.tar.gz) |
| Windows x86_64 | [aether-proxy-windows-amd64.zip](https://github.com/fawney19/Aether/releases/download/proxy-v0.2.5/aether-proxy-windows-amd64.zip) |
<!-- DOWNLOAD_TABLE_END -->
## 快速开始
+3 -1
View File
@@ -182,7 +182,9 @@ where
MsgType::StreamEnd | MsgType::StreamError => {
// Client-side cancellation or end
streams.remove(&frame.stream_id);
if let Some(tx) = streams.remove(&frame.stream_id) {
let _ = tx.send(frame).await;
}
}
MsgType::Ping => {
+158 -67
View File
@@ -3,22 +3,27 @@
//! Receives request frames, executes the upstream HTTP request,
//! and sends response frames back through the writer channel.
use std::io;
use std::sync::atomic::AtomicUsize;
use std::sync::atomic::Ordering;
use std::sync::Arc;
use std::time::{Duration, Instant};
use bytes::Bytes;
use futures_util::stream;
use futures_util::StreamExt;
use http_body_util::BodyExt;
use hyper::body::Frame as BodyFrame;
use tokio::sync::mpsc;
use tracing::{debug, warn};
use crate::state::{AppState, ServerContext};
use crate::target_filter;
use crate::upstream_client::{self, UpstreamRequestBody};
use crate::upstream_client;
use super::protocol::{
compress_payload, decompress_if_gzip, flags, Frame, MsgType, RequestMeta, ResponseMeta,
compress_payload, decompress_if_gzip, flags, Frame as TunnelFrame, MsgType, RequestMeta,
ResponseMeta,
};
use super::writer::FrameSender;
@@ -64,13 +69,13 @@ pub async fn handle_stream(
server: Arc<ServerContext>,
stream_id: u32,
meta: RequestMeta,
mut body_rx: mpsc::Receiver<Frame>,
body_rx: mpsc::Receiver<TunnelFrame>,
frame_tx: FrameSender,
) {
server.active_connections.fetch_add(1, Ordering::Release);
let connect_elapsed =
handle_stream_inner(&state, &server, stream_id, meta, &mut body_rx, &frame_tx).await;
handle_stream_inner(&state, &server, stream_id, meta, body_rx, &frame_tx).await;
server.active_connections.fetch_sub(1, Ordering::Release);
if let Some(d) = connect_elapsed {
@@ -79,7 +84,7 @@ pub async fn handle_stream(
}
/// Send a frame to the writer with a timeout. Returns false if send failed.
async fn send_frame(tx: &FrameSender, frame: Frame) -> bool {
async fn send_frame(tx: &FrameSender, frame: TunnelFrame) -> bool {
match tokio::time::timeout(FRAME_SEND_TIMEOUT, tx.send(frame)).await {
Ok(Ok(())) => true,
Ok(Err(_)) => {
@@ -102,62 +107,9 @@ async fn handle_stream_inner(
server: &ServerContext,
stream_id: u32,
meta: RequestMeta,
body_rx: &mut mpsc::Receiver<Frame>,
body_rx: mpsc::Receiver<TunnelFrame>,
frame_tx: &FrameSender,
) -> Option<Duration> {
// Collect request body
let mut body_parts: Vec<Bytes> = Vec::new();
let mut body_done = false;
// Drain body frames
while !body_done {
match body_rx.recv().await {
Some(frame) => {
if frame.msg_type == MsgType::RequestBody {
let payload = match decompress_if_gzip(&frame) {
Ok(d) => d,
Err(e) => {
send_error(
frame_tx,
stream_id,
&format!("gzip decompress failed: {e}"),
)
.await;
return None;
}
};
if !payload.is_empty() {
body_parts.push(payload);
}
if frame.is_end_stream() {
body_done = true;
}
} else if frame.msg_type == MsgType::StreamEnd
|| frame.msg_type == MsgType::StreamError
{
body_done = true;
if frame.msg_type == MsgType::StreamError {
return None; // Client cancelled
}
}
}
None => return None, // Channel closed
}
}
let body: Bytes = if body_parts.is_empty() {
Bytes::new()
} else if body_parts.len() == 1 {
body_parts.into_iter().next().unwrap()
} else {
let total: usize = body_parts.iter().map(|b| b.len()).sum();
let mut combined = Vec::with_capacity(total);
for part in &body_parts {
combined.extend_from_slice(part);
}
Bytes::from(combined)
};
// Validate target
let target_url = match url::Url::parse(&meta.url) {
Ok(u) => u,
@@ -207,12 +159,14 @@ async fn handle_stream_inner(
// Execute upstream request
let client = &state.upstream_client;
let timeout = Duration::from_secs(meta.timeout.clamp(MIN_TIMEOUT_SECS, MAX_TIMEOUT_SECS));
let request_body_size = Arc::new(AtomicUsize::new(0));
let request_body = build_streaming_request_body(body_rx, Arc::clone(&request_body_size));
let method: hyper::Method = meta.method.parse().unwrap_or(hyper::Method::GET);
let mut request = match hyper::Request::builder()
.method(method)
.uri(meta.url.as_str())
.body(UpstreamRequestBody::new(body.clone()))
.body(request_body)
{
Ok(request) => request,
Err(e) => {
@@ -240,7 +194,6 @@ async fn handle_stream_inner(
}
}
let body_size = body.len();
let mut captured_connection = upstream_client::capture_connection(&mut request);
let connection_start = Instant::now();
let connection_capture = tokio::spawn(async move {
@@ -313,7 +266,7 @@ async fn handle_stream_inner(
"upstream_processing_ms": request_timing.response_wait_ms,
"timing_source": "instrumented_connector",
"total_ms": connect_elapsed.as_millis() as u64,
"body_size": body_size,
"body_size": request_body_size.load(Ordering::Relaxed),
"mode": "tunnel",
});
resp_headers.push(("x-proxy-timing".to_string(), timing.to_string()));
@@ -325,7 +278,7 @@ async fn handle_stream_inner(
let (meta_payload, meta_flags) = compress_payload(meta_json);
if !send_frame(
frame_tx,
Frame::new(
TunnelFrame::new(
stream_id,
MsgType::ResponseHeaders,
meta_flags,
@@ -350,7 +303,7 @@ async fn handle_stream_inner(
let (payload, extra_flags) = compress_payload(chunk);
if !send_frame(
frame_tx,
Frame::new(stream_id, MsgType::ResponseBody, extra_flags, payload),
TunnelFrame::new(stream_id, MsgType::ResponseBody, extra_flags, payload),
)
.await
{
@@ -365,7 +318,12 @@ async fn handle_stream_inner(
let (payload, extra_flags) = compress_payload(slice);
if !send_frame(
frame_tx,
Frame::new(stream_id, MsgType::ResponseBody, extra_flags, payload),
TunnelFrame::new(
stream_id,
MsgType::ResponseBody,
extra_flags,
payload,
),
)
.await
{
@@ -387,7 +345,7 @@ async fn handle_stream_inner(
// Send STREAM_END
let _ = send_frame(
frame_tx,
Frame::new(
TunnelFrame::new(
stream_id,
MsgType::StreamEnd,
flags::END_STREAM,
@@ -404,7 +362,7 @@ async fn send_error(tx: &FrameSender, stream_id: u32, msg: &str) {
// Error frames use best-effort delivery — don't block if writer is congested
let _ = send_frame(
tx,
Frame::new(
TunnelFrame::new(
stream_id,
MsgType::StreamError,
0,
@@ -413,3 +371,136 @@ async fn send_error(tx: &FrameSender, stream_id: u32, msg: &str) {
)
.await;
}
fn build_streaming_request_body(
body_rx: mpsc::Receiver<TunnelFrame>,
body_size: Arc<AtomicUsize>,
) -> upstream_client::UpstreamRequestBody {
let body_stream = stream::unfold(
(body_rx, body_size, false),
|(mut body_rx, body_size, finished)| async move {
if finished {
return None;
}
loop {
let frame = match body_rx.recv().await {
Some(frame) => frame,
None => return None,
};
match frame.msg_type {
MsgType::RequestBody => {
let end_stream = frame.is_end_stream();
let payload = match decompress_if_gzip(&frame) {
Ok(payload) => payload,
Err(error) => {
let err =
io::Error::other(format!("gzip decompress failed: {error}"));
return Some((Err(err), (body_rx, body_size, true)));
}
};
if payload.is_empty() {
if end_stream {
return None;
}
continue;
}
body_size.fetch_add(payload.len(), Ordering::Relaxed);
return Some((
Ok(BodyFrame::data(payload)),
(body_rx, body_size, end_stream),
));
}
MsgType::StreamError => {
let message = String::from_utf8(frame.payload.to_vec())
.unwrap_or_else(|_| "client cancelled request body".to_string());
return Some((Err(io::Error::other(message)), (body_rx, body_size, true)));
}
MsgType::StreamEnd => return None,
_ => continue,
}
}
},
);
upstream_client::stream_request_body(body_stream)
}
#[cfg(test)]
mod tests {
use super::*;
#[tokio::test]
async fn streaming_request_body_yields_chunks_and_tracks_size() {
let (tx, rx) = mpsc::channel(4);
let body_size = Arc::new(AtomicUsize::new(0));
let mut body = build_streaming_request_body(rx, Arc::clone(&body_size));
tx.send(TunnelFrame::new(
1,
MsgType::RequestBody,
0,
Bytes::from_static(b"abc"),
))
.await
.expect("send first chunk");
tx.send(TunnelFrame::new(
1,
MsgType::RequestBody,
flags::END_STREAM,
Bytes::from_static(b"def"),
))
.await
.expect("send final chunk");
drop(tx);
let first = body
.frame()
.await
.expect("first frame")
.expect("first frame ok")
.into_data()
.expect("first data frame");
let second = body
.frame()
.await
.expect("second frame")
.expect("second frame ok")
.into_data()
.expect("second data frame");
assert_eq!(first, Bytes::from_static(b"abc"));
assert_eq!(second, Bytes::from_static(b"def"));
assert!(body.frame().await.is_none());
assert_eq!(body_size.load(Ordering::Relaxed), 6);
}
#[tokio::test]
async fn streaming_request_body_surfaces_client_cancel_as_error() {
let (tx, rx) = mpsc::channel(4);
let body_size = Arc::new(AtomicUsize::new(0));
let mut body = build_streaming_request_body(rx, Arc::clone(&body_size));
tx.send(TunnelFrame::new(
1,
MsgType::StreamError,
0,
Bytes::from_static(b"client cancelled"),
))
.await
.expect("send cancel frame");
drop(tx);
let err = body
.frame()
.await
.expect("error frame present")
.expect_err("body should surface cancellation error");
assert!(err.to_string().contains("client cancelled"));
assert!(body.frame().await.is_none());
assert_eq!(body_size.load(Ordering::Relaxed), 0);
}
}
+12 -2
View File
@@ -7,7 +7,10 @@ use std::task::{Context, Poll};
use std::time::Duration;
use bytes::Bytes;
use http_body_util::Full;
use futures_util::Stream;
use http_body_util::combinators::UnsyncBoxBody;
use http_body_util::{BodyExt, StreamBody};
use hyper::body::Frame;
use hyper::rt;
use hyper::Response;
use hyper::Uri;
@@ -30,9 +33,16 @@ type BoxError = Box<dyn std::error::Error + Send + Sync>;
type PlainStream = TokioIo<TcpStream>;
type TlsStream = TokioIo<tokio_rustls::client::TlsStream<TcpStream>>;
pub type UpstreamRequestBody = Full<Bytes>;
pub type UpstreamRequestBody = UnsyncBoxBody<Bytes, io::Error>;
pub type UpstreamClient = Client<InstrumentedConnector, UpstreamRequestBody>;
pub fn stream_request_body<S>(stream: S) -> UpstreamRequestBody
where
S: Stream<Item = Result<Frame<Bytes>, io::Error>> + Send + 'static,
{
StreamBody::new(stream).boxed_unsync()
}
#[derive(Clone, Copy, Debug, Default)]
pub struct ConnectTiming {
pub connect_ms: u64,
@@ -0,0 +1,64 @@
"""decouple request_candidates.key_id foreign key from provider_api_keys lifecycle
Revision ID: b7c8d9e0f1a2
Revises: c1d2e3f4a5b6
Create Date: 2026-03-12 19:15:00.000000+00:00
"""
from __future__ import annotations
import sqlalchemy as sa
from alembic import op
# revision identifiers, used by Alembic.
revision = "b7c8d9e0f1a2"
down_revision = "c1d2e3f4a5b6"
branch_labels = None
depends_on = None
def _fk_exists(constraint_name: str, table_name: str) -> bool:
bind = op.get_bind()
result = bind.execute(
sa.text(
"SELECT 1 FROM pg_constraint c "
"JOIN pg_class r ON c.conrelid = r.oid "
"JOIN pg_namespace n ON r.relnamespace = n.oid "
"WHERE c.conname = :name AND r.relname = :table "
" AND n.nspname = current_schema() AND c.contype = 'f'"
),
{"name": constraint_name, "table": table_name},
)
return result.scalar() is not None
def upgrade() -> None:
if _fk_exists("request_candidates_key_id_fkey", "request_candidates"):
op.drop_constraint(
"request_candidates_key_id_fkey", "request_candidates", type_="foreignkey"
)
def downgrade() -> None:
bind = op.get_bind()
bind.execute(
sa.text(
"UPDATE request_candidates rc "
"SET key_id = NULL "
"WHERE key_id IS NOT NULL "
" AND NOT EXISTS ("
" SELECT 1 FROM provider_api_keys pak WHERE pak.id = rc.key_id"
" )"
)
)
if not _fk_exists("request_candidates_key_id_fkey", "request_candidates"):
op.create_foreign_key(
"request_candidates_key_id_fkey",
"request_candidates",
"provider_api_keys",
["key_id"],
["id"],
ondelete="CASCADE",
)
@@ -0,0 +1,54 @@
"""add user rate_limit and backfill normal api key limits
Revision ID: b7e8f9a0c1d2
Revises: b7c8d9e0f1a2
Create Date: 2026-03-13 12:00:00.000000+00:00
"""
from __future__ import annotations
from collections.abc import Sequence
import sqlalchemy as sa
from sqlalchemy import inspect
from alembic import op
# revision identifiers, used by Alembic.
revision: str = "b7e8f9a0c1d2"
down_revision: str | None = "b7c8d9e0f1a2"
branch_labels: str | Sequence[str] | None = None
depends_on: str | Sequence[str] | None = None
def column_exists(table_name: str, column_name: str) -> bool:
bind = op.get_bind()
inspector = inspect(bind)
columns = [c["name"] for c in inspector.get_columns(table_name)]
return column_name in columns
def upgrade() -> None:
if not column_exists("users", "rate_limit"):
op.add_column("users", sa.Column("rate_limit", sa.Integer(), nullable=True))
# 普通 Key 新语义不再允许 NULL;存量 NULL 统一回填为 0(不限制)。
op.execute(sa.text("""
UPDATE api_keys
SET rate_limit = 0
WHERE is_standalone = FALSE
AND rate_limit IS NULL
"""))
def downgrade() -> None:
# 恢复普通 Key 的 rate_limit 为 NULL(与 upgrade 中回填 0 对应)
op.execute(sa.text("""
UPDATE api_keys
SET rate_limit = NULL
WHERE is_standalone = FALSE
AND rate_limit = 0
"""))
if column_exists("users", "rate_limit"):
op.drop_column("users", "rate_limit")
@@ -0,0 +1,87 @@
"""add user sessions table for device-level auth
Revision ID: f6e7d8c9b0a1
Revises: b7e8f9a0c1d2
Create Date: 2026-03-15 12:00:00.000000+00:00
"""
from __future__ import annotations
import sqlalchemy as sa
from alembic import op
# revision identifiers, used by Alembic.
revision = "f6e7d8c9b0a1"
down_revision = "b7e8f9a0c1d2"
branch_labels = None
depends_on = None
def upgrade() -> None:
bind = op.get_bind()
inspector = sa.inspect(bind)
if "user_sessions" in inspector.get_table_names():
return
op.create_table(
"user_sessions",
sa.Column("id", sa.String(length=36), nullable=False),
sa.Column("user_id", sa.String(length=36), nullable=False),
sa.Column("client_device_id", sa.String(length=128), nullable=False),
sa.Column("device_label", sa.String(length=120), nullable=True),
sa.Column("device_type", sa.String(length=20), nullable=False, server_default="unknown"),
sa.Column("browser_name", sa.String(length=50), nullable=True),
sa.Column("browser_version", sa.String(length=50), nullable=True),
sa.Column("os_name", sa.String(length=50), nullable=True),
sa.Column("os_version", sa.String(length=50), nullable=True),
sa.Column("device_model", sa.String(length=100), nullable=True),
sa.Column("ip_address", sa.String(length=45), nullable=True),
sa.Column("user_agent", sa.String(length=1000), nullable=True),
sa.Column("client_hints", sa.JSON(), nullable=True),
sa.Column("refresh_token_hash", sa.String(length=64), nullable=False),
sa.Column("prev_refresh_token_hash", sa.String(length=64), nullable=True),
sa.Column("rotated_at", sa.DateTime(timezone=True), nullable=True),
sa.Column(
"last_seen_at", sa.DateTime(timezone=True), nullable=False, server_default=sa.func.now()
),
sa.Column("expires_at", sa.DateTime(timezone=True), nullable=False),
sa.Column("revoked_at", sa.DateTime(timezone=True), nullable=True),
sa.Column("revoke_reason", sa.String(length=100), nullable=True),
sa.Column(
"created_at", sa.DateTime(timezone=True), nullable=False, server_default=sa.func.now()
),
sa.Column(
"updated_at", sa.DateTime(timezone=True), nullable=False, server_default=sa.func.now()
),
sa.ForeignKeyConstraint(["user_id"], ["users.id"], ondelete="CASCADE"),
sa.PrimaryKeyConstraint("id"),
)
op.create_index("ix_user_sessions_user_id", "user_sessions", ["user_id"], unique=False)
op.create_index(
"ix_user_sessions_client_device_id",
"user_sessions",
["client_device_id"],
unique=False,
)
op.create_index(
"idx_user_sessions_user_active",
"user_sessions",
["user_id", "revoked_at", "expires_at"],
unique=False,
)
op.create_index(
"idx_user_sessions_user_device",
"user_sessions",
["user_id", "client_device_id"],
unique=False,
)
def downgrade() -> None:
op.drop_index("idx_user_sessions_user_device", table_name="user_sessions")
op.drop_index("idx_user_sessions_user_active", table_name="user_sessions")
op.drop_index("ix_user_sessions_client_device_id", table_name="user_sessions")
op.drop_index("ix_user_sessions_user_id", table_name="user_sessions")
op.drop_table("user_sessions")
@@ -0,0 +1,41 @@
"""add status_snapshot column to provider_api_keys
Revision ID: c9d8e7f6a5b4
Revises: f6e7d8c9b0a1
Create Date: 2026-03-20 12:00:00.000000
"""
from __future__ import annotations
from collections.abc import Sequence
import sqlalchemy as sa
from sqlalchemy import inspect
from alembic import op
# revision identifiers, used by Alembic.
revision: str = "c9d8e7f6a5b4"
down_revision: str | None = "f6e7d8c9b0a1"
branch_labels: str | Sequence[str] | None = None
depends_on: str | Sequence[str] | None = None
def column_exists(table_name: str, column_name: str) -> bool:
bind = op.get_bind()
inspector = inspect(bind)
columns = [c["name"] for c in inspector.get_columns(table_name)]
return column_name in columns
def upgrade() -> None:
if not column_exists("provider_api_keys", "status_snapshot"):
op.add_column(
"provider_api_keys",
sa.Column("status_snapshot", sa.JSON(), nullable=True),
)
def downgrade() -> None:
if column_exists("provider_api_keys", "status_snapshot"):
op.drop_column("provider_api_keys", "status_snapshot")
@@ -0,0 +1,317 @@
"""usage token semantics v2
Revision ID: c3d4e5f6a7b8
Revises: c9d8e7f6a5b4
Create Date: 2026-03-24 14:00:00.000000+00:00
"""
from __future__ import annotations
from collections.abc import Sequence
import sqlalchemy as sa
from sqlalchemy import inspect
from alembic import op
# revision identifiers, used by Alembic.
revision: str = "c3d4e5f6a7b8"
down_revision: str | None = "c9d8e7f6a5b4"
branch_labels: str | Sequence[str] | None = None
depends_on: str | Sequence[str] | None = None
BACKFILL_BATCH_SIZE = 2000
# 最大批次数,防止因数据异常导致死循环(2000 * 500000 = 10亿行上限)
_MAX_BATCHES = 500000
# 使用子查询中间层展开 input_output_total_tokens 的计算,
# 确保 total_tokens 引用的是本次 SET 后的新值而非旧值。
_UPGRADE_BACKFILL_SQL = sa.text(
"""
UPDATE usage
SET
input_output_total_tokens = src.new_iot,
input_context_tokens = src.new_ict,
total_tokens = src.new_total,
cache_creation_cost_usd_5m = src.new_cc5m,
cache_creation_cost_usd_1h = src.new_cc1h,
actual_cache_creation_cost_usd_5m = src.new_acc5m,
actual_cache_creation_cost_usd_1h = src.new_acc1h,
actual_cache_cost_usd = src.new_accu,
cache_creation_price_per_1m_5m = src.new_cp5m,
cache_creation_price_per_1m_1h = src.new_cp1h,
cache_cost_usd = src.new_ccu
FROM (
SELECT
id,
COALESCE(input_output_total_tokens,
COALESCE(input_tokens, 0) + COALESCE(output_tokens, 0))
AS new_iot,
COALESCE(input_tokens, 0) + COALESCE(cache_read_input_tokens, 0)
AS new_ict,
/* total_tokens 引用本行计算出的 new_iot,避免依赖 SET 顺序 */
COALESCE(input_output_total_tokens,
COALESCE(input_tokens, 0) + COALESCE(output_tokens, 0))
+ COALESCE(cache_creation_input_tokens, 0)
+ COALESCE(cache_read_input_tokens, 0)
AS new_total,
CASE
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens_1h, 0) = 0
THEN COALESCE(cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_5m, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END AS new_cc5m,
CASE
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens_5m, 0) = 0
THEN COALESCE(cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_1h, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END AS new_cc1h,
CASE
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens_1h, 0) = 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_5m, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END AS new_acc5m,
CASE
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens_5m, 0) = 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_1h, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END AS new_acc1h,
COALESCE(actual_cache_creation_cost_usd, 0)
+ COALESCE(actual_cache_read_cost_usd, 0) AS new_accu,
CASE
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens_1h, 0) = 0
THEN cache_creation_price_per_1m
ELSE NULL
END AS new_cp5m,
CASE
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens_5m, 0) = 0
THEN cache_creation_price_per_1m
ELSE NULL
END AS new_cp1h,
COALESCE(cache_creation_cost_usd, 0)
+ COALESCE(cache_read_cost_usd, 0) AS new_ccu
FROM usage
WHERE id IN (
SELECT id FROM usage
WHERE
input_output_total_tokens IS DISTINCT FROM
COALESCE(input_output_total_tokens,
COALESCE(input_tokens, 0) + COALESCE(output_tokens, 0))
OR input_context_tokens IS DISTINCT FROM
COALESCE(input_tokens, 0) + COALESCE(cache_read_input_tokens, 0)
OR total_tokens IS DISTINCT FROM (
COALESCE(input_output_total_tokens,
COALESCE(input_tokens, 0) + COALESCE(output_tokens, 0))
+ COALESCE(cache_creation_input_tokens, 0)
+ COALESCE(cache_read_input_tokens, 0)
)
OR cache_creation_cost_usd_5m IS DISTINCT FROM (
CASE
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens_1h, 0) = 0
THEN COALESCE(cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_5m, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END
)
OR cache_creation_cost_usd_1h IS DISTINCT FROM (
CASE
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens_5m, 0) = 0
THEN COALESCE(cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_1h, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END
)
OR actual_cache_creation_cost_usd_5m IS DISTINCT FROM (
CASE
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens_1h, 0) = 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_5m, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END
)
OR actual_cache_creation_cost_usd_1h IS DISTINCT FROM (
CASE
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens_5m, 0) = 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens, 0) > 0
THEN COALESCE(actual_cache_creation_cost_usd, 0)
* (COALESCE(cache_creation_input_tokens_1h, 0) * 1.0
/ GREATEST(COALESCE(cache_creation_input_tokens, 0), 1))
ELSE 0
END
)
OR actual_cache_cost_usd IS DISTINCT FROM (
COALESCE(actual_cache_creation_cost_usd, 0)
+ COALESCE(actual_cache_read_cost_usd, 0)
)
OR cache_creation_price_per_1m_5m IS DISTINCT FROM (
CASE
WHEN COALESCE(cache_creation_input_tokens_5m, 0) > 0
AND COALESCE(cache_creation_input_tokens_1h, 0) = 0
THEN cache_creation_price_per_1m
ELSE NULL
END
)
OR cache_creation_price_per_1m_1h IS DISTINCT FROM (
CASE
WHEN COALESCE(cache_creation_input_tokens_1h, 0) > 0
AND COALESCE(cache_creation_input_tokens_5m, 0) = 0
THEN cache_creation_price_per_1m
ELSE NULL
END
)
OR cache_cost_usd IS DISTINCT FROM (
COALESCE(cache_creation_cost_usd, 0) + COALESCE(cache_read_cost_usd, 0)
)
ORDER BY id
LIMIT :batch_size
)
) AS src
WHERE usage.id = src.id
"""
)
_DOWNGRADE_BACKFILL_SQL = sa.text(
"""
UPDATE usage
SET total_tokens = COALESCE(input_output_total_tokens, COALESCE(input_tokens, 0) + COALESCE(output_tokens, 0))
WHERE id IN (
SELECT id
FROM usage
WHERE total_tokens IS DISTINCT FROM
COALESCE(input_output_total_tokens, COALESCE(input_tokens, 0) + COALESCE(output_tokens, 0))
ORDER BY id
LIMIT :batch_size
)
"""
)
def column_exists(table_name: str, column_name: str) -> bool:
bind = op.get_bind()
inspector = inspect(bind)
columns = [c["name"] for c in inspector.get_columns(table_name)]
return column_name in columns
def run_backfill_in_batches(sql: sa.TextClause, batch_size: int = BACKFILL_BATCH_SIZE) -> None:
context = op.get_context()
for _ in range(_MAX_BATCHES):
# Commit the preceding schema transaction before each batch so PostgreSQL
# does not keep ALTER TABLE locks for the entire data backfill.
with context.autocommit_block():
rowcount = op.get_bind().execute(sql, {"batch_size": batch_size}).rowcount
if rowcount == 0:
break
else:
raise RuntimeError(
f"Backfill did not converge after {_MAX_BATCHES} batches "
f"(batch_size={batch_size}). Possible infinite loop due to data anomaly."
)
def upgrade() -> None:
if column_exists("usage", "total_tokens") and not column_exists("usage", "input_output_total_tokens"):
with op.batch_alter_table("usage") as batch_op:
batch_op.alter_column(
"total_tokens",
new_column_name="input_output_total_tokens",
existing_type=sa.Integer(),
existing_nullable=True,
)
with op.batch_alter_table("usage") as batch_op:
if not column_exists("usage", "input_context_tokens"):
batch_op.add_column(sa.Column("input_context_tokens", sa.Integer(), nullable=False, server_default="0"))
if not column_exists("usage", "total_tokens"):
batch_op.add_column(sa.Column("total_tokens", sa.Integer(), nullable=False, server_default="0"))
if not column_exists("usage", "cache_creation_cost_usd_5m"):
batch_op.add_column(sa.Column("cache_creation_cost_usd_5m", sa.Numeric(20, 8), nullable=False, server_default="0"))
if not column_exists("usage", "cache_creation_cost_usd_1h"):
batch_op.add_column(sa.Column("cache_creation_cost_usd_1h", sa.Numeric(20, 8), nullable=False, server_default="0"))
if not column_exists("usage", "actual_cache_creation_cost_usd_5m"):
batch_op.add_column(sa.Column("actual_cache_creation_cost_usd_5m", sa.Numeric(20, 8), nullable=False, server_default="0"))
if not column_exists("usage", "actual_cache_creation_cost_usd_1h"):
batch_op.add_column(sa.Column("actual_cache_creation_cost_usd_1h", sa.Numeric(20, 8), nullable=False, server_default="0"))
if not column_exists("usage", "actual_cache_cost_usd"):
batch_op.add_column(sa.Column("actual_cache_cost_usd", sa.Numeric(20, 8), nullable=False, server_default="0"))
if not column_exists("usage", "cache_creation_price_per_1m_5m"):
batch_op.add_column(sa.Column("cache_creation_price_per_1m_5m", sa.Numeric(20, 8), nullable=True))
if not column_exists("usage", "cache_creation_price_per_1m_1h"):
batch_op.add_column(sa.Column("cache_creation_price_per_1m_1h", sa.Numeric(20, 8), nullable=True))
run_backfill_in_batches(_UPGRADE_BACKFILL_SQL)
def downgrade() -> None:
run_backfill_in_batches(_DOWNGRADE_BACKFILL_SQL)
with op.batch_alter_table("usage") as batch_op:
if column_exists("usage", "cache_creation_price_per_1m_1h"):
batch_op.drop_column("cache_creation_price_per_1m_1h")
if column_exists("usage", "cache_creation_price_per_1m_5m"):
batch_op.drop_column("cache_creation_price_per_1m_5m")
if column_exists("usage", "actual_cache_cost_usd"):
batch_op.drop_column("actual_cache_cost_usd")
if column_exists("usage", "actual_cache_creation_cost_usd_1h"):
batch_op.drop_column("actual_cache_creation_cost_usd_1h")
if column_exists("usage", "actual_cache_creation_cost_usd_5m"):
batch_op.drop_column("actual_cache_creation_cost_usd_5m")
if column_exists("usage", "cache_creation_cost_usd_1h"):
batch_op.drop_column("cache_creation_cost_usd_1h")
if column_exists("usage", "cache_creation_cost_usd_5m"):
batch_op.drop_column("cache_creation_cost_usd_5m")
if column_exists("usage", "input_context_tokens"):
batch_op.drop_column("input_context_tokens")
if column_exists("usage", "total_tokens"):
batch_op.drop_column("total_tokens")
if column_exists("usage", "input_output_total_tokens"):
batch_op.alter_column(
"input_output_total_tokens",
new_column_name="total_tokens",
existing_type=sa.Integer(),
existing_nullable=True,
)
+35 -24
View File
@@ -5,6 +5,7 @@
# 部署/更新: ./deploy.sh (自动检测所有变化)
# 指定 Hub 版本: ./deploy.sh --hub-tag hub-v0.1.0
# 更新 Hub: ./deploy.sh --update-hub
# GitHub 镜像: ./deploy.sh --mirror https://ghfast.top
# 强制重建: ./deploy.sh --rebuild-base
# 强制全部重建: ./deploy.sh --force
@@ -44,6 +45,7 @@ Usage: ./deploy.sh [options]
Options:
--hub-tag <hub-vX.Y.Z> 指定 Hub Release tag(例如 hub-v0.1.0)
--update-hub 强制刷新 Hub 版本标记(下次构建会重新下载)
--mirror <url> GitHub 下载镜像(例如 https://ghfast.top)
--rebuild-base, -r 仅重建 base 镜像
--force, -f 强制重建全部(hub/base/app)并重启
-h, --help 显示帮助
@@ -54,6 +56,7 @@ FORCE_REBUILD_ALL=false
REBUILD_BASE_ONLY=false
FORCE_UPDATE_HUB=false
HUB_TAG="${HUB_TAG:-}"
GITHUB_MIRROR="${GITHUB_MIRROR:-}"
RESOLVED_HUB_TAG=""
while [ $# -gt 0 ]; do
@@ -70,6 +73,14 @@ while [ $# -gt 0 ]; do
FORCE_UPDATE_HUB=true
shift
;;
--mirror)
if [ $# -lt 2 ]; then
echo "ERROR: --mirror needs a URL, e.g. https://ghfast.top"
exit 1
fi
GITHUB_MIRROR="$2"
shift 2
;;
--rebuild-base|-r)
REBUILD_BASE_ONLY=true
shift
@@ -97,27 +108,26 @@ if [ -n "$HUB_TAG" ]; then
esac
fi
# 提取 pyproject.toml 中"会影响运行时依赖安装"的最小指纹(与 CI 保持一致):
# - [build-system] requires / build-backend
# - [project] requires-python / dependencies
# 使用 Python tomllib 解析,不受 TOML 格式变化影响。
# 提取 pyproject.toml 中会影响运行时依赖安装的字段指纹(纯 shell,无需 Python)
# 用 sed 提取 dependencies / requires 数组块和单值字段,排序后输出稳定文本
pyproject_deps_fingerprint() {
python3 - <<'PY'
import json, pathlib, tomllib
data = tomllib.loads(pathlib.Path("pyproject.toml").read_text("utf-8"))
project = data.get("project") or {}
build = data.get("build-system") or {}
fingerprint = {
"requires-python": project.get("requires-python"),
"dependencies": sorted(project.get("dependencies") or []),
"build-backend": build.get("build-backend"),
"build-requires": sorted(build.get("requires") or []),
}
print(json.dumps(fingerprint, sort_keys=True, separators=(",", ":")))
PY
local file="pyproject.toml"
# 提取 "key = [..." 多行数组块(从 key 行到 ] 行)
extract_array() {
sed -n "/^$1[[:space:]]*=[[:space:]]*\[/,/\]/p" "$file" | grep '"' | sed 's/.*"\(.*\)".*/\1/' | sort
}
# 提取 "key = "value"" 单行值
extract_value() {
grep -m1 "^$1[[:space:]]*=" "$file" 2>/dev/null | sed 's/.*"\(.*\)".*/\1/'
}
{
echo "requires-python=$(extract_value requires-python)"
echo "build-backend=$(extract_value build-backend)"
echo "dependencies:"
extract_array dependencies
echo "build-requires:"
extract_array requires
}
}
# 计算依赖文件的哈希值(包含 Dockerfile.base.local)
@@ -286,10 +296,15 @@ build_app() {
if [ -n "${GITHUB_TOKEN:-}" ]; then
token_args=(--build-arg "GITHUB_TOKEN=${GITHUB_TOKEN}")
fi
local mirror_args=()
if [ -n "${GITHUB_MIRROR:-}" ]; then
mirror_args=(--build-arg "GITHUB_MIRROR=${GITHUB_MIRROR}")
fi
docker build --pull=false \
--build-arg HUB_RELEASE_REPO="$GITHUB_REPO" \
--build-arg HUB_TAG="$RESOLVED_HUB_TAG" \
"${token_args[@]}" \
"${mirror_args[@]}" \
-f Dockerfile.app.local \
-t aether-app:latest .
save_code_hash
@@ -372,10 +387,6 @@ if [ "$FORCE_UPDATE_HUB" = true ]; then
exit 0
fi
# 拉取最新代码
echo ">>> Pulling latest code..."
git pull
# 标记是否需要重启
NEED_RESTART=false
BASE_REBUILT=false
+66 -2
View File
@@ -5,12 +5,14 @@
</template>
<script setup lang="ts">
import { onMounted, onErrorCaptured } from 'vue'
import { onMounted, onErrorCaptured, onUnmounted } from 'vue'
import { useAuthStore } from '@/stores/auth'
import ToastContainer from '@/components/ToastContainer.vue'
import ConfirmContainer from '@/components/ConfirmContainer.vue'
import apiClient from '@/api/client'
import apiClient, { AUTH_STATE_CHANGE_EVENT } from '@/api/client'
import { NETWORK_CONFIG, AUTH_CONFIG } from '@/config/constants'
import router from '@/router'
import { hasAuthIdentityChanged } from '@/utils/authToken'
import { log } from '@/utils/logger'
const authStore = useAuthStore()
@@ -86,7 +88,59 @@ if (typeof window !== 'undefined') {
})
}
async function syncExternalAuthState(nextToken: string | null): Promise<void> {
const previousToken = authStore.token
const previousUser = authStore.user
? {
id: authStore.user.id,
role: authStore.user.role,
}
: null
authStore.syncToken()
if (!nextToken) {
if (previousToken || previousUser) {
authStore.applyExternalLogout()
await router.replace('/')
}
return
}
const identityChanged = hasAuthIdentityChanged(previousToken, nextToken, previousUser)
if (!identityChanged && previousUser) {
return
}
const user = await authStore.fetchCurrentUser()
if (!user) {
return
}
if (router.currentRoute.value.path.startsWith('/admin') && user.role !== 'admin') {
await router.replace('/dashboard')
}
}
function handleAuthStorageChange(event: StorageEvent): void {
if (event.key !== 'access_token') {
return
}
syncExternalAuthState(event.newValue).catch((err) => log.error('syncExternalAuthState failed', err))
}
function handleLocalAuthStateChange(event: Event): void {
const authEvent = event as CustomEvent<{ token: string | null }>
syncExternalAuthState(authEvent.detail?.token ?? apiClient.getToken()).catch((err) => log.error('syncExternalAuthState failed', err))
}
onMounted(async () => {
if (typeof window !== 'undefined') {
window.addEventListener('storage', handleAuthStorageChange)
window.addEventListener(AUTH_STATE_CHANGE_EVENT, handleLocalAuthStateChange as (event: Event) => void)
}
// 延迟检查认证状态,让页面先加载
setTimeout(async () => {
try {
@@ -97,4 +151,14 @@ onMounted(async () => {
}
}, AUTH_CONFIG.TOKEN_REFRESH_INTERVAL)
})
onUnmounted(() => {
if (typeof window !== 'undefined') {
window.removeEventListener('storage', handleAuthStorageChange)
window.removeEventListener(
AUTH_STATE_CHANGE_EVENT,
handleLocalAuthStateChange as (event: Event) => void,
)
}
})
</script>
+31
View File
@@ -0,0 +1,31 @@
import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest'
import apiClient, { AUTH_STATE_CHANGE_EVENT } from '@/api/client'
describe('apiClient auth state change event', () => {
beforeEach(() => {
localStorage.clear()
apiClient.clearAuth()
})
afterEach(() => {
localStorage.clear()
apiClient.clearAuth()
})
it('dispatches a same-tab auth change event when clearing auth', () => {
const handler = vi.fn()
window.addEventListener(AUTH_STATE_CHANGE_EVENT, handler as EventListener)
apiClient.setToken('access-token')
apiClient.clearAuth()
expect(localStorage.getItem('access_token')).toBeNull()
expect(handler).toHaveBeenCalledTimes(1)
const event = handler.mock.calls[0][0] as CustomEvent<{ token: string | null }>
expect(event.detail).toEqual({ token: null })
window.removeEventListener(AUTH_STATE_CHANGE_EVENT, handler as EventListener)
})
})
+4 -312
View File
@@ -86,6 +86,7 @@ export interface UserExport {
allowed_providers?: string[] | null
allowed_api_formats?: string[] | null
allowed_models?: string[] | null
rate_limit?: number | null // null = 跟随系统默认,0 = 不限制
model_capability_settings?: Record<string, Record<string, boolean>>
unlimited?: boolean
wallet?: BillingSummary | null
@@ -102,7 +103,7 @@ export interface UserApiKeyExport {
allowed_providers?: string[] | null
allowed_api_formats?: string[] | null
allowed_models?: string[] | null
rate_limit?: number | null // null = 无限制
rate_limit?: number | null // legacy/null 兼容;1.3+ standalone null = 跟随系统默认
concurrent_limit?: number | null
force_capabilities?: Record<string, boolean>
is_active: boolean
@@ -349,7 +350,7 @@ export interface AdminApiKey {
total_requests?: number
total_tokens?: number
total_cost_usd?: number
rate_limit?: number | null // null = 无限制
rate_limit?: number | null // null = 跟随系统默认,0 = 不限制
allowed_providers?: string[] | null // 允许的提供商列表
allowed_api_formats?: string[] | null // 允许的 API 格式列表
allowed_models?: string[] | null // 允许的模型列表
@@ -365,7 +366,7 @@ export interface CreateStandaloneApiKeyRequest {
allowed_providers?: string[] | null
allowed_api_formats?: string[] | null
allowed_models?: string[] | null
rate_limit?: number | null // null = 无限制
rate_limit?: number | null // null = 跟随系统默认,0 = 不限制
expires_at?: string | null // ISO 日期字符串,如 "2025-12-31",null = 永不过期
initial_balance_usd: number | null // 初始余额,null = 无限制
unlimited_balance?: boolean | null // 编辑时仅切换额度模式,不调整余额数值
@@ -379,102 +380,6 @@ export interface AdminApiKeysResponse {
skip: number
}
export interface LeaderboardItem {
rank: number
id: string
name: string
value: number
requests: number
tokens: number
cost: number
}
export interface LeaderboardResponse {
items: LeaderboardItem[]
total: number
metric: string
start_date?: string | null
end_date?: string | null
}
export interface CostForecastResponse {
history: Array<{ date: string; total_cost: number }>
forecast: Array<{ date: string; total_cost: number }>
slope: number
intercept: number
start_date: string
end_date: string
}
export interface CostSavingsResponse {
cache_read_tokens: number
cache_read_cost: number
cache_creation_cost: number
estimated_full_cost: number
cache_savings: number
}
export interface QuotaUsageProvider {
id: string
name: string
quota_usd: number
used_usd: number
remaining_usd: number
usage_percent: number
quota_expires_at?: string | null
estimated_exhaust_at?: string | null
}
export interface QuotaUsageResponse {
providers: QuotaUsageProvider[]
}
export interface PercentileItem {
date: string
p50_response_time_ms?: number | null
p90_response_time_ms?: number | null
p99_response_time_ms?: number | null
p50_first_byte_time_ms?: number | null
p90_first_byte_time_ms?: number | null
p99_first_byte_time_ms?: number | null
}
export interface ErrorDistributionItem {
category: string
count: number
}
export interface ErrorTrendItem {
date: string
total: number
categories: Record<string, number>
}
export interface ErrorDistributionResponse {
distribution: ErrorDistributionItem[]
trend: ErrorTrendItem[]
}
export interface ComparisonMetric {
total_requests: number
total_tokens: number
total_cost: number
actual_total_cost: number
avg_response_time_ms: number
error_requests: number
}
export interface ComparisonResponse {
current: ComparisonMetric
comparison: ComparisonMetric
change_percent: Record<string, number | null>
current_start: string
current_end: string
comparison_start: string
comparison_end: string
}
export interface ApiKeyToggleResponse {
id: string // UUID
is_active: boolean
@@ -752,197 +657,6 @@ export const adminApi = {
return response.data
},
// Stats / Leaderboards
async getLeaderboardUsers(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
metric?: 'requests' | 'tokens' | 'cost'
order?: 'asc' | 'desc'
limit?: number
offset?: number
provider_name?: string
model?: string
include_inactive?: boolean
exclude_admin?: boolean
}): Promise<LeaderboardResponse> {
const cacheKey = buildCacheKey('admin:stats:leaderboard:users', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<LeaderboardResponse>('/api/admin/stats/leaderboard/users', {
params
})
return response.data
},
20 * 1000
)
},
async getLeaderboardApiKeys(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
metric?: 'requests' | 'tokens' | 'cost'
order?: 'asc' | 'desc'
limit?: number
offset?: number
provider_name?: string
model?: string
include_inactive?: boolean
exclude_admin?: boolean
}): Promise<LeaderboardResponse> {
const cacheKey = buildCacheKey('admin:stats:leaderboard:api-keys', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<LeaderboardResponse>('/api/admin/stats/leaderboard/api-keys', {
params
})
return response.data
},
20 * 1000
)
},
async getLeaderboardModels(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
metric?: 'requests' | 'tokens' | 'cost'
order?: 'asc' | 'desc'
limit?: number
offset?: number
provider_name?: string
model?: string
}): Promise<LeaderboardResponse> {
const cacheKey = buildCacheKey('admin:stats:leaderboard:models', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<LeaderboardResponse>('/api/admin/stats/leaderboard/models', {
params
})
return response.data
},
20 * 1000
)
},
async getCostForecast(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
days?: number
forecast_days?: number
}): Promise<CostForecastResponse> {
const cacheKey = buildCacheKey('admin:stats:cost:forecast', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<CostForecastResponse>('/api/admin/stats/cost/forecast', {
params
})
return response.data
},
30 * 1000
)
},
async getCostSavings(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
provider_name?: string
model?: string
}): Promise<CostSavingsResponse> {
const cacheKey = buildCacheKey('admin:stats:cost:savings', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<CostSavingsResponse>('/api/admin/stats/cost/savings', {
params
})
return response.data
},
30 * 1000
)
},
async getQuotaUsage(): Promise<QuotaUsageResponse> {
return cachedRequest(
'admin:stats:providers:quota-usage',
async () => {
const response = await apiClient.get<QuotaUsageResponse>('/api/admin/stats/providers/quota-usage')
return response.data
},
30 * 1000
)
},
async getPercentiles(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
}): Promise<PercentileItem[]> {
const cacheKey = buildCacheKey('admin:stats:performance:percentiles', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<PercentileItem[]>('/api/admin/stats/performance/percentiles', {
params
})
return response.data
},
20 * 1000
)
},
async getErrorDistribution(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
}): Promise<ErrorDistributionResponse> {
const cacheKey = buildCacheKey('admin:stats:errors:distribution', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<ErrorDistributionResponse>('/api/admin/stats/errors/distribution', {
params
})
return response.data
},
20 * 1000
)
},
async getComparison(params: {
current_start: string
current_end: string
comparison_type?: 'period' | 'year'
timezone?: string
tz_offset_minutes?: number
}): Promise<ComparisonResponse> {
const response = await apiClient.get<ComparisonResponse>('/api/admin/stats/comparison', {
params
})
return response.data
},
// 数据清空
purgeConfig: () => purge<{ message: string; deleted: Record<string, number> }>('config'),
purgeUsers: () => purge<{ message: string; deleted: Record<string, number> }>('users'),
@@ -951,26 +665,4 @@ export const adminApi = {
purgeRequestBodies: () => purge<{ message: string; cleaned: Record<string, number> }>('request-bodies'),
purgeStats: () => purge<{ message: string }>('stats'),
async getTimeSeries(params?: {
start_date?: string
end_date?: string
preset?: string
granularity?: 'hour' | 'day' | 'week' | 'month'
timezone?: string
tz_offset_minutes?: number
user_id?: string
model?: string
provider_name?: string
}): Promise<Array<Record<string, unknown>>> {
const cacheKey = buildCacheKey('admin:stats:time-series', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<Array<Record<string, unknown>>>('/api/admin/stats/time-series', { params })
return response.data
},
20 * 1000
)
},
}
+396
View File
@@ -0,0 +1,396 @@
import apiClient from './client'
import { cachedRequest, buildCacheKey } from '@/utils/cache'
import type { ActivityHeatmap } from '@/types/activity'
export type AnalyticsScopeKind = 'global' | 'me' | 'user' | 'api_key'
export type AnalyticsGranularity = 'hour' | 'day' | 'week' | 'month'
export type AnalyticsBreakdownDimension = 'model' | 'provider' | 'api_format' | 'api_key' | 'user'
export type AnalyticsBreakdownMetric = 'requests_total' | 'total_tokens' | 'total_cost_usd' | 'actual_total_cost_usd'
export type AnalyticsLeaderboardEntity = 'user' | 'api_key'
export type AnalyticsLeaderboardMetric = 'requests_total' | 'total_tokens' | 'total_cost_usd'
export interface AnalyticsScope {
kind: AnalyticsScopeKind
user_id?: string | null
api_key_id?: string | null
}
export interface AnalyticsTimeRange {
start_date?: string
end_date?: string
preset?: string
granularity?: AnalyticsGranularity
timezone?: string | null
tz_offset_minutes?: number
}
export interface AnalyticsFilters {
user_ids?: string[]
provider_names?: string[]
models?: string[]
target_models?: string[]
api_key_ids?: string[]
api_formats?: string[]
request_types?: string[]
statuses?: string[]
error_categories?: string[]
is_stream?: boolean | null
has_format_conversion?: boolean | null
}
export interface AnalyticsBaseRequest {
scope: AnalyticsScope
time_range: AnalyticsTimeRange
filters?: AnalyticsFilters
}
export interface AnalyticsSummary {
requests_total: number
requests_success: number
requests_error: number
requests_stream: number
success_rate: number
input_tokens: number
output_tokens: number
input_output_total_tokens: number
cache_creation_input_tokens: number
cache_creation_input_tokens_5m: number
cache_creation_input_tokens_1h: number
cache_read_input_tokens: number
input_context_tokens: number
total_tokens: number
cache_hit_rate: number
input_cost_usd: number
output_cost_usd: number
cache_creation_cost_usd: number
cache_creation_cost_usd_5m: number
cache_creation_cost_usd_1h: number
cache_read_cost_usd: number
cache_cost_usd: number
request_cost_usd: number
total_cost_usd: number
actual_total_cost_usd: number
actual_cache_cost_usd: number
avg_response_time_ms: number
avg_first_byte_time_ms: number
format_conversion_count: number
models_used_count: number
}
export interface AnalyticsCompositionSegment {
key: string
value: number
percentage: number
}
export interface AnalyticsOverviewResponse {
query_context: {
scope: AnalyticsScope
time_range: AnalyticsTimeRange
}
summary: AnalyticsSummary
composition: {
token_segments: AnalyticsCompositionSegment[]
cost_segments: AnalyticsCompositionSegment[]
}
}
export interface AnalyticsTimeseriesBucket extends AnalyticsSummary {
bucket_start: string
bucket_end: string
}
export interface AnalyticsTimeseriesResponse {
buckets: AnalyticsTimeseriesBucket[]
}
export interface AnalyticsBreakdownRequest extends AnalyticsBaseRequest {
dimension: AnalyticsBreakdownDimension
metric?: AnalyticsBreakdownMetric
limit?: number
}
export interface AnalyticsBreakdownRow extends AnalyticsSummary {
key: string
label: string
share_of_total_cost: number
share_of_total_tokens: number
}
export interface AnalyticsBreakdownResponse {
dimension: AnalyticsBreakdownDimension
metric: AnalyticsBreakdownMetric
rows: AnalyticsBreakdownRow[]
}
export interface AnalyticsRecordsRequest extends AnalyticsBaseRequest {
search?: {
text?: string | null
request_id?: string | null
}
pagination?: {
limit?: number
offset?: number
}
}
export interface AnalyticsRecord {
id: string
request_id: string
created_at: string | null
user_id: string | null
username: string | null
api_key_id: string | null
api_key_name: string | null
provider_api_key_name: string | null
provider_name: string | null
model: string
target_model: string | null
api_format: string | null
request_type: string | null
status: string
billing_status: string
is_stream: boolean
has_format_conversion: boolean | null
has_fallback?: boolean
has_retry?: boolean
status_code: number | null
error_message: string | null
error_category: string | null
response_time_ms: number | null
first_byte_time_ms: number | null
input_tokens: number
output_tokens: number
input_output_total_tokens: number
cache_creation_input_tokens: number
cache_creation_input_tokens_5m: number
cache_creation_input_tokens_1h: number
cache_read_input_tokens: number
input_context_tokens: number
total_tokens: number
input_cost_usd: number
output_cost_usd: number
cache_creation_cost_usd: number
cache_creation_cost_usd_5m: number
cache_creation_cost_usd_1h: number
cache_read_cost_usd: number
cache_cost_usd: number
request_cost_usd: number
total_cost_usd: number
actual_total_cost_usd: number
actual_cache_cost_usd: number
rate_multiplier: number
}
export interface AnalyticsRecordsResponse {
total: number
limit: number
offset: number
records: AnalyticsRecord[]
}
export interface AnalyticsFilterOption {
value: string
label: string
}
export interface AnalyticsFilterOptionsResponse {
providers: AnalyticsFilterOption[]
models: AnalyticsFilterOption[]
target_models: AnalyticsFilterOption[]
api_formats: AnalyticsFilterOption[]
request_types: AnalyticsFilterOption[]
error_categories: AnalyticsFilterOption[]
statuses: AnalyticsFilterOption[]
users?: AnalyticsFilterOption[]
api_keys?: AnalyticsFilterOption[]
}
export interface AnalyticsLeaderboardRequest extends AnalyticsBaseRequest {
entity: AnalyticsLeaderboardEntity
metric: AnalyticsLeaderboardMetric
limit?: number
}
export interface AnalyticsLeaderboardItem {
rank: number
id: string
label: string
requests_total: number
total_tokens: number
total_cost_usd: number
actual_total_cost_usd: number
metric_value: number
}
export interface AnalyticsLeaderboardResponse {
entity: AnalyticsLeaderboardEntity
metric: AnalyticsLeaderboardMetric
items: AnalyticsLeaderboardItem[]
}
export interface AnalyticsPercentilePoint {
date: string
p50_response_time_ms: number | null
p90_response_time_ms: number | null
p99_response_time_ms: number | null
p50_first_byte_time_ms: number | null
p90_first_byte_time_ms: number | null
p99_first_byte_time_ms: number | null
}
export interface AnalyticsErrorCategory {
category: string
label: string
count: number
}
export interface AnalyticsErrorTrendItem {
date: string
total: number
}
export interface AnalyticsProviderHealthItem {
provider_name: string
requests_total: number
success_rate: number
error_rate: number
avg_response_time_ms: number
avg_first_byte_time_ms: number
}
export interface AnalyticsPerformanceResponse {
latency: {
response_time_ms: {
avg: number
p50: number | null
p90: number | null
p99: number | null
}
first_byte_time_ms: {
avg: number
p50: number | null
p90: number | null
p99: number | null
}
}
percentiles: AnalyticsPercentilePoint[]
errors: {
total: number
rate: number
categories: AnalyticsErrorCategory[]
trend: AnalyticsErrorTrendItem[]
}
provider_health: AnalyticsProviderHealthItem[]
}
export interface AnalyticsActiveRequest {
id: string
status: 'pending' | 'streaming' | 'completed' | 'failed' | 'cancelled'
input_tokens: number
output_tokens: number
cache_creation_input_tokens?: number | null
cache_read_input_tokens?: number | null
cost?: number
total_cost_usd?: number
actual_cost?: number | null
actual_total_cost_usd?: number | null
rate_multiplier?: number | null
response_time_ms: number | null
first_byte_time_ms: number | null
provider?: string | null
provider_name?: string | null
api_key_name?: string | null
api_format?: string | null
endpoint_api_format?: string | null
has_format_conversion?: boolean | null
target_model?: string | null
}
export interface AnalyticsActiveRequestsResponse {
requests: AnalyticsActiveRequest[]
}
export const analyticsApi = {
async getOverview(payload: AnalyticsBaseRequest): Promise<AnalyticsOverviewResponse> {
const response = await apiClient.post<AnalyticsOverviewResponse>('/api/analytics/overview', payload)
return response.data
},
async getTimeseries(payload: AnalyticsBaseRequest): Promise<AnalyticsTimeseriesResponse> {
const response = await apiClient.post<AnalyticsTimeseriesResponse>('/api/analytics/timeseries', payload)
return response.data
},
async getBreakdown(payload: AnalyticsBreakdownRequest): Promise<AnalyticsBreakdownResponse> {
const response = await apiClient.post<AnalyticsBreakdownResponse>('/api/analytics/breakdown', payload)
return response.data
},
async getRecords(payload: AnalyticsRecordsRequest): Promise<AnalyticsRecordsResponse> {
const response = await apiClient.post<AnalyticsRecordsResponse>('/api/analytics/records', payload)
return response.data
},
async getFilterOptions(payload: AnalyticsBaseRequest): Promise<AnalyticsFilterOptionsResponse> {
const response = await apiClient.post<AnalyticsFilterOptionsResponse>('/api/analytics/filter-options', payload)
return response.data
},
async getLeaderboard(payload: AnalyticsLeaderboardRequest): Promise<AnalyticsLeaderboardResponse> {
const response = await apiClient.post<AnalyticsLeaderboardResponse>('/api/analytics/leaderboard', payload)
return response.data
},
async getPerformance(payload: AnalyticsBaseRequest): Promise<AnalyticsPerformanceResponse> {
const response = await apiClient.post<AnalyticsPerformanceResponse>('/api/analytics/performance', payload)
return response.data
},
async getHeatmap(payload: {
scope: AnalyticsScope
user_id?: string | null
api_key_id?: string | null
}): Promise<ActivityHeatmap> {
const cacheKey = buildCacheKey('analytics:heatmap', payload as Record<string, unknown>)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.post<ActivityHeatmap>('/api/analytics/heatmap', payload)
return response.data
},
60000,
)
},
async getActiveRequests(payload: {
scope: AnalyticsScope
ids?: string[]
}): Promise<AnalyticsActiveRequestsResponse> {
const response = await apiClient.post<AnalyticsActiveRequestsResponse>('/api/analytics/active-requests', {
...payload,
ids: payload.ids ?? [],
})
return response.data
},
async analyzeCacheAffinityTTL(payload: {
scope: AnalyticsScope
user_id?: string | null
api_key_id?: string | null
hours?: number
}): Promise<Record<string, unknown>> {
const response = await apiClient.post('/api/analytics/cache-affinity/ttl-analysis', payload)
return response.data
},
async analyzeCacheAffinityHit(payload: {
scope: AnalyticsScope
user_id?: string | null
api_key_id?: string | null
hours?: number
}): Promise<Record<string, unknown>> {
const response = await apiClient.post('/api/analytics/cache-affinity/hit-analysis', payload)
return response.data
},
}
+202
View File
@@ -33,6 +33,203 @@ export interface AuditFilters {
offset?: number
}
export type MonitoringMetricStatus = 'ok' | 'warning' | 'danger' | 'degraded' | 'error' | 'unknown'
export interface MonitoringCpuMetric {
status: MonitoringMetricStatus
label: string
usage_percent: number | null
load_percent: number | null
core_count: number
message?: string | null
}
export interface MonitoringMemoryMetric {
status: MonitoringMetricStatus
label: string
used_percent: number | null
used_bytes: number | null
available_bytes: number | null
total_bytes: number | null
message?: string | null
}
export interface MonitoringServiceMetric {
status: MonitoringMetricStatus
label: string
latency_ms: number | null
memory_status?: MonitoringMetricStatus
memory_label?: string | null
used_memory_bytes?: number | null
peak_memory_bytes?: number | null
maxmemory_bytes?: number | null
memory_ceiling_bytes?: number | null
memory_source?: 'configured' | 'maxmemory' | 'system' | 'unknown'
available_memory_bytes?: number | null
memory_percent?: number | null
message?: string | null
}
export interface MonitoringPostgresMetric {
status: MonitoringMetricStatus
label: string
usage_percent: number | null
pool_usage_percent?: number | null
checked_out: number
pool_size: number
overflow: number
max_capacity: number
pool_timeout: number
server_connections?: number | null
server_max_connections?: number | null
server_usage_percent?: number | null
storage_status?: MonitoringMetricStatus
storage_label?: string | null
storage_total_bytes?: number | null
storage_free_bytes?: number | null
storage_free_percent?: number | null
database_size_bytes?: number | null
storage_message?: string | null
message?: string | null
}
export interface MonitoringSystemStatus {
timestamp: string
users: {
total: number
active: number
}
providers: {
total: number
active: number
}
api_keys: {
total: number
active: number
}
today_stats: {
requests: number
tokens: number
cost_usd: number
}
recent_errors: number
system_metrics?: {
cpu: MonitoringCpuMetric
memory: MonitoringMemoryMetric
redis: MonitoringServiceMetric
postgres: MonitoringPostgresMetric
}
}
function parseNumericLike(value: unknown): number {
if (typeof value === 'number') {
return Number.isFinite(value) ? value : 0
}
if (typeof value === 'string') {
const normalized = value.trim().replace(/[$,\s]/g, '')
if (!normalized) return 0
const parsed = Number(normalized)
return Number.isFinite(parsed) ? parsed : 0
}
return 0
}
function parseNullableNumericLike(value: unknown): number | null {
if (value == null) {
return null
}
if (typeof value === 'number') {
return Number.isFinite(value) ? value : null
}
if (typeof value === 'string') {
const normalized = value.trim().replace(/[$,\s]/g, '')
if (!normalized) return null
const parsed = Number(normalized)
return Number.isFinite(parsed) ? parsed : null
}
return null
}
function normalizeMonitoringSystemStatus(data: MonitoringSystemStatus): MonitoringSystemStatus {
return {
...data,
users: {
total: parseNumericLike(data.users?.total),
active: parseNumericLike(data.users?.active),
},
providers: {
total: parseNumericLike(data.providers?.total),
active: parseNumericLike(data.providers?.active),
},
api_keys: {
total: parseNumericLike(data.api_keys?.total),
active: parseNumericLike(data.api_keys?.active),
},
today_stats: {
requests: parseNumericLike(data.today_stats?.requests),
tokens: parseNumericLike(data.today_stats?.tokens),
cost_usd: parseNumericLike(data.today_stats?.cost_usd),
},
recent_errors: parseNumericLike(data.recent_errors),
system_metrics: data.system_metrics ? {
cpu: {
status: data.system_metrics.cpu?.status ?? 'unknown',
label: data.system_metrics.cpu?.label ?? '未知',
usage_percent: parseNullableNumericLike(data.system_metrics.cpu?.usage_percent),
load_percent: parseNullableNumericLike(data.system_metrics.cpu?.load_percent),
core_count: parseNumericLike(data.system_metrics.cpu?.core_count),
message: data.system_metrics.cpu?.message ?? null,
},
memory: {
status: data.system_metrics.memory?.status ?? 'unknown',
label: data.system_metrics.memory?.label ?? '未知',
used_percent: parseNullableNumericLike(data.system_metrics.memory?.used_percent),
used_bytes: parseNullableNumericLike(data.system_metrics.memory?.used_bytes),
available_bytes: parseNullableNumericLike(data.system_metrics.memory?.available_bytes),
total_bytes: parseNullableNumericLike(data.system_metrics.memory?.total_bytes),
message: data.system_metrics.memory?.message ?? null,
},
redis: {
status: data.system_metrics.redis?.status ?? 'unknown',
label: data.system_metrics.redis?.label ?? '未知',
latency_ms: parseNullableNumericLike(data.system_metrics.redis?.latency_ms),
memory_status: data.system_metrics.redis?.memory_status ?? 'unknown',
memory_label: data.system_metrics.redis?.memory_label ?? '未知',
used_memory_bytes: parseNullableNumericLike(data.system_metrics.redis?.used_memory_bytes),
peak_memory_bytes: parseNullableNumericLike(data.system_metrics.redis?.peak_memory_bytes),
maxmemory_bytes: parseNullableNumericLike(data.system_metrics.redis?.maxmemory_bytes),
memory_ceiling_bytes: parseNullableNumericLike(data.system_metrics.redis?.memory_ceiling_bytes),
memory_source: (data.system_metrics.redis?.memory_source as MonitoringServiceMetric['memory_source']) ?? 'unknown',
available_memory_bytes: parseNullableNumericLike(data.system_metrics.redis?.available_memory_bytes),
memory_percent: parseNullableNumericLike(data.system_metrics.redis?.memory_percent),
message: data.system_metrics.redis?.message ?? null,
},
postgres: {
status: data.system_metrics.postgres?.status ?? 'unknown',
label: data.system_metrics.postgres?.label ?? '未知',
usage_percent: parseNullableNumericLike(data.system_metrics.postgres?.usage_percent),
pool_usage_percent: parseNullableNumericLike(data.system_metrics.postgres?.pool_usage_percent),
checked_out: parseNumericLike(data.system_metrics.postgres?.checked_out),
pool_size: parseNumericLike(data.system_metrics.postgres?.pool_size),
overflow: parseNumericLike(data.system_metrics.postgres?.overflow),
max_capacity: parseNumericLike(data.system_metrics.postgres?.max_capacity),
pool_timeout: parseNumericLike(data.system_metrics.postgres?.pool_timeout),
server_connections: parseNullableNumericLike(data.system_metrics.postgres?.server_connections),
server_max_connections: parseNullableNumericLike(data.system_metrics.postgres?.server_max_connections),
server_usage_percent: parseNullableNumericLike(data.system_metrics.postgres?.server_usage_percent),
storage_status: data.system_metrics.postgres?.storage_status ?? 'unknown',
storage_label: data.system_metrics.postgres?.storage_label ?? '未知',
storage_total_bytes: parseNullableNumericLike(data.system_metrics.postgres?.storage_total_bytes),
storage_free_bytes: parseNullableNumericLike(data.system_metrics.postgres?.storage_free_bytes),
storage_free_percent: parseNullableNumericLike(data.system_metrics.postgres?.storage_free_percent),
database_size_bytes: parseNullableNumericLike(data.system_metrics.postgres?.database_size_bytes),
storage_message: data.system_metrics.postgres?.storage_message ?? null,
message: data.system_metrics.postgres?.message ?? null,
},
} : undefined,
}
}
function normalizeAuditResponse(data: Record<string, unknown>): AuditLogsResponse {
const items: AuditLog[] = (data.items ?? data.logs ?? []) as AuditLog[]
const meta: PaginationMeta = (data.meta as PaginationMeta) ?? {
@@ -78,6 +275,11 @@ export const auditApi = {
return response.data
},
async getSystemStatus(): Promise<MonitoringSystemStatus> {
const response = await apiClient.get<MonitoringSystemStatus>('/api/admin/monitoring/system-status')
return normalizeMonitoringSystemStatus(response.data)
},
// 分析用户行为 (管理员)
async analyzeUserBehavior(userId: number, days: number = 7): Promise<{
analysis: Record<string, unknown>
+2 -23
View File
@@ -9,7 +9,6 @@ export interface LoginRequest {
export interface LoginResponse {
access_token: string
refresh_token?: string
token_type?: string
expires_in?: number
user_id?: string // UUID
@@ -25,13 +24,6 @@ export interface UserPreferences {
[key: string]: unknown // 允许扩展其他偏好设置
}
export interface UserStats {
total_requests?: number
total_cost?: number
last_request_at?: string
[key: string]: unknown // 允许扩展其他统计数据
}
export interface SendVerificationCodeRequest {
email: string
}
@@ -114,23 +106,15 @@ export interface User {
role: string // 'admin' or 'user'
is_active: boolean
billing?: BillingSummary
allowed_providers?: string[] | null // 允许使用的提供商 ID 列表
allowed_api_formats?: string[] | null // 允许使用的 API 格式列表
allowed_models?: string[] | null // 允许使用的模型名称列表
created_at: string
last_login_at?: string
preferences?: UserPreferences
stats?: UserStats
}
export const authApi = {
async login(credentials: LoginRequest): Promise<LoginResponse> {
const response = await apiClient.post<LoginResponse>('/api/auth/login', credentials)
apiClient.setToken(response.data.access_token)
// 后端暂时没有返回 refresh_token
if (response.data.refresh_token) {
localStorage.setItem('refresh_token', response.data.refresh_token)
}
return response.data
},
@@ -152,14 +136,9 @@ export const authApi = {
return response.data
},
async refreshToken(refreshToken: string): Promise<LoginResponse> {
const response = await apiClient.post<LoginResponse>('/api/auth/refresh', {
refresh_token: refreshToken
})
async refreshToken(): Promise<LoginResponse> {
const response = await apiClient.post<LoginResponse>('/api/auth/refresh', {})
apiClient.setToken(response.data.access_token)
if (response.data.refresh_token) {
localStorage.setItem('refresh_token', response.data.refresh_token)
}
return response.data
},
+19 -3
View File
@@ -288,7 +288,12 @@ export const cacheAnalysisApi = {
api_key_id?: string
hours?: number
}): Promise<TTLAnalysisResponse> {
const response = await api.get('/api/admin/usage/cache-affinity/ttl-analysis', { params })
const response = await api.post('/api/analytics/cache-affinity/ttl-analysis', {
scope: { kind: 'global' },
user_id: params?.user_id ?? null,
api_key_id: params?.api_key_id ?? null,
hours: params?.hours,
})
return response.data
},
@@ -300,7 +305,12 @@ export const cacheAnalysisApi = {
api_key_id?: string
hours?: number
}): Promise<CacheHitAnalysisResponse> {
const response = await api.get('/api/admin/usage/cache-affinity/hit-analysis', { params })
const response = await api.post('/api/analytics/cache-affinity/hit-analysis', {
scope: { kind: 'global' },
user_id: params?.user_id ?? null,
api_key_id: params?.api_key_id ?? null,
hours: params?.hours,
})
return response.data
},
@@ -319,7 +329,13 @@ export const cacheAnalysisApi = {
return cachedRequest(
cacheKey,
async () => {
const response = await api.get('/api/admin/usage/cache-affinity/interval-timeline', { params })
const response = await api.post('/api/analytics/interval-timeline', {
scope: params?.user_id ? { kind: 'user', user_id: params.user_id } : { kind: 'global' },
user_id: params?.user_id ?? null,
hours: params?.hours,
limit: params?.limit,
include_user_info: params?.include_user_info ?? false,
})
return response.data
},
30000
+87 -48
View File
@@ -3,10 +3,13 @@ import type { AxiosInstance, AxiosRequestConfig, AxiosResponse, InternalAxiosReq
import { NETWORK_CONFIG, AUTH_CONFIG } from '@/config/constants'
import { isDemoMode } from '@/config/demo'
import { handleMockRequest, setMockUserToken } from '@/mocks'
import { getClientDeviceId } from '@/utils/deviceId'
import { CrossTabRefreshCoordinator } from '@/utils/crossTabRefresh'
import { log } from '@/utils/logger'
// 在开发环境下使用代理,生产环境使用环境变量
const API_BASE_URL = import.meta.env.VITE_API_URL || ''
export const AUTH_STATE_CHANGE_EVENT = 'aether-auth-state-change'
/**
* 判断请求是否为公共端点
@@ -31,17 +34,15 @@ function isAuthRequest(url?: string): boolean {
}
/**
* 判断是否为可刷新的认证错误
* 判断 403 错误是否表示用户账号级别的问题(需要清除认证并跳转)
*/
function isRefreshableAuthError(errorDetail: string): boolean {
const nonRefreshableErrors = [
function isAccountLevelForbidden(status: number, errorDetail: string): boolean {
if (status !== 403) return false
const accountErrors = [
'用户不存在或已禁用',
'需要管理员权限',
'权限不足',
'用户已禁用',
]
return !nonRefreshableErrors.some((msg) => errorDetail.includes(msg))
return accountErrors.some((msg) => errorDetail.includes(msg))
}
/**
@@ -83,12 +84,21 @@ class ApiClient {
private client: AxiosInstance
private token: string | null = null
private isRefreshing = false
private refreshPromise: Promise<AxiosResponse> | null = null
private refreshPromise: Promise<string> | null = null
private readonly refreshCoordinator = new CrossTabRefreshCoordinator()
private readonly onStorageSync = (event: StorageEvent): void => {
if (event.key !== 'access_token') {
return
}
this.syncTokenState(event.newValue)
}
constructor() {
this.client = axios.create({
baseURL: API_BASE_URL,
timeout: NETWORK_CONFIG.API_TIMEOUT,
withCredentials: true,
headers: {
'Content-Type': 'application/json',
},
@@ -99,6 +109,7 @@ class ApiClient {
this.client.defaults.adapter = createDemoAdapter(defaultAdapter)
this.setupInterceptors()
this.setupCrossTabAuthSync()
}
/**
@@ -108,6 +119,10 @@ class ApiClient {
// 请求拦截器 - 仅处理认证
this.client.interceptors.request.use(
(config) => {
if (config.url?.includes('/api/')) {
config.headers['X-Client-Device-Id'] = getClientDeviceId()
}
const requiresAuth = !isPublicEndpoint(config.url, config.method) &&
config.url?.includes('/api/')
@@ -129,6 +144,23 @@ class ApiClient {
)
}
private setupCrossTabAuthSync(): void {
if (typeof window !== 'undefined') {
window.addEventListener('storage', this.onStorageSync)
}
}
private emitAuthStateChange(token: string | null): void {
if (typeof window === 'undefined') {
return
}
window.dispatchEvent(
new CustomEvent<{ token: string | null }>(AUTH_STATE_CHANGE_EVENT, {
detail: { token },
})
)
}
/**
* 处理响应错误
*/
@@ -155,8 +187,22 @@ class ApiClient {
return Promise.reject(error)
}
const status = error.response?.status ?? 0
// 处理 403 用户账号级别错误(被禁用/删除)
if (status === 403) {
const rawDetail = (error.response?.data as Record<string, unknown>)?.detail
const errorDetail = typeof rawDetail === 'string' ? rawDetail : ''
if (isAccountLevelForbidden(status, errorDetail)) {
log.info('User account issue detected, clearing auth', { errorDetail })
this.clearAuth()
window.location.href = '/'
return Promise.reject(error)
}
}
// 处理401错误
if (error.response?.status === 401) {
if (status === 401) {
return this.handle401Error(error, originalRequest)
}
@@ -177,25 +223,7 @@ class ApiClient {
return Promise.reject(error)
}
const errorDetail = (error.response?.data as Record<string, unknown>)?.detail as string || ''
log.debug('Got 401 error, attempting token refresh', { errorDetail })
// 检查是否为业务相关的401错误(用户被禁用/删除等)
if (!isRefreshableAuthError(errorDetail)) {
log.info('User account issue detected, logging out and redirecting to home', { errorDetail })
this.clearAuth()
// 跳转到首页
window.location.href = '/'
return Promise.reject(error)
}
// 获取refresh token
const refreshToken = localStorage.getItem('refresh_token')
if (!refreshToken) {
log.info('No refresh token available, clearing invalid token')
this.clearAuth()
return Promise.reject(error)
}
log.debug('Got 401 error, attempting token refresh')
// 标记为已重试
originalRequest._retry = true
@@ -210,8 +238,8 @@ class ApiClient {
// 如果正在刷新,等待刷新完成
if (this.isRefreshing) {
try {
await this.refreshPromise
originalRequest.headers.Authorization = `Bearer ${this.getToken()}`
const accessToken = await this.refreshPromise
originalRequest.headers.Authorization = `Bearer ${accessToken}`
return this.client.request(originalRequest)
} catch {
return Promise.reject(error)
@@ -219,29 +247,27 @@ class ApiClient {
}
// 开始刷新token
return this.refreshTokenAndRetry(refreshToken, originalRequest, error)
return this.refreshTokenAndRetry(originalRequest, error)
}
/**
* 刷新token并重试原始请求
*/
private async refreshTokenAndRetry(
refreshToken: string,
originalRequest: InternalAxiosRequestConfig,
originalError: import('axios').AxiosError
): Promise<AxiosResponse> {
this.isRefreshing = true
this.refreshPromise = this.refreshToken(refreshToken)
this.refreshPromise = this.coordinatedRefresh()
try {
const response = await this.refreshPromise
this.setToken(response.data.access_token)
localStorage.setItem('refresh_token', response.data.refresh_token)
const accessToken = await this.refreshPromise
this.setToken(accessToken)
this.isRefreshing = false
this.refreshPromise = null
// 重试原始请求
originalRequest.headers.Authorization = `Bearer ${response.data.access_token}`
originalRequest.headers.Authorization = `Bearer ${accessToken}`
return this.client.request(originalRequest)
} catch (refreshError: unknown) {
log.error('Token refresh failed', refreshError instanceof Error ? refreshError.message : String(refreshError))
@@ -252,15 +278,29 @@ class ApiClient {
}
}
setToken(token: string): void {
private async coordinatedRefresh(): Promise<string> {
return this.refreshCoordinator.run(async () => {
const response = await this.refreshToken()
const accessToken = response.data.access_token
if (!accessToken) {
throw new Error('Refresh response missing access token')
}
return accessToken
})
}
private syncTokenState(token: string | null): void {
this.token = token
localStorage.setItem('access_token', token)
// 同步到 mock handler
if (isDemoMode()) {
setMockUserToken(token)
}
}
setToken(token: string): void {
this.syncTokenState(token)
localStorage.setItem('access_token', token)
}
getToken(): string | null {
if (!this.token) {
this.token = localStorage.getItem('access_token')
@@ -273,18 +313,17 @@ class ApiClient {
}
clearAuth(): void {
this.token = null
const hadAuth = this.token !== null || localStorage.getItem('access_token') !== null
this.syncTokenState(null)
localStorage.removeItem('access_token')
localStorage.removeItem('refresh_token')
// 同步清除 mock token
if (isDemoMode()) {
setMockUserToken(null)
// 同标签页内清理认证状态时不会触发 storage 事件,这里主动广播一次。
if (hadAuth) {
this.emitAuthStateChange(null)
}
}
async refreshToken(refreshToken: string): Promise<AxiosResponse> {
// refreshToken 会通过 adapter 处理 Demo 模式
return this.client.post('/api/auth/refresh', { refresh_token: refreshToken })
async refreshToken(): Promise<AxiosResponse> {
return this.client.post('/api/auth/refresh', {})
}
// 以下方法直接委托给 axios client,Demo 模式由 adapter 统一处理
-370
View File
@@ -1,370 +0,0 @@
import apiClient from './client'
import { cachedRequest, buildCacheKey } from '@/utils/cache'
export interface DashboardStat {
name: string
value: string
subValue?: string
change?: string
changeType?: 'increase' | 'decrease' | 'neutral'
extraBadge?: string
icon: string
}
export interface RecentRequest {
id: string // UUID
user: string
model: string
tokens: number
time: string
}
export interface ProviderStatus {
name: string
status: 'active' | 'inactive'
requests: number
}
// 系统健康指标(管理员专用)
export interface SystemHealth {
avg_response_time: number
error_rate: number
error_requests: number
fallback_count: number
total_requests: number
}
// 成本统计(管理员专用)
export interface CostStats {
total_cost: number
total_actual_cost: number
cost_savings: number
}
// 缓存统计
export interface CacheStats {
cache_creation_tokens: number
cache_read_tokens: number
cache_creation_cost?: number
cache_read_cost?: number
cache_hit_rate?: number
total_cache_tokens: number
}
// 用户统计(管理员专用)
export interface UserStats {
total: number
active: number
}
// Token 详细分类
export interface TokenBreakdown {
input: number
output: number
cache_creation: number
cache_read: number
}
export interface DashboardStatsResponse {
stats: DashboardStat[]
today?: {
requests: number
tokens: number
cost: number
actual_cost?: number
cache_creation_tokens?: number
cache_read_tokens?: number
}
api_keys?: {
total: number
active: number
}
tokens?: {
month: number
}
// 管理员专用字段
system_health?: SystemHealth
cost_stats?: CostStats
cache_stats?: CacheStats
users?: UserStats
token_breakdown?: TokenBreakdown
// 普通用户专用字段
monthly_cost?: number
}
export interface RecentRequestsResponse {
requests: RecentRequest[]
}
export interface ProviderStatusResponse {
providers: ProviderStatus[]
}
// 视频/图像/音频计费信息
export interface VideoBilling {
task_type: 'video' | 'image' | 'audio'
duration_seconds?: number // 视频时长(秒)
resolution?: string // 分辨率
video_price_per_second?: number // 每秒单价
video_cost?: number // 视频费用
cost?: number // 总费用
rule_name?: string // 计费规则名称
expression?: string // 计费公式
status?: string // 计费状态
}
export interface RequestDetail {
id: string // UUID
request_id: string
user: {
id: string // UUID
username: string
email: string
}
api_key: {
id: string // UUID
name: string
display: string
}
provider: string
api_format?: string
model: string
target_model?: string | null // 映射后的目标模型名
tokens: {
input: number
output: number
total: number
}
cost: {
input: number
output: number
total: number
}
// Additional token fields
input_tokens?: number
output_tokens?: number
total_tokens?: number
cache_creation_input_tokens?: number
cache_creation_input_tokens_5m?: number
cache_creation_input_tokens_1h?: number
cache_read_input_tokens?: number
// Additional cost fields
input_cost?: number
output_cost?: number
total_cost?: number
cache_creation_cost?: number
cache_read_cost?: number
request_cost?: number // 按次计费费用
// Historical pricing fields (per 1M tokens)
input_price_per_1m?: number
output_price_per_1m?: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
price_per_request?: number // 按次计费价格
request_type: string
is_stream: boolean
status_code: number
status?: string // pending, streaming, completed, failed, cancelled
error_message?: string
response_time_ms: number
created_at: string
request_headers?: Record<string, unknown>
request_body?: Record<string, unknown>
provider_request_headers?: Record<string, unknown>
provider_request_body?: Record<string, unknown>
response_headers?: Record<string, unknown>
client_response_headers?: Record<string, unknown>
response_body?: Record<string, unknown>
client_response_body?: Record<string, unknown>
has_request_body?: boolean
has_provider_request_body?: boolean
has_response_body?: boolean
has_client_response_body?: boolean
metadata?: Record<string, unknown>
// 阶梯计费信息
tiered_pricing?: {
total_input_context: number // 总输入上下文 (input + cache_read)
tier_index: number // 命中的阶梯索引 (0-based)
tier_count: number // 阶梯总数
source?: 'provider' | 'global' // 定价来源: 提供商或全局
current_tier: { // 当前命中的阶梯配置
up_to?: number | null
input_price_per_1m: number
output_price_per_1m: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
cache_ttl_pricing?: Array<{
ttl_minutes: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
}>
}
tiers: Array<{ // 完整阶梯配置列表
up_to?: number | null
input_price_per_1m: number
output_price_per_1m: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
cache_ttl_pricing?: Array<{
ttl_minutes: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
}>
}>
} | null
// 视频/图像/音频计费信息
video_billing?: VideoBilling | null
}
export interface CurlData {
url: string
method: string
headers: Record<string, string>
body: Record<string, unknown>
curl: string
}
export interface ReplayRequest {
provider_id?: string
endpoint_id?: string
api_key_id?: string
body_override?: Record<string, unknown>
}
export interface ReplayResponse {
url: string
provider: string
status_code: number
response_headers: Record<string, string>
response_body: Record<string, unknown>
response_time_ms: number
}
export interface ModelBreakdown {
model: string
requests: number
tokens: number
cost: number
}
export interface ModelSummary {
model: string
requests: number
tokens: number
cost: number
avg_response_time: number
cost_per_request: number
tokens_per_request: number
}
export interface ProviderSummary {
provider: string
requests: number
tokens: number
cost: number
}
export interface DailyStat {
date: string // ISO date string
requests: number
tokens: number
cost: number
avg_response_time: number // in seconds
unique_models: number
unique_providers?: number // 仅管理员返回
model_breakdown: ModelBreakdown[]
}
export interface DailyStatsResponse {
daily_stats: DailyStat[]
model_summary: ModelSummary[]
provider_summary?: ProviderSummary[] // 仅管理员返回
period: {
start_date: string
end_date: string
days: number
}
}
export interface TimeRangeParams {
start_date?: string
end_date?: string
preset?: string
granularity?: 'hour' | 'day' | 'week' | 'month'
timezone?: string
tz_offset_minutes?: number
}
export const dashboardApi = {
// 获取仪表盘统计数据
async getStats(params?: TimeRangeParams): Promise<DashboardStatsResponse> {
const cacheKey = buildCacheKey('dashboard:stats', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<DashboardStatsResponse>('/api/dashboard/stats', { params })
return response.data
},
10 * 1000
)
},
// 获取最近的请求记录
async getRecentRequests(limit: number = 10): Promise<RecentRequest[]> {
const response = await apiClient.get<RecentRequestsResponse>('/api/dashboard/recent-requests', {
params: { limit }
})
return response.data.requests
},
// 获取提供商状态
async getProviderStatus(): Promise<ProviderStatus[]> {
return cachedRequest(
'dashboard:provider-status',
async () => {
const response = await apiClient.get<ProviderStatusResponse>('/api/dashboard/provider-status')
return response.data.providers
},
20 * 1000
)
},
// 获取请求详情
// NOTE: This method now calls the new RESTful API at /api/admin/usage/{id}
async getRequestDetail(requestId: string, options: { includeBodies?: boolean } = {}): Promise<RequestDetail> {
const response = await apiClient.get<RequestDetail>(`/api/admin/usage/${requestId}`, {
params: { include_bodies: options.includeBodies ?? true },
})
return response.data
},
// 获取每日统计数据
async getDailyStats(params?: TimeRangeParams & { days?: number }): Promise<DailyStatsResponse> {
const cacheKey = buildCacheKey('dashboard:daily-stats', params)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<DailyStatsResponse>('/api/dashboard/daily-stats', {
params
})
return response.data
},
20 * 1000
)
},
// 获取 cURL 命令数据(含明文 API Key)
async getCurlData(requestId: string): Promise<CurlData> {
const response = await apiClient.get<CurlData>(`/api/admin/usage/${requestId}/curl`)
return response.data
},
// 回放请求到提供商
async replayRequest(requestId: string, params?: ReplayRequest): Promise<ReplayResponse> {
const response = await apiClient.post<ReplayResponse>(
`/api/admin/usage/${requestId}/replay`,
params || {}
)
return response.data
}
}
+73
View File
@@ -4,6 +4,34 @@ import type { EndpointAPIKey, AllowedModels } from './types'
// Re-export types for convenience
export type { EndpointAPIKey, AllowedModels }
export interface GroupedFormatKey {
id: string
provider_id: string
name: string
auth_type?: string
api_key_masked: string
internal_priority: number
global_priority_by_format: Record<string, number> | null
format_priority: number | null
rate_multipliers: Record<string, number> | null
is_active: boolean
provider_active: boolean
pool_enabled: boolean
circuit_breaker_open: boolean
provider_name: string
api_format: string
api_formats: string[]
capabilities: string[]
health_score: number | null
success_rate: number | null
avg_response_time_ms: number | null
request_count: number
}
function toNumberOrNull(value: unknown): number | null {
return typeof value === 'number' && Number.isFinite(value) ? value : null
}
/**
* 能力定义类型
*/
@@ -125,6 +153,51 @@ export async function getProviderKeys(providerId: string): Promise<EndpointAPIKe
return allKeys
}
/**
* 获取按 API 格式分组的 Key 列表
*/
export async function getKeysGroupedByFormat(): Promise<Record<string, GroupedFormatKey[]>> {
const response = await client.get('/api/admin/endpoints/keys/grouped-by-format')
const grouped = response.data as Record<string, Array<Record<string, unknown>>>
const result: Record<string, GroupedFormatKey[]> = {}
for (const [apiFormat, keys] of Object.entries(grouped || {})) {
if (!Array.isArray(keys)) continue
result[apiFormat] = keys.map((key) => ({
id: String(key.id || ''),
provider_id: String(key.provider_id || ''),
name: String(key.name || 'Unnamed Key'),
auth_type: typeof key.auth_type === 'string' ? key.auth_type : undefined,
api_key_masked: String(key.api_key_masked || '***'),
internal_priority: toNumberOrNull(key.internal_priority) ?? 0,
global_priority_by_format:
key.global_priority_by_format && typeof key.global_priority_by_format === 'object'
? (key.global_priority_by_format as Record<string, number>)
: null,
format_priority: toNumberOrNull(key.format_priority),
rate_multipliers:
key.rate_multipliers && typeof key.rate_multipliers === 'object'
? (key.rate_multipliers as Record<string, number>)
: null,
is_active: key.is_active !== false,
provider_active: key.provider_active !== false,
pool_enabled: key.pool_enabled === true,
circuit_breaker_open: key.circuit_breaker_open === true,
provider_name: String(key.provider_name || 'Unknown Provider'),
api_format: typeof key.api_format === 'string' ? key.api_format : apiFormat,
api_formats: Array.isArray(key.api_formats) ? key.api_formats.map(item => String(item)) : [apiFormat],
capabilities: Array.isArray(key.capabilities) ? key.capabilities.map(item => String(item)) : [],
health_score: toNumberOrNull(key.health_score),
success_rate: toNumberOrNull(key.success_rate),
avg_response_time_ms: toNumberOrNull(key.avg_response_time_ms),
request_count: toNumberOrNull(key.request_count) ?? 0,
}))
}
return result
}
/**
* 为 Provider 添加 Key
*/
+11 -2
View File
@@ -1,6 +1,7 @@
import client from '../client'
import { dedupedRequest } from '@/utils/cache'
import type { AllowedModels, OAuthOrganizationInfo, ProxyConfig } from './types/provider'
import type { ProviderKeyStatusSnapshot } from './types/statusSnapshot'
const POOL_BATCH_ACTION_TIMEOUT_MS = 5 * 60 * 1000
@@ -99,12 +100,20 @@ export interface PoolKeyDetail {
is_active: boolean
auth_type: string
oauth_expires_at?: number | null
oauth_invalid_at?: number | null
oauth_invalid_reason?: string | null
oauth_invalid_at?: number | null // 兼容字段;优先使用 status_snapshot.oauth
oauth_invalid_reason?: string | null // 兼容字段;优先使用 status_snapshot.oauth
oauth_plan_type?: string | null
oauth_account_id?: string | null
oauth_account_user_id?: string | null
oauth_account_name?: string | null
oauth_organizations?: OAuthOrganizationInfo[] | null
account_status_code?: string | null // 兼容字段;优先使用 status_snapshot.account
account_status_label?: string | null // 兼容字段;优先使用 status_snapshot.account
account_status_reason?: string | null // 兼容字段;优先使用 status_snapshot.account
account_status_blocked?: boolean // 兼容字段;优先使用 status_snapshot.account
account_status_recoverable?: boolean // 兼容字段;优先使用 status_snapshot.account
account_status_source?: string | null // 兼容字段;优先使用 status_snapshot.account
status_snapshot?: ProviderKeyStatusSnapshot | null
quota_updated_at?: number | null
health_score?: number
circuit_breaker_open?: boolean
+3 -1
View File
@@ -18,6 +18,8 @@ export interface ProviderOAuthCompleteResponse {
expires_at?: number | null
has_refresh_token: boolean
email?: string | null
account_state_recheck_attempted?: boolean
account_state_recheck_error?: string | null
}
export interface ProviderOAuthCompleteResponseWithKey {
@@ -92,7 +94,7 @@ export async function completeProviderLevelOAuth(
export async function importProviderRefreshToken(
providerId: string,
data: { refresh_token: string; name?: string; proxy_node_id?: string }
data: { refresh_token?: string; access_token?: string; name?: string; proxy_node_id?: string }
): Promise<ProviderOAuthCompleteResponseWithKey> {
const resp = await client.post(`/api/admin/provider-oauth/providers/${providerId}/import-refresh-token`, data)
return resp.data
+7
View File
@@ -194,6 +194,8 @@ export interface TestModelFailoverRequest {
api_format?: string
endpoint_id?: string
message?: string
request_headers?: Record<string, unknown>
request_body?: Record<string, unknown>
request_id?: string
concurrency?: number
}
@@ -212,6 +214,11 @@ export interface TestAttemptDetail {
error_message?: string | null
status_code?: number | null
latency_ms?: number | null
request_url?: string | null
request_headers?: Record<string, unknown> | null
request_body?: unknown
response_headers?: Record<string, unknown> | null
response_body?: unknown
}
export interface TestModelFailoverResponse {
+27 -5
View File
@@ -1,3 +1,5 @@
import type { ProviderKeyStatusSnapshot } from './statusSnapshot'
/**
* 代理配置类型
* 支持两种模式:
@@ -42,8 +44,6 @@ export interface HeaderRuleRename {
to: string
}
export type HeaderRule = HeaderRuleSet | HeaderRuleDrop | HeaderRuleRename
/**
* 请求体规则类型
* - set: 设置/覆盖字段
@@ -136,10 +136,28 @@ export type BodyRuleConditionOp =
| 'exists' | 'not_exists'
| 'in' | 'type_is'
export interface BodyRuleCondition {
export interface BodyRuleConditionLeaf {
path: string
op: BodyRuleConditionOp
value?: unknown // exists / not_exists 不需要 value
source?: 'original' | 'current'
}
export interface BodyRuleConditionAll {
all: BodyRuleCondition[]
}
export interface BodyRuleConditionAny {
any: BodyRuleCondition[]
}
export type BodyRuleCondition =
| BodyRuleConditionLeaf
| BodyRuleConditionAll
| BodyRuleConditionAny
export type HeaderRule = (HeaderRuleSet | HeaderRuleDrop | HeaderRuleRename) & {
condition?: BodyRuleCondition
}
/**
@@ -272,9 +290,12 @@ export interface EndpointAPIKey {
oauth_plan_type?: string | null // Codex 订阅类型: plus/free/team/enterprise
oauth_account_id?: string | null // Codex ChatGPT 账号 ID
oauth_account_user_id?: string | null // Codex ChatGPT account-user 联合 ID
oauth_account_name?: string | null
oauth_organizations?: OAuthOrganizationInfo[] | null // OAuth 关联组织/工作区摘要
oauth_invalid_at?: number | null // OAuth Token 失效时间(Unix 时间戳)
oauth_invalid_reason?: string | null // OAuth Token 失效原因
oauth_temporary?: boolean | null // 是否为仅 Access Token 导入的临时 OAuth 账号
oauth_invalid_at?: number | null // 兼容字段;优先使用 status_snapshot.oauth
oauth_invalid_reason?: string | null // 兼容字段;优先使用 status_snapshot.oauth
status_snapshot?: ProviderKeyStatusSnapshot | null
// 上游元数据(由上游响应采集,如 Codex 额度信息 / Antigravity 配额信息)
upstream_metadata?: UpstreamMetadata | null
// Key 级别代理配置(覆盖 Provider 级别代理)
@@ -574,6 +595,7 @@ export interface HealthSummary {
total: number
active: number
unhealthy: number
circuit_open?: number
}
}
@@ -0,0 +1,36 @@
export interface OAuthStatusSnapshot {
code: 'none' | 'valid' | 'expiring' | 'expired' | 'invalid' | 'check_failed'
label?: string | null
reason?: string | null
expires_at?: number | null
invalid_at?: number | null
source?: string | null
requires_reauth?: boolean
expiring_soon?: boolean
}
export interface AccountStatusSnapshot {
code: string
label?: string | null
reason?: string | null
blocked: boolean
source?: string | null
recoverable?: boolean
}
export interface QuotaStatusSnapshot {
code: 'unknown' | 'ok' | 'exhausted'
label?: string | null
reason?: string | null
exhausted: boolean
usage_ratio?: number | null
updated_at?: number | null
reset_seconds?: number | null
plan_type?: string | null
}
export interface ProviderKeyStatusSnapshot {
oauth: OAuthStatusSnapshot
account: AccountStatusSnapshot
quota: QuotaStatusSnapshot
}
+36 -175
View File
@@ -1,8 +1,9 @@
import apiClient from './client'
import type { ActivityHeatmap } from '@/types/activity'
import type { TieredPricingConfig } from './endpoints/types'
import { cachedRequest, buildCacheKey } from '@/utils/cache'
import type { BillingSummary } from './auth'
import type { UserSession } from '@/types/session'
export type { UserSession }
export interface Profile {
id: string // UUID
@@ -22,8 +23,6 @@ export interface Profile {
export interface UserPreferences {
avatar_url?: string
bio?: string
default_provider_id?: string // UUID
default_provider?: Record<string, unknown>
theme: string
language: string
timezone?: string
@@ -34,87 +33,6 @@ export interface UserPreferences {
}
}
// 提供商配置接口
export interface ProviderConfig {
provider_id: string
priority: number // 优先级(越高越优先)
weight: number // 负载均衡权重
enabled: boolean // 是否启用
}
// 使用记录接口
export interface UsageRecordDetail {
id: string
provider: string
model: string
input_tokens: number
output_tokens: number
total_tokens: number
cost: number // 官方费率
actual_cost?: number // 倍率消耗(仅管理员可见)
rate_multiplier?: number // 成本倍率(仅管理员可见)
response_time_ms?: number
is_stream: boolean
created_at: string
cache_creation_input_tokens?: number
cache_read_input_tokens?: number
status_code: number
error_message?: string
input_price_per_1m: number
output_price_per_1m: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
price_per_request?: number // 按次计费价格
api_key?: {
id: string
name: string
display: string
}
}
// 模型统计接口
export interface ModelSummary {
model: string
requests: number
input_tokens: number
output_tokens: number
total_tokens: number
total_cost_usd: number
actual_total_cost_usd?: number // 倍率消耗(仅管理员可见)
}
// 提供商统计接口
export interface ProviderSummary {
provider: string
requests: number
total_tokens: number
total_cost_usd: number
success_rate: number | null
avg_response_time_ms: number | null
}
// 使用统计响应接口
export interface UsageResponse {
total_requests: number
total_input_tokens: number
total_output_tokens: number
total_tokens: number
total_cost: number // 官方费率
total_actual_cost?: number // 倍率消耗(仅管理员可见)
avg_response_time: number
billing: BillingSummary
summary_by_model: ModelSummary[]
summary_by_provider?: ProviderSummary[]
pagination?: {
total: number
limit: number
offset: number
has_more: boolean
}
records: UsageRecordDetail[]
activity_heatmap?: ActivityHeatmap | null
}
export interface ApiKey {
id: string // UUID
name: string
@@ -126,7 +44,7 @@ export interface ApiKey {
created_at: string
total_requests?: number
total_cost_usd?: number
allowed_providers?: ProviderConfig[]
rate_limit?: number | null
force_capabilities?: Record<string, boolean> | null // 强制能力配置
}
@@ -159,14 +77,36 @@ export const meApi = {
return response.data
},
async listSessions(): Promise<UserSession[]> {
const response = await apiClient.get<UserSession[]>('/api/users/me/sessions')
return response.data
},
async updateSessionLabel(sessionId: string, deviceLabel: string): Promise<UserSession> {
const response = await apiClient.patch<UserSession>(`/api/users/me/sessions/${sessionId}`, {
device_label: deviceLabel,
})
return response.data
},
async revokeSession(sessionId: string): Promise<{ message: string }> {
const response = await apiClient.delete(`/api/users/me/sessions/${sessionId}`)
return response.data
},
async revokeOtherSessions(): Promise<{ message: string; revoked_count: number }> {
const response = await apiClient.delete('/api/users/me/sessions/others')
return response.data
},
// API密钥管理
async getApiKeys(): Promise<ApiKey[]> {
const response = await apiClient.get<ApiKey[]>('/api/users/me/api-keys')
return response.data
},
async createApiKey(name: string): Promise<ApiKey> {
const response = await apiClient.post<ApiKey>('/api/users/me/api-keys', { name })
async createApiKey(data: { name: string; rate_limit?: number }): Promise<ApiKey> {
const response = await apiClient.post<ApiKey>('/api/users/me/api-keys', data)
return response.data
},
@@ -196,48 +136,14 @@ export const meApi = {
return response.data
},
// 使用统计
async getUsage(params?: {
start_date?: string
end_date?: string
preset?: string
timezone?: string
tz_offset_minutes?: number
search?: string // 通用搜索:密钥名、模型名
limit?: number
offset?: number
}): Promise<UsageResponse> {
const response = await apiClient.get<UsageResponse>('/api/users/me/usage', { params })
return response.data
},
// 获取活跃请求状态(用于轮询更新)
async getActiveRequests(ids?: string): Promise<{
requests: Array<{
id: string
status: 'pending' | 'streaming' | 'completed' | 'failed' | 'cancelled'
input_tokens: number
output_tokens: number
cache_creation_input_tokens?: number | null
cache_read_input_tokens?: number | null
cost: number
actual_cost?: number | null
rate_multiplier?: number | null
response_time_ms: number | null
first_byte_time_ms: number | null
api_format?: string | null
endpoint_api_format?: string | null
has_format_conversion?: boolean | null
}>
}> {
const params = ids ? { ids } : {}
const response = await apiClient.get('/api/users/me/usage/active', { params })
return response.data
},
// 获取可用的提供商
async getAvailableProviders(): Promise<Array<Record<string, unknown>>> {
const response = await apiClient.get('/api/users/me/providers')
async updateApiKey(
keyId: string,
data: { name?: string; rate_limit?: number | null }
): Promise<ApiKey & { message: string }> {
const response = await apiClient.put<ApiKey & { message: string }>(
`/api/users/me/api-keys/${keyId}`,
data
)
return response.data
},
@@ -281,16 +187,6 @@ export const meApi = {
return response.data
},
// 提供商绑定管理相关方法已移除,改为直接从可用提供商中选择
// API密钥提供商关联
async updateApiKeyProviders(keyId: string, data: {
allowed_providers?: ProviderConfig[]
}): Promise<{ message: string }> {
const response = await apiClient.put(`/api/users/me/api-keys/${keyId}/providers`, data)
return response.data
},
// API密钥能力配置
async updateApiKeyCapabilities(keyId: string, data: {
force_capabilities?: Record<string, boolean> | null
@@ -317,39 +213,4 @@ export const meApi = {
return response.data
},
// 获取请求间隔时间线(用于散点图)
async getIntervalTimeline(params?: {
hours?: number
limit?: number
}): Promise<{
analysis_period_hours: number
total_points: number
points: Array<{ x: string; y: number; model?: string }>
models?: string[]
}> {
const cacheKey = buildCacheKey('me:interval-timeline', params as Record<string, unknown> | undefined)
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get('/api/users/me/usage/interval-timeline', { params })
return response.data
},
30000
)
},
/**
* 获取活跃度热力图数据(用户)
* 后端已缓存5分钟
*/
async getActivityHeatmap(): Promise<ActivityHeatmap> {
return cachedRequest(
'me-activity-heatmap',
async () => {
const response = await apiClient.get<ActivityHeatmap>('/api/users/me/usage/heatmap')
return response.data
},
60000
)
}
}
+161
View File
@@ -0,0 +1,161 @@
import apiClient from './client'
export interface VideoBilling {
task_type: 'video' | 'image' | 'audio'
duration_seconds?: number
resolution?: string
video_price_per_second?: number
video_cost?: number
cost?: number
rule_name?: string
expression?: string
status?: string
}
export interface RequestDetail {
id: string
request_id: string
user: {
id: string | null
username: string | null
email: string | null
}
api_key: {
id: string | null
name: string | null
display: string | null
}
provider_api_key: {
id: string | null
name: string | null
}
provider: string
api_format?: string
model: string
target_model?: string | null
tokens: {
input: number
output: number
total: number
}
cost: {
input: number
output: number
total: number
}
input_tokens?: number
output_tokens?: number
total_tokens?: number
cache_creation_input_tokens?: number
cache_creation_input_tokens_5m?: number
cache_creation_input_tokens_1h?: number
cache_read_input_tokens?: number
input_cost?: number
output_cost?: number
total_cost?: number
cache_creation_cost?: number
cache_read_cost?: number
request_cost?: number
input_price_per_1m?: number
output_price_per_1m?: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
price_per_request?: number
request_type: string
is_stream: boolean
status_code: number
status?: string
error_message?: string
response_time_ms: number
created_at: string
request_headers?: Record<string, unknown>
request_body?: Record<string, unknown>
provider_request_headers?: Record<string, unknown>
provider_request_body?: Record<string, unknown>
response_headers?: Record<string, unknown>
client_response_headers?: Record<string, unknown>
response_body?: Record<string, unknown>
client_response_body?: Record<string, unknown>
has_request_body?: boolean
has_provider_request_body?: boolean
has_response_body?: boolean
has_client_response_body?: boolean
metadata?: Record<string, unknown>
tiered_pricing?: {
total_input_context: number
tier_index: number
tier_count: number
source?: 'provider' | 'global'
current_tier: {
up_to?: number | null
input_price_per_1m: number
output_price_per_1m: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
cache_ttl_pricing?: Array<{
ttl_minutes: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
}>
}
tiers: Array<{
up_to?: number | null
input_price_per_1m: number
output_price_per_1m: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
cache_ttl_pricing?: Array<{
ttl_minutes: number
cache_creation_price_per_1m?: number
cache_read_price_per_1m?: number
}>
}>
} | null
video_billing?: VideoBilling | null
}
export interface CurlData {
url: string
method: string
headers: Record<string, string>
body: Record<string, unknown>
curl: string
}
export interface ReplayRequest {
provider_id?: string
endpoint_id?: string
api_key_id?: string
body_override?: Record<string, unknown>
}
export interface ReplayResponse {
url: string
provider: string
status_code: number
response_headers: Record<string, string>
response_body: Record<string, unknown>
response_time_ms: number
}
export const requestDetailsApi = {
async getRequestDetail(requestId: string, options: { includeBodies?: boolean } = {}): Promise<RequestDetail> {
const response = await apiClient.get<RequestDetail>(`/api/admin/usage/${requestId}`, {
params: { include_bodies: options.includeBodies ?? true },
})
return response.data
},
async getCurlData(requestId: string): Promise<CurlData> {
const response = await apiClient.get<CurlData>(`/api/admin/usage/${requestId}/curl`)
return response.data
},
async replayRequest(requestId: string, params?: ReplayRequest): Promise<ReplayResponse> {
const response = await apiClient.post<ReplayResponse>(
`/api/admin/usage/${requestId}/replay`,
params || {},
)
return response.data
},
}
-241
View File
@@ -1,241 +0,0 @@
import apiClient from './client'
import { cachedRequest, dedupedRequest, buildCacheKey } from '@/utils/cache'
import type { ActivityHeatmap } from '@/types/activity'
export interface UsageRecord {
id: string // UUID
user_id: string // UUID
username?: string
provider_id?: string // UUID
provider_name?: string
model: string
input_tokens: number
output_tokens: number
cache_creation_input_tokens?: number
cache_read_input_tokens?: number
total_tokens: number
cost?: number
response_time?: number
created_at: string
has_fallback?: boolean // 🆕 是否发生了 fallback
}
export interface UsageStats {
total_requests: number
total_tokens: number
total_cost: number
total_actual_cost?: number
avg_response_time: number
today?: {
requests: number
tokens: number
cost: number
}
activity_heatmap?: ActivityHeatmap | null
}
export interface UsageByModel {
model: string
request_count: number
total_tokens: number
total_cost: number
avg_response_time?: number
}
export interface UsageByUser {
user_id: string // UUID
email: string
username: string
request_count: number
total_tokens: number
total_cost: number
}
export interface UsageByProvider {
provider_id: string
provider: string
request_count: number
total_tokens: number
total_cost: number
actual_cost: number
avg_response_time_ms: number
success_rate: number
error_count: number
}
export interface UsageByApiFormat {
api_format: string
request_count: number
total_tokens: number
total_cost: number
actual_cost: number
avg_response_time_ms: number
}
export interface UsageFilters {
user_id?: string // UUID
provider_id?: string // UUID
model?: string
start_date?: string
end_date?: string
preset?: string
granularity?: 'hour' | 'day' | 'week' | 'month'
timezone?: string
tz_offset_minutes?: number
page?: number
page_size?: number
}
export const usageApi = {
async getUsageRecords(filters?: UsageFilters): Promise<{
records: UsageRecord[]
total: number
page: number
page_size: number
}> {
const response = await apiClient.get('/api/usage', { params: filters })
return response.data
},
async getUsageStats(filters?: UsageFilters): Promise<UsageStats> {
// 为统计数据添加30秒缓存
const cacheKey = `usage-stats-${JSON.stringify(filters || {})}`
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<UsageStats>('/api/admin/usage/stats', { params: filters })
return response.data
},
30000 // 30秒缓存
)
},
/**
* Get usage aggregation by dimension (RESTful API)
* @param groupBy Aggregation dimension: 'model', 'user', 'provider', or 'api_format'
* @param filters Optional filters
*/
async getUsageAggregation<T = UsageByModel[] | UsageByUser[] | UsageByProvider[] | UsageByApiFormat[]>(
groupBy: 'model' | 'user' | 'provider' | 'api_format',
filters?: UsageFilters & { limit?: number }
): Promise<T> {
const cacheKey = `usage-aggregation-${groupBy}-${JSON.stringify(filters || {})}`
return cachedRequest(
cacheKey,
async () => {
const response = await apiClient.get<T>('/api/admin/usage/aggregation/stats', {
params: { group_by: groupBy, ...filters }
})
return response.data
},
30000 // 30秒缓存
)
},
// Shorthand methods using getUsageAggregation
async getUsageByModel(filters?: UsageFilters & { limit?: number }): Promise<UsageByModel[]> {
return this.getUsageAggregation<UsageByModel[]>('model', filters)
},
async getUsageByUser(filters?: UsageFilters & { limit?: number }): Promise<UsageByUser[]> {
return this.getUsageAggregation<UsageByUser[]>('user', filters)
},
async getUsageByProvider(filters?: UsageFilters & { limit?: number }): Promise<UsageByProvider[]> {
return this.getUsageAggregation<UsageByProvider[]>('provider', filters)
},
async getUsageByApiFormat(filters?: UsageFilters & { limit?: number }): Promise<UsageByApiFormat[]> {
return this.getUsageAggregation<UsageByApiFormat[]>('api_format', filters)
},
async getUserUsage(userId: string, filters?: UsageFilters): Promise<{
records: UsageRecord[]
stats: UsageStats
}> {
const response = await apiClient.get(`/api/users/${userId}/usage`, { params: filters })
return response.data
},
async exportUsage(format: 'csv' | 'json', filters?: UsageFilters): Promise<Blob> {
const response = await apiClient.get('/api/usage/export', {
params: { ...filters, format },
responseType: 'blob'
})
return response.data
},
async getAllUsageRecords(params?: {
start_date?: string
end_date?: string
preset?: string
granularity?: 'hour' | 'day' | 'week' | 'month'
timezone?: string
tz_offset_minutes?: number
search?: string // 通用搜索:用户名、密钥名、模型名、提供商名
user_id?: string // UUID
username?: string
model?: string
provider?: string
api_format?: string // API 格式筛选(如 openai:chat, claude:chat)
status?: string // 'stream' | 'standard' | 'error'
limit?: number
offset?: number
}): Promise<{
records: Array<Record<string, unknown>>
total: number
limit: number
offset: number
}> {
const key = buildCacheKey('usage:records', params as Record<string, unknown> | undefined)
return dedupedRequest(key, async () => {
const response = await apiClient.get('/api/admin/usage/records', { params })
return response.data
})
},
/**
* 获取活跃请求的状态(轻量级接口,用于轮询更新)
* @param ids 可选,逗号分隔的请求 ID 列表
*/
async getActiveRequests(ids?: string[]): Promise<{
requests: Array<{
id: string
status: 'pending' | 'streaming' | 'completed' | 'failed' | 'cancelled'
input_tokens: number
output_tokens: number
cache_creation_input_tokens?: number | null
cache_read_input_tokens?: number | null
cost: number
actual_cost?: number | null
rate_multiplier?: number | null
response_time_ms: number | null
first_byte_time_ms: number | null
provider?: string | null
api_key_name?: string | null
api_format?: string | null
endpoint_api_format?: string | null
has_format_conversion?: boolean | null
target_model?: string | null
}>
}> {
const params = ids?.length ? { ids: ids.join(',') } : {}
const response = await apiClient.get('/api/admin/usage/active', { params })
return response.data
},
/**
* 获取活跃度热力图数据(管理员)
* 后端已缓存5分钟
*/
async getActivityHeatmap(): Promise<ActivityHeatmap> {
return cachedRequest(
'admin-usage-activity-heatmap',
async () => {
const response = await apiClient.get<ActivityHeatmap>('/api/admin/usage/heatmap')
return response.data
},
60000
)
}
}
+46 -8
View File
@@ -10,6 +10,7 @@ export interface User {
allowed_providers: string[] | null // 允许使用的提供商 ID 列表
allowed_api_formats: string[] | null // 允许使用的 API 格式列表
allowed_models: string[] | null // 允许使用的模型名称列表
rate_limit?: number | null // null = 跟随系统默认,0 = 不限制
created_at: string
updated_at?: string
last_login_at?: string | null
@@ -25,6 +26,7 @@ export interface CreateUserRequest {
allowed_providers?: string[] | null
allowed_api_formats?: string[] | null
allowed_models?: string[] | null
rate_limit?: number | null
}
export interface UpdateUserRequest {
@@ -36,6 +38,7 @@ export interface UpdateUserRequest {
allowed_providers?: string[] | null
allowed_api_formats?: string[] | null
allowed_models?: string[] | null
rate_limit?: number | null
}
export interface ApiKey {
@@ -49,11 +52,18 @@ export interface ApiKey {
is_active: boolean
is_locked: boolean // 管理员锁定标志
is_standalone: boolean // 是否为独立余额Key
rate_limit?: number // 速率限制(请求/分钟)
rate_limit?: number | null // 普通Key: 0 = 不限制,历史 null 视为跟随系统默认
total_requests?: number // 总请求数
total_cost_usd?: number // 总费用
}
export interface UpsertUserApiKeyRequest {
name?: string
rate_limit?: number | null
}
export type { UserSession } from '@/types/session'
export const usersApi = {
async getAllUsers(): Promise<User[]> {
const response = await apiClient.get<User[]>('/api/admin/users')
@@ -84,8 +94,41 @@ export const usersApi = {
return response.data.api_keys
},
async createApiKey(userId: string, name?: string): Promise<ApiKey & { key: string }> {
const response = await apiClient.post<ApiKey & { key: string }>(`/api/admin/users/${userId}/api-keys`, { name })
async getUserSessions(userId: string): Promise<UserSession[]> {
const response = await apiClient.get<UserSession[]>(`/api/admin/users/${userId}/sessions`)
return response.data
},
async revokeUserSession(userId: string, sessionId: string): Promise<{ message: string }> {
const response = await apiClient.delete(`/api/admin/users/${userId}/sessions/${sessionId}`)
return response.data
},
async revokeAllUserSessions(userId: string): Promise<{ message: string; revoked_count: number }> {
const response = await apiClient.delete(`/api/admin/users/${userId}/sessions`)
return response.data
},
async createApiKey(
userId: string,
data: UpsertUserApiKeyRequest
): Promise<ApiKey & { key: string }> {
const response = await apiClient.post<ApiKey & { key: string }>(
`/api/admin/users/${userId}/api-keys`,
data
)
return response.data
},
async updateApiKey(
userId: string,
keyId: string,
data: UpsertUserApiKeyRequest
): Promise<ApiKey & { message: string }> {
const response = await apiClient.put<ApiKey & { message: string }>(
`/api/admin/users/${userId}/api-keys/${keyId}`,
data
)
return response.data
},
@@ -98,10 +141,5 @@ export const usersApi = {
`/api/admin/users/${userId}/api-keys/${keyId}/full-key`
)
return response.data
},
// 管理员统计
async getUsageStats(): Promise<Record<string, unknown>> {
const response = await apiClient.get('/api/admin/usage/stats')
return response.data
}
}
+30 -16
View File
@@ -1,5 +1,8 @@
<template>
<div class="w-full h-full">
<div
class="w-full"
:style="{ height: `${props.height}px` }"
>
<canvas ref="chartRef" />
</div>
</template>
@@ -18,6 +21,7 @@ import {
type ChartData,
type ChartOptions
} from 'chart.js'
import { observeChartThemeChanges, resolveChartTheme } from '@/utils/chartTheme'
const props = withDefaults(defineProps<Props>(), {
height: 300,
@@ -44,6 +48,7 @@ interface Props {
const chartRef = ref<HTMLCanvasElement>()
let chart: ChartJS<'bar'> | null = null
let stopThemeObserver: (() => void) | null = null
const defaultOptions: ChartOptions<'bar'> = {
responsive: true,
@@ -91,9 +96,7 @@ const defaultOptions: ChartOptions<'bar'> = {
}
}
function createChart() {
if (!chartRef.value) return
function buildChartOptions(): ChartOptions<'bar'> {
const stackedOptions = props.stacked ? {
scales: {
x: { ...defaultOptions.scales?.x, stacked: true },
@@ -106,20 +109,26 @@ function createChart() {
}
}
return resolveChartTheme({
...defaultOptions,
...stackedOptions,
...props.options
})
}
function createChart() {
if (!chartRef.value) return
chart = new ChartJS(chartRef.value, {
type: 'bar',
data: props.data,
options: {
...defaultOptions,
...stackedOptions,
...props.options
}
data: resolveChartTheme(props.data),
options: buildChartOptions()
})
}
function updateChart() {
if (chart) {
chart.data = props.data
chart.data = resolveChartTheme(props.data)
chart.update('none')
}
}
@@ -127,9 +136,17 @@ function updateChart() {
onMounted(async () => {
await nextTick()
createChart()
stopThemeObserver = observeChartThemeChanges(() => {
if (!chart) return
chart.data = resolveChartTheme(props.data)
chart.options = buildChartOptions()
chart.update('none')
})
})
onUnmounted(() => {
stopThemeObserver?.()
stopThemeObserver = null
if (chart) {
chart.destroy()
chart = null
@@ -139,11 +156,8 @@ onUnmounted(() => {
watch(() => props.data, updateChart, { deep: true })
watch(() => props.options, () => {
if (chart) {
chart.options = {
...defaultOptions,
...props.options
}
chart.update()
chart.options = buildChartOptions()
chart.update('none')
}
}, { deep: true })
</script>
@@ -1,5 +1,8 @@
<template>
<div class="w-full h-full">
<div
class="w-full"
:style="{ height: `${props.height}px` }"
>
<canvas ref="chartRef" />
</div>
</template>
@@ -16,10 +19,12 @@ import {
type ChartData,
type ChartOptions
} from 'chart.js'
import { observeChartThemeChanges, resolveChartTheme } from '@/utils/chartTheme'
const props = withDefaults(defineProps<Props>(), {
height: 300,
options: undefined
options: undefined,
showLegend: true,
})
ChartJS.register(
@@ -34,15 +39,22 @@ interface Props {
data: ChartData<'doughnut'>
options?: ChartOptions<'doughnut'>
height?: number
showLegend?: boolean
}
const chartRef = ref<HTMLCanvasElement>()
let chart: ChartJS<'doughnut'> | null = null
let stopThemeObserver: (() => void) | null = null
const defaultOptions: ChartOptions<'doughnut'> = {
responsive: true,
maintainAspectRatio: false,
cutout: '60%',
events: ['mousemove', 'mouseout', 'click', 'touchstart', 'touchmove'],
interaction: {
mode: 'nearest',
intersect: true,
},
plugins: {
legend: {
position: 'right',
@@ -71,22 +83,59 @@ const defaultOptions: ChartOptions<'doughnut'> = {
}
}
function resolvedOptions(): ChartOptions<'doughnut'> {
const incomingOptions = props.options ?? {}
const incomingPlugins = (incomingOptions.plugins ?? {}) as Record<string, unknown>
const incomingLegend = (incomingPlugins.legend ?? {}) as Record<string, unknown>
const incomingTooltip = (incomingPlugins.tooltip ?? {}) as Record<string, unknown>
const defaultPlugins = (defaultOptions.plugins ?? {}) as Record<string, unknown>
const defaultLegend = (defaultPlugins.legend ?? {}) as Record<string, unknown>
const defaultTooltip = (defaultPlugins.tooltip ?? {}) as Record<string, unknown>
return resolveChartTheme({
...defaultOptions,
...incomingOptions,
interaction: {
...defaultOptions.interaction,
...incomingOptions.interaction,
},
plugins: {
...defaultPlugins,
...incomingPlugins,
legend: {
...defaultLegend,
...incomingLegend,
display: incomingLegend.display ?? props.showLegend,
labels: {
...(defaultLegend.labels as Record<string, unknown> | undefined),
...(incomingLegend.labels as Record<string, unknown> | undefined),
},
},
tooltip: {
...defaultTooltip,
...incomingTooltip,
callbacks: {
...(defaultTooltip.callbacks as Record<string, unknown> | undefined),
...(incomingTooltip.callbacks as Record<string, unknown> | undefined),
},
},
},
})
}
function createChart() {
if (!chartRef.value) return
chart = new ChartJS(chartRef.value, {
type: 'doughnut',
data: props.data,
options: {
...defaultOptions,
...props.options
}
data: resolveChartTheme(props.data),
options: resolvedOptions(),
})
}
function updateChart() {
if (chart) {
chart.data = props.data
chart.data = resolveChartTheme(props.data)
chart.update('none')
}
}
@@ -94,9 +143,17 @@ function updateChart() {
onMounted(async () => {
await nextTick()
createChart()
stopThemeObserver = observeChartThemeChanges(() => {
if (!chart) return
chart.data = resolveChartTheme(props.data)
chart.options = resolvedOptions()
chart.update('none')
})
})
onUnmounted(() => {
stopThemeObserver?.()
stopThemeObserver = null
if (chart) {
chart.destroy()
chart = null
@@ -106,8 +163,15 @@ onUnmounted(() => {
watch(() => props.data, updateChart, { deep: true })
watch(() => props.options, () => {
if (chart) {
chart.options = { ...defaultOptions, ...props.options }
chart.update()
chart.options = resolvedOptions()
chart.update('none')
}
}, { deep: true })
watch(() => props.showLegend, () => {
if (chart) {
chart.options = resolvedOptions()
chart.update('none')
}
})
</script>
+25 -6
View File
@@ -1,5 +1,8 @@
<template>
<div class="w-full h-full">
<div
class="w-full"
:style="{ height: `${props.height}px` }"
>
<canvas ref="chartRef" />
</div>
</template>
@@ -16,9 +19,11 @@ import {
Title,
Tooltip,
Legend,
Filler,
type ChartData,
type ChartOptions
} from 'chart.js'
import { observeChartThemeChanges, resolveChartTheme } from '@/utils/chartTheme'
const props = withDefaults(defineProps<Props>(), {
height: 300,
@@ -34,7 +39,8 @@ ChartJS.register(
LineController,
Title,
Tooltip,
Legend
Legend,
Filler
)
interface Props {
@@ -45,12 +51,17 @@ interface Props {
const chartRef = ref<HTMLCanvasElement>()
let chart: ChartJS<'line'> | null = null
let stopThemeObserver: (() => void) | null = null
function buildChartOptions(): ChartOptions<'line'> {
return {
return resolveChartTheme({
...defaultOptions,
...props.options
}
})
}
function buildChartData(): ChartData<'line'> {
return resolveChartTheme(props.data)
}
const defaultOptions: ChartOptions<'line'> = {
@@ -95,14 +106,14 @@ function createChart() {
chart = new ChartJS(chartRef.value, {
type: 'line',
data: props.data,
data: buildChartData(),
options: buildChartOptions()
})
}
function updateChart() {
if (chart) {
chart.data = props.data
chart.data = buildChartData()
chart.update('none') // 禁用动画以提高性能
}
}
@@ -110,9 +121,17 @@ function updateChart() {
onMounted(async () => {
await nextTick()
createChart()
stopThemeObserver = observeChartThemeChanges(() => {
if (!chart) return
chart.data = buildChartData()
chart.options = buildChartOptions()
chart.update('none')
})
})
onUnmounted(() => {
stopThemeObserver?.()
stopThemeObserver = null
if (chart) {
chart.destroy()
chart = null
@@ -4,7 +4,7 @@
ref="fileInputRef"
type="file"
:accept="accept"
multiple
:multiple="multiple"
class="hidden"
@change="handleFileSelect"
>
@@ -94,6 +94,7 @@ const props = withDefaults(defineProps<{
disabled?: boolean
resetKey?: string | number
accept?: string
multiple?: boolean
dropTitle?: string
dropHint?: string
manualLabel?: string
@@ -106,6 +107,7 @@ const props = withDefaults(defineProps<{
disabled: false,
resetKey: '',
accept: '.json,.txt',
multiple: true,
dropTitle: '拖入导入文件或点击选择',
dropHint: '支持 .json / .txt,可多选',
manualLabel: '',
@@ -218,13 +220,19 @@ function mergeFileContents(contents: string[]): string {
}
async function readFiles(files: File[]) {
const validFiles = files.filter(isValidFileType)
const sourceFiles = props.multiple ? files : files.slice(0, 1)
if (!props.multiple && files.length > 1) {
emitError('仅支持选择 1 个文件,已读取第一个文件', '提示')
}
const validFiles = sourceFiles.filter(isValidFileType)
if (validFiles.length === 0) {
emitError('仅支持 .json 或 .txt 文件', '格式错误')
return
}
if (validFiles.length < files.length) {
emitError(`已忽略 ${files.length - validFiles.length} 个不支持的文件`, '提示')
if (validFiles.length < sourceFiles.length) {
emitError(`已忽略 ${sourceFiles.length - validFiles.length} 个不支持的文件`, '提示')
}
try {
@@ -3,39 +3,24 @@
<Select
v-model="selectedPreset"
>
<SelectTrigger class="h-8 w-32 text-xs border-border/60">
<SelectTrigger :class="presetTriggerClass">
<SelectValue placeholder="选择时间段" />
</SelectTrigger>
<SelectContent>
<SelectItem value="today">
今天
</SelectItem>
<SelectItem value="yesterday">
昨天
</SelectItem>
<SelectItem value="last7days">
最近7天
</SelectItem>
<SelectItem value="last30days">
最近30天
</SelectItem>
<SelectItem value="last90days">
最近90天
<SelectItem value="last180days">
最近180天
</SelectItem>
<SelectItem value="this_week">
本周
</SelectItem>
<SelectItem value="last_week">
上周
</SelectItem>
<SelectItem value="this_month">
本月
</SelectItem>
<SelectItem value="last_month">
上月
</SelectItem>
<SelectItem value="this_year">
今年
<SelectItem value="last1year">
最近一年
</SelectItem>
<SelectItem value="custom">
自定义
@@ -50,13 +35,13 @@
<Input
v-model="startDate"
type="date"
class="h-8 w-36 text-xs border-border/60"
:class="dateInputClass"
/>
<span class="text-xs text-muted-foreground">至</span>
<Input
v-model="endDate"
type="date"
class="h-8 w-36 text-xs border-border/60"
:class="dateInputClass"
/>
</div>
@@ -64,10 +49,16 @@
v-if="showGranularity"
v-model="selectedGranularity"
>
<SelectTrigger class="h-8 w-24 text-xs border-border/60">
<SelectTrigger :class="granularityTriggerClass">
<SelectValue placeholder="粒度" />
</SelectTrigger>
<SelectContent>
<SelectItem
v-if="includeAutoGranularity"
value="auto"
>
自动
</SelectItem>
<SelectItem
v-if="allowHourly && canUseHourly"
value="hour"
@@ -100,26 +91,51 @@ import {
} from '@/components/ui'
import type { DateRangeParams } from '@/features/usage/types'
type PickerGranularity = NonNullable<DateRangeParams['granularity']>
const props = defineProps<{
modelValue: DateRangeParams
showGranularity?: boolean
allowHourly?: boolean
includeAutoGranularity?: boolean
compact?: boolean
}>()
const emit = defineEmits<{
'update:modelValue': [value: DateRangeParams]
}>()
const selectedPreset = ref(props.modelValue.preset || 'last7days')
function getDefaultGranularity(): PickerGranularity {
return props.includeAutoGranularity ? 'auto' : 'day'
}
const selectedPreset = ref(props.modelValue.preset || 'today')
const startDate = ref(props.modelValue.start_date || '')
const endDate = ref(props.modelValue.end_date || '')
const selectedGranularity = ref(props.modelValue.granularity || 'day')
const selectedGranularity = ref<PickerGranularity>(props.modelValue.granularity || getDefaultGranularity())
const showGranularity = computed(() => props.showGranularity !== false)
const allowHourly = computed(() => props.allowHourly === true)
const includeAutoGranularity = computed(() => props.includeAutoGranularity === true)
const compact = computed(() => props.compact === true)
const presetTriggerClass = computed(() =>
compact.value
? 'h-8 w-28 text-xs border-border/60'
: 'h-8 w-32 text-xs border-border/60',
)
const dateInputClass = computed(() =>
compact.value
? 'h-8 w-32 text-xs border-border/60'
: 'h-8 w-36 text-xs border-border/60',
)
const granularityTriggerClass = computed(() =>
compact.value
? 'h-8 w-20 text-xs border-border/60'
: 'h-8 w-24 text-xs border-border/60',
)
const canUseHourly = computed(() => {
if (selectedPreset.value === 'today' || selectedPreset.value === 'yesterday') return true
if (selectedPreset.value === 'today') return true
if (selectedPreset.value === 'custom' && startDate.value && endDate.value) {
return startDate.value === endDate.value
}
@@ -165,7 +181,7 @@ watch(() => props.modelValue, (value) => {
if (value.preset) selectedPreset.value = value.preset
if (value.start_date !== undefined) startDate.value = value.start_date || ''
if (value.end_date !== undefined) endDate.value = value.end_date || ''
if (value.granularity) selectedGranularity.value = value.granularity
selectedGranularity.value = value.granularity || getDefaultGranularity()
// 同步更新 lastEmittedValue,避免外部设置值后触发重复 emit
lastEmittedValue = getValueKey(value)
}, { deep: true })
@@ -173,7 +189,7 @@ watch(() => props.modelValue, (value) => {
watch([selectedPreset, startDate, endDate, selectedGranularity], () => {
if (!allowHourly.value || !canUseHourly.value) {
if (selectedGranularity.value === 'hour') {
selectedGranularity.value = 'day'
selectedGranularity.value = includeAutoGranularity.value ? 'auto' : 'day'
}
}
@@ -13,8 +13,11 @@
<p class="mt-0.5">
{{ tooltip.day.requests }} 次请求 · {{ formatTokens(tooltip.day.total_tokens) }}
</p>
<p class="text-[11px] text-muted-foreground">
成本 {{ formatCurrency(tooltip.day.total_cost) }}
<p class="text-[11px] text-muted-foreground whitespace-nowrap">
{{ valueLabel }} {{ formatCurrency(tooltip.day.total_cost) }}
<template v-if="actualValueLabel && tooltip.day.actual_total_cost !== undefined">
· {{ actualValueLabel }} {{ formatCurrency(tooltip.day.actual_total_cost) }}
</template>
</p>
</div>
</Teleport>
@@ -90,7 +93,7 @@
class="flex items-center invisible"
>周六</span>
</div>
<div class="flex-1 min-w-[200px]">
<div class="flex-1 min-w-0">
<div
ref="heatmapWrapper"
class="relative block w-full"
@@ -100,12 +103,12 @@
:style="horizontalGapStyle"
>
<div
v-for="(week, weekIndex) in weekColumns"
v-for="(week, weekIndex) in visibleWeekColumns"
:key="`month-${weekIndex}`"
:style="monthCellStyle"
class="text-center"
>
<span v-if="monthMarkers[weekIndex]">{{ monthMarkers[weekIndex] }}</span>
<span v-if="visibleMonthMarkers[weekIndex]">{{ visibleMonthMarkers[weekIndex] }}</span>
</div>
</div>
<div
@@ -113,7 +116,7 @@
:style="horizontalGapStyle"
>
<div
v-for="(week, weekIndex) in weekColumns"
v-for="(week, weekIndex) in visibleWeekColumns"
:key="weekIndex"
class="flex flex-col"
:style="verticalGapStyle"
@@ -161,11 +164,15 @@ const props = withDefaults(defineProps<{
title?: string
subtitle?: string
showHeader?: boolean
valueLabel?: string
actualValueLabel?: string
}>(), {
data: undefined,
title: undefined,
subtitle: undefined,
showHeader: true
showHeader: true,
valueLabel: '费用',
actualValueLabel: undefined,
})
const legendLevels = [0.08, 0.25, 0.45, 0.65, 0.85]
@@ -270,7 +277,47 @@ const monthMarkers = computed(() => {
if (month === lastMonth) {
return
}
markers[index] = `${month + 1}月`
markers[index] = String(month + 1)
lastMonth = month
})
return markers
})
const MIN_CELL_SIZE = 6
// 根据容器宽度计算最多能放多少周,然后从后面截取(保留最新数据)
const maxVisibleWeeks = computed(() => {
const total = weekColumns.value.length
if (!heatmapWidth.value || !total) return total
// 先尝试用理想尺寸放全部
const totalGapAll = Math.max(total - 1, 0) * cellGap.value
const rawSizeAll = (heatmapWidth.value - totalGapAll) / total
if (rawSizeAll >= MIN_CELL_SIZE) return total
// 放不下,算容器最多能容纳几周(用最小尺寸)
// n * MIN_CELL_SIZE + (n-1) * gap <= width
const maxN = Math.floor((heatmapWidth.value + cellGap.value) / (MIN_CELL_SIZE + cellGap.value))
return Math.max(1, Math.min(maxN, total))
})
const visibleWeekColumns = computed(() => {
const all = weekColumns.value
const n = maxVisibleWeeks.value
if (n >= all.length) return all
return all.slice(all.length - n)
})
const visibleMonthMarkers = computed(() => {
const markers: Record<number, string> = {}
const columns = visibleWeekColumns.value
let lastMonth: number | null = null
columns.forEach((week, index) => {
const firstValid = week.find((day): day is DayWithMeta => day !== null)
if (!firstValid) return
const month = firstValid.dateObj.getUTCMonth()
if (month === lastMonth) return
markers[index] = String(month + 1)
lastMonth = month
})
@@ -282,7 +329,7 @@ let mediaQuery: MediaQueryList | null = null
let mediaQueryHandler: ((event?: MediaQueryListEvent) => void) | null = null
const recalcCellSize = () => {
const columnCount = weekColumns.value.length
const columnCount = visibleWeekColumns.value.length
if (!columnCount || !heatmapWidth.value) {
return
}
@@ -290,12 +337,12 @@ const recalcCellSize = () => {
const totalGap = Math.max(columnCount - 1, 0) * cellGap.value
const availableSpace = Math.max(heatmapWidth.value - totalGap, 0)
const rawSize = availableSpace / columnCount
// 自适应尺寸,最小 6px
cellSize.value = Math.max(6, rawSize)
// 当所有列都能放下时,让格子继续随容器放大,避免宽屏下出现大块留白。
cellSize.value = Math.max(MIN_CELL_SIZE, rawSize)
}
watch(
[() => heatmapWidth.value, () => weekColumns.value.length, () => cellGap.value],
[() => heatmapWidth.value, () => visibleWeekColumns.value.length, () => cellGap.value],
() => {
recalcCellSize()
},
@@ -345,7 +392,7 @@ onBeforeUnmount(() => {
function handleHover(day: ActivityHeatmapDay, event: MouseEvent) {
const cellRect = (event.currentTarget as HTMLElement).getBoundingClientRect()
const tooltipWidth = 200
const tooltipWidth = 280
const tooltipHeight = 72
// Calculate horizontal position (centered on cell)
@@ -402,10 +449,10 @@ function getCellStyle(requests: number) {
function buildTooltip(day: ActivityHeatmapDay): string {
const dateLabel = day.date
const costLabel = formatCurrency(day.total_cost || 0)
const costLabel = `${props.valueLabel} ${formatCurrency(day.total_cost || 0)}`
const parts = [`${dateLabel}`, `${day.requests} 次请求`, `${formatTokens(day.total_tokens)} tokens`, costLabel]
if (day.actual_total_cost !== undefined) {
parts.push(`倍率: ${formatCurrency(day.actual_total_cost)}`)
if (props.actualValueLabel && day.actual_total_cost !== undefined) {
parts.push(`${props.actualValueLabel} ${formatCurrency(day.actual_total_cost)}`)
}
return parts.join(' · ')
}
@@ -1,38 +0,0 @@
<template>
<Card class="p-4 space-y-2">
<div class="text-xs text-muted-foreground">
{{ label }}
</div>
<div class="text-lg font-semibold">
{{ value }}
</div>
<div
class="text-xs"
:class="changeClass"
>
<span v-if="changePercent !== null">{{ changePercent }}%</span>
<span v-else>--</span>
<span class="ml-1 text-muted-foreground">vs 对比期</span>
</div>
</Card>
</template>
<script setup lang="ts">
import { computed } from 'vue'
import { Card } from '@/components/ui'
interface Props {
label: string
value: string
changePercent: number | null
}
const props = defineProps<Props>()
const changeClass = computed(() => {
if (props.changePercent === null) return 'text-muted-foreground'
if (props.changePercent > 0) return 'text-emerald-500'
if (props.changePercent < 0) return 'text-rose-500'
return 'text-muted-foreground'
})
</script>
@@ -1,93 +0,0 @@
<template>
<div class="space-y-3">
<div class="flex items-center justify-between">
<h3 class="text-sm font-semibold">
{{ title }}
</h3>
<span
v-if="subtitle"
class="text-xs text-muted-foreground"
>{{ subtitle }}</span>
</div>
<div
v-if="loading"
class="p-6"
>
<LoadingState />
</div>
<div
v-else
class="h-[280px]"
>
<LineChart
:data="chartData"
:options="chartOptions"
/>
</div>
</div>
</template>
<script setup lang="ts">
import { computed } from 'vue'
import LineChart from '@/components/charts/LineChart.vue'
import { LoadingState } from '@/components/common'
import { formatCurrency } from '@/utils/format'
interface Props {
title: string
subtitle?: string
history: Array<{ date: string; total_cost: number }>
forecast: Array<{ date: string; total_cost: number }>
loading?: boolean
}
const props = withDefaults(defineProps<Props>(), {
subtitle: undefined,
loading: false
})
const labels = computed(() => [
...props.history.map(item => item.date),
...props.forecast.map(item => item.date)
])
const chartData = computed(() => {
const historyValues = props.history.map(item => item.total_cost)
const forecastValues = props.forecast.map(item => item.total_cost)
return {
labels: labels.value,
datasets: [
{
label: '实际成本',
data: historyValues.concat(new Array(forecastValues.length).fill(null)),
borderColor: 'rgb(59, 130, 246)',
backgroundColor: 'rgba(59, 130, 246, 0.15)',
tension: 0.25,
pointRadius: 2
},
{
label: '预测成本',
data: new Array(historyValues.length).fill(null).concat(forecastValues),
borderColor: 'rgb(234, 179, 8)',
backgroundColor: 'rgba(234, 179, 8, 0.15)',
borderDash: [6, 4],
tension: 0.25,
pointRadius: 2
}
]
}
})
const chartOptions = computed(() => ({
plugins: {
tooltip: {
callbacks: {
label: (context: { parsed?: { y?: number }; dataset: { label?: string } }) => {
const value = context.parsed?.y ?? 0
return `${context.dataset.label}: ${formatCurrency(value)}`
}
}
}
}
}))
</script>
@@ -1,14 +1,34 @@
<template>
<div class="space-y-3">
<div class="flex items-center justify-between">
<h3 class="text-sm font-semibold">
{{ title }}
</h3>
<span
v-if="subtitle"
class="text-xs text-muted-foreground"
>{{ subtitle }}</span>
<div class="space-y-4">
<div class="flex flex-col gap-3 sm:flex-row sm:items-start sm:justify-between">
<div>
<h3 class="text-sm font-semibold">
{{ title }}
</h3>
<p
v-if="subtitle"
class="mt-1 text-[11px] text-muted-foreground"
>
{{ subtitle }}
</p>
</div>
<div
v-if="!unavailable"
class="flex flex-wrap gap-2"
>
<div class="rounded-full border border-rose-500/20 bg-rose-500/8 px-2.5 py-1 text-[10px] font-medium text-rose-700 dark:text-rose-300">
总错误 {{ totalErrors }}
</div>
<div
v-if="topCategory"
class="rounded-full border border-border/70 bg-muted/20 px-2.5 py-1 text-[10px] font-medium text-muted-foreground"
>
TOP {{ topCategory.label }} · {{ topCategory.count }}
</div>
</div>
</div>
<div
v-if="loading"
class="p-6"
@@ -16,57 +36,116 @@
<LoadingState />
</div>
<div
v-else
class="h-[260px]"
v-else-if="unavailable"
class="flex h-[272px] items-center justify-center rounded-2xl border border-dashed border-border/70 bg-muted/10 text-xs text-muted-foreground"
>
<DoughnutChart
:data="chartData"
:options="chartOptions"
/>
{{ unavailableText }}
</div>
<div
v-else-if="!hasData"
class="flex h-[272px] items-center justify-center rounded-2xl border border-dashed border-border/70 bg-muted/10 text-xs text-muted-foreground"
>
暂无错误分布数据
</div>
<div
v-else
class="rounded-2xl border border-border/60 bg-background p-3"
>
<div class="mx-auto w-full max-w-[320px]">
<DoughnutChart
:data="chartData"
:options="chartOptions"
:show-legend="true"
:height="248"
/>
</div>
</div>
</div>
</template>
<script setup lang="ts">
import { computed } from 'vue'
import type { ChartData, ChartOptions } from 'chart.js'
import DoughnutChart from '@/components/charts/DoughnutChart.vue'
import { LoadingState } from '@/components/common'
import type { ErrorDistributionItem } from '@/api/admin'
interface ErrorDistributionItem {
category: string
label: string
count: number
}
interface Props {
title: string
subtitle?: string
distribution: ErrorDistributionItem[]
loading?: boolean
unavailable?: boolean
unavailableText?: string
}
const props = withDefaults(defineProps<Props>(), {
subtitle: undefined,
loading: false
loading: false,
unavailable: false,
unavailableText: '数据暂不可用',
})
const chartData = computed(() => ({
labels: props.distribution.map(item => item.category),
const chartColors = [
'rgba(226, 85, 118, 0.88)',
'rgba(84, 135, 237, 0.86)',
'rgba(225, 165, 52, 0.84)',
'rgba(68, 187, 132, 0.82)',
'rgba(148, 163, 184, 0.8)',
'rgba(145, 118, 214, 0.82)',
]
const totalErrors = computed(() => props.distribution.reduce((sum, item) => sum + item.count, 0))
const hasData = computed(() => totalErrors.value > 0)
const distributionWithShare = computed(() => props.distribution.map(item => ({
...item,
label: item.label?.trim() || item.category,
share: totalErrors.value > 0 ? `${((item.count / totalErrors.value) * 100).toFixed(1)}%` : '0.0%',
})))
const topCategory = computed(() => (
distributionWithShare.value.slice().sort((left, right) => right.count - left.count)[0] ?? null
))
const chartData = computed<ChartData<'doughnut'>>(() => ({
labels: distributionWithShare.value.map(item => item.label),
datasets: [
{
data: props.distribution.map(item => item.count),
backgroundColor: [
'rgba(239, 68, 68, 0.7)',
'rgba(59, 130, 246, 0.7)',
'rgba(234, 179, 8, 0.7)',
'rgba(34, 197, 94, 0.7)',
'rgba(148, 163, 184, 0.7)'
],
borderWidth: 0
}
]
data: distributionWithShare.value.map(item => item.count),
backgroundColor: distributionWithShare.value.map((_, index) => chartColors[index % chartColors.length]),
borderColor: 'rgba(255,255,255,0.92)',
borderWidth: 2,
hoverOffset: 6,
},
],
}))
const chartOptions = computed(() => ({
const chartOptions = computed<ChartOptions<'doughnut'>>(() => ({
cutout: '68%',
plugins: {
legend: {
position: 'bottom' as const
}
}
position: 'bottom',
align: 'center',
labels: {
usePointStyle: true,
pointStyle: 'circle',
boxWidth: 10,
padding: 14,
},
},
tooltip: {
callbacks: {
label: (context) => {
const value = Number(context.raw ?? 0)
const percentage = totalErrors.value > 0 ? ((value / totalErrors.value) * 100).toFixed(1) : '0.0'
return `${context.label ?? '错误'}: ${value} (${percentage}%)`
},
},
},
},
}))
</script>
@@ -1,34 +1,30 @@
<template>
<TableCard :title="title">
<template #actions>
<Select
v-if="showMetricSelect"
:model-value="metric"
@update:model-value="emitMetric"
<div
v-if="isRefreshing"
class="inline-flex h-7 items-center rounded-lg border border-border/70 bg-background px-2.5 text-[10px] text-muted-foreground"
>
<SelectTrigger class="h-8 text-xs w-28">
<SelectValue placeholder="指标" />
</SelectTrigger>
<SelectContent>
<SelectItem value="requests">
请求数
</SelectItem>
<SelectItem value="tokens">
Tokens
</SelectItem>
<SelectItem value="cost">
成本
</SelectItem>
</SelectContent>
</Select>
更新中
</div>
</template>
<div
v-if="loading"
v-if="showLoadingState"
class="p-6"
>
<LoadingState />
</div>
<div
v-else-if="showUnavailableState"
class="p-6"
>
<EmptyState
type="error"
:title="unavailableTitle"
:description="unavailableDescription"
/>
</div>
<div
v-else-if="items.length === 0"
class="p-6"
@@ -38,20 +34,29 @@
description="当前时间范围内没有统计结果"
/>
</div>
<Table v-else>
<Table
v-else
class="table-fixed"
:class="{ 'opacity-60 transition-opacity': isRefreshing }"
>
<TableHeader>
<TableRow>
<TableHead class="w-16">
<TableHead class="w-14 px-2.5">
排名
</TableHead>
<TableHead>名称</TableHead>
<TableHead class="text-right">
<TableHead class="w-[34%] px-2.5">
名称
</TableHead>
<TableHead class="w-[15%] px-2.5 text-right">
请求数
</TableHead>
<TableHead class="text-right">
<TableHead class="w-[18%] px-2.5 text-right">
Tokens
</TableHead>
<TableHead class="text-right">
<TableHead class="w-[16%] px-2.5 text-right">
费用
</TableHead>
<TableHead class="w-[17%] px-2.5 text-right">
成本
</TableHead>
</TableRow>
@@ -60,68 +65,118 @@
<TableRow
v-for="item in items"
:key="item.id"
class="cursor-pointer"
:class="item.id === selectedId
? 'bg-[linear-gradient(90deg,rgba(232,145,89,0.16),rgba(232,145,89,0.05),transparent)] shadow-[inset_4px_0_0_0_rgba(232,145,89,0.9)] hover:bg-[linear-gradient(90deg,rgba(232,145,89,0.2),rgba(232,145,89,0.06),transparent)]'
: 'hover:bg-muted/25'"
@click="emit('select', item.id)"
>
<TableCell class="font-medium">
{{ item.rank }}
<TableCell class="px-2.5 py-2.5">
<div
class="inline-flex h-7 min-w-7 items-center justify-center rounded-full border px-2 text-[11px] font-semibold tabular-nums transition-colors"
:class="item.id === selectedId
? 'border-amber-500/70 bg-amber-500/15 text-amber-700 dark:text-amber-300'
: 'border-border/70 bg-muted/35 text-muted-foreground'"
>
{{ item.rank }}
</div>
</TableCell>
<TableCell>{{ item.name }}</TableCell>
<TableCell class="text-right">
{{ item.requests }}
<TableCell class="px-2.5 py-2.5">
<div class="min-w-0">
<div
class="truncate text-sm font-medium"
:title="item.name"
>
{{ item.name }}
</div>
<div class="mt-0.5 truncate text-[10px] text-muted-foreground">
费用优先
</div>
</div>
</TableCell>
<TableCell class="text-right">
<TableCell class="px-2.5 py-2.5 text-right text-[11px] tabular-nums">
{{ item.requests.toLocaleString() }}
</TableCell>
<TableCell class="px-2.5 py-2.5 text-right text-[11px] tabular-nums">
{{ formatTokens(item.tokens) }}
</TableCell>
<TableCell class="text-right">
<TableCell class="px-2.5 py-2.5 text-right text-[11px] font-medium tabular-nums text-amber-700 dark:text-amber-300">
{{ formatCurrency(item.cost) }}
</TableCell>
<TableCell class="px-2.5 py-2.5 text-right text-[11px] tabular-nums text-muted-foreground">
{{ formatCurrency(item.actualCost) }}
</TableCell>
</TableRow>
</TableBody>
</Table>
<template v-if="showPagination" #pagination>
<Pagination
:current="currentPage"
:total="totalItems"
:page-size="pageSize"
:show-page-size-selector="false"
@update:current="emit('update:currentPage', $event)"
/>
</template>
</TableCard>
</template>
<script setup lang="ts">
import { computed } from 'vue'
import { EmptyState, LoadingState } from '@/components/common'
import { TableCard } from '@/components/ui'
import {
Select,
SelectContent,
SelectItem,
SelectTrigger,
SelectValue,
Table,
TableBody,
TableCell,
TableHead,
TableHeader,
TableRow
} from '@/components/ui'
import { Pagination, Table, TableBody, TableCell, TableHead, TableHeader, TableRow, TableCard } from '@/components/ui'
import { formatCurrency, formatTokens } from '@/utils/format'
import type { LeaderboardItem } from '@/api/admin'
interface LeaderboardItem {
rank: number
id: string
name: string
requests: number
tokens: number
cost: number
actualCost: number
}
interface Props {
title: string
items: LeaderboardItem[]
metric: 'requests' | 'tokens' | 'cost'
selectedId?: string
loading?: boolean
showMetricSelect?: boolean
hasLoaded?: boolean
currentPage?: number
totalItems?: number
pageSize?: number
showPagination?: boolean
unavailable?: boolean
unavailableTitle?: string
unavailableDescription?: string
}
const props = withDefaults(defineProps<Props>(), {
selectedId: undefined,
loading: false,
showMetricSelect: true
hasLoaded: false,
currentPage: 1,
totalItems: 0,
pageSize: 5,
showPagination: false,
unavailable: false,
unavailableTitle: '排行榜暂不可用',
unavailableDescription: '接口未返回结果,请稍后重试',
})
const emit = defineEmits<{
(e: 'update:metric', value: 'requests' | 'tokens' | 'cost'): void
(e: 'select', value: string): void
(e: 'update:currentPage', value: number): void
}>()
const metric = computed(() => props.metric)
function emitMetric(value: string) {
if (value === 'requests' || value === 'tokens' || value === 'cost') {
emit('update:metric', value)
}
}
const selectedId = computed(() => props.selectedId)
const showLoadingState = computed(() => props.loading && (!props.hasLoaded || props.items.length === 0))
const showUnavailableState = computed(() => props.unavailable && !props.loading && props.items.length === 0)
const isRefreshing = computed(() => props.loading && props.hasLoaded && props.items.length > 0)
const currentPage = computed(() => props.currentPage)
const totalItems = computed(() => props.totalItems)
const pageSize = computed(() => props.pageSize)
const showPagination = computed(() => props.showPagination && props.totalItems > props.pageSize)
</script>
+193 -53
View File
@@ -1,27 +1,59 @@
<template>
<div class="space-y-3">
<div class="flex items-center justify-between">
<h3 class="text-sm font-semibold">
{{ title }}
</h3>
<span
v-if="subtitle"
class="text-xs text-muted-foreground"
>{{ subtitle }}</span>
<div class="space-y-4">
<div class="flex flex-col gap-3 sm:flex-row sm:items-start sm:justify-between">
<div>
<h3 class="text-sm font-semibold">
{{ title }}
</h3>
<p
v-if="subtitle"
class="mt-1 text-[11px] text-muted-foreground"
>
{{ subtitle }}
</p>
</div>
<div
v-if="summaryMetrics.length > 0 && !unavailable"
class="flex flex-wrap gap-2"
>
<div
v-for="metric in summaryMetrics"
:key="metric.label"
class="rounded-full border px-2.5 py-1 text-[10px] font-medium"
:class="metric.className"
>
{{ metric.label }} {{ metric.value }}
</div>
</div>
</div>
<div
v-if="loading"
class="p-6"
>
<LoadingState />
</div>
<div
v-else-if="unavailable"
class="flex h-[272px] items-center justify-center rounded-2xl border border-dashed border-border/70 bg-muted/10 text-xs text-muted-foreground"
>
{{ unavailableText }}
</div>
<div
v-else-if="!hasData"
class="flex h-[272px] items-center justify-center rounded-2xl border border-dashed border-border/70 bg-muted/10 text-xs text-muted-foreground"
>
暂无延迟数据
</div>
<div
v-else
class="h-[260px]"
class="rounded-2xl border border-border/60 bg-background p-3"
>
<LineChart
:data="chartData"
:options="chartOptions"
:height="248"
/>
</div>
</div>
@@ -29,9 +61,19 @@
<script setup lang="ts">
import { computed } from 'vue'
import type { ChartData, ChartOptions } from 'chart.js'
import LineChart from '@/components/charts/LineChart.vue'
import { LoadingState } from '@/components/common'
import type { PercentileItem } from '@/api/admin'
interface PercentileItem {
date: string
p50_response_time_ms?: number | null
p90_response_time_ms?: number | null
p99_response_time_ms?: number | null
p50_first_byte_time_ms?: number | null
p90_first_byte_time_ms?: number | null
p99_first_byte_time_ms?: number | null
}
interface Props {
title: string
@@ -39,61 +81,159 @@ interface Props {
series: PercentileItem[]
mode: 'response' | 'ttfb'
loading?: boolean
unavailable?: boolean
unavailableText?: string
}
const props = withDefaults(defineProps<Props>(), {
subtitle: undefined,
loading: false
loading: false,
unavailable: false,
unavailableText: '数据暂不可用',
})
const labels = computed(() => props.series.map(item => item.date))
// 毫秒转秒
function msToSeconds(ms: number | null | undefined): number | null {
if (ms == null) return null
return ms / 1000
return Number((ms / 1000).toFixed(3))
}
const chartData = computed(() => {
const p50Key = props.mode === 'response' ? 'p50_response_time_ms' : 'p50_first_byte_time_ms'
const p90Key = props.mode === 'response' ? 'p90_response_time_ms' : 'p90_first_byte_time_ms'
const p99Key = props.mode === 'response' ? 'p99_response_time_ms' : 'p99_first_byte_time_ms'
return {
labels: labels.value,
datasets: [
{
label: 'P50',
data: props.series.map(item => msToSeconds(item[p50Key])),
borderColor: 'rgb(59, 130, 246)',
tension: 0.25,
pointRadius: 2
},
{
label: 'P90',
data: props.series.map(item => msToSeconds(item[p90Key])),
borderColor: 'rgb(234, 179, 8)',
tension: 0.25,
pointRadius: 2
},
{
label: 'P99',
data: props.series.map(item => msToSeconds(item[p99Key])),
borderColor: 'rgb(239, 68, 68)',
tension: 0.25,
pointRadius: 2
}
]
function getPercentileValue(item: PercentileItem, percentile: '50' | '90' | '99') {
if (props.mode === 'response') {
if (percentile === '50') return msToSeconds(item.p50_response_time_ms)
if (percentile === '90') return msToSeconds(item.p90_response_time_ms)
return msToSeconds(item.p99_response_time_ms)
}
})
const chartOptions = computed(() => ({
if (percentile === '50') return msToSeconds(item.p50_first_byte_time_ms)
if (percentile === '90') return msToSeconds(item.p90_first_byte_time_ms)
return msToSeconds(item.p99_first_byte_time_ms)
}
function getLatestMetricValue(percentile: '50' | '90' | '99') {
for (let index = props.series.length - 1; index >= 0; index -= 1) {
const value = getPercentileValue(props.series[index], percentile)
if (value != null) return value
}
return null
}
function formatSeconds(value: number | null) {
if (value == null) return '--'
if (value < 1) return `${Math.round(value * 1000)}ms`
return `${value.toFixed(2)}s`
}
const labels = computed(() => props.series.map(item => item.date?.slice(5) ?? ''))
const hasData = computed(() => props.series.some(item => (
getPercentileValue(item, '50') != null
|| getPercentileValue(item, '90') != null
|| getPercentileValue(item, '99') != null
)))
const summaryMetrics = computed(() => ([
{
label: 'P50',
value: formatSeconds(getLatestMetricValue('50')),
className: 'border-sky-500/20 bg-sky-500/8 text-sky-700 dark:text-sky-300',
},
{
label: 'P90',
value: formatSeconds(getLatestMetricValue('90')),
className: 'border-amber-500/20 bg-amber-500/8 text-amber-700 dark:text-amber-300',
},
{
label: 'P99',
value: formatSeconds(getLatestMetricValue('99')),
className: 'border-rose-500/20 bg-rose-500/8 text-rose-700 dark:text-rose-300',
},
]))
const chartData = computed<ChartData<'line'>>(() => ({
labels: labels.value,
datasets: [
{
label: 'P50',
data: props.series.map(item => getPercentileValue(item, '50')),
borderColor: 'rgb(70, 136, 240)',
backgroundColor: 'rgba(70, 136, 240, 0.10)',
borderWidth: 2.4,
fill: false,
tension: 0.32,
pointRadius: 0,
pointHoverRadius: 4,
pointHitRadius: 14,
},
{
label: 'P90',
data: props.series.map(item => getPercentileValue(item, '90')),
borderColor: 'rgb(223, 166, 55)',
backgroundColor: 'rgba(223, 166, 55, 0.08)',
borderWidth: 2.1,
fill: false,
tension: 0.3,
pointRadius: 0,
pointHoverRadius: 4,
pointHitRadius: 14,
},
{
label: 'P99',
data: props.series.map(item => getPercentileValue(item, '99')),
borderColor: 'rgb(214, 92, 111)',
backgroundColor: 'rgba(214, 92, 111, 0.08)',
borderWidth: 2,
borderDash: [6, 4],
fill: false,
tension: 0.28,
pointRadius: 0,
pointHoverRadius: 4,
pointHitRadius: 14,
},
],
}))
const chartOptions = computed<ChartOptions<'line'>>(() => ({
responsive: true,
maintainAspectRatio: false,
interaction: {
mode: 'index',
intersect: false,
},
plugins: {
legend: {
position: 'bottom',
labels: {
usePointStyle: true,
boxWidth: 10,
padding: 14,
},
},
tooltip: {
callbacks: {
title: items => props.series[items[0]?.dataIndex ?? -1]?.date ?? '',
label: context => `${context.dataset.label}: ${formatSeconds(Number(context.raw ?? 0))}`,
},
},
},
scales: {
y: {
x: {
grid: {
color: 'rgba(148, 163, 184, 0.06)',
},
ticks: {
callback: (value: number) => `${value}s`
}
}
}
color: 'rgb(100, 116, 139)',
maxRotation: 0,
},
},
y: {
beginAtZero: true,
grid: {
color: 'rgba(148, 163, 184, 0.07)',
},
ticks: {
color: 'rgb(100, 116, 139)',
callback: value => formatSeconds(Number(value)),
},
},
},
}))
</script>
@@ -1,81 +0,0 @@
<template>
<Card class="p-4 space-y-4">
<div class="flex items-center justify-between">
<h3 class="text-sm font-semibold">
{{ title }}
</h3>
<span
v-if="subtitle"
class="text-xs text-muted-foreground"
>{{ subtitle }}</span>
</div>
<div
v-if="loading"
class="p-4"
>
<LoadingState />
</div>
<div
v-else-if="providers.length === 0"
class="p-4"
>
<EmptyState
title="暂无数据"
description="暂无月卡配额数据"
/>
</div>
<div
v-else
class="space-y-4"
>
<div
v-for="provider in providers"
:key="provider.id"
class="space-y-2"
>
<div class="flex items-center justify-between text-xs">
<span class="font-medium">{{ provider.name }}</span>
<span class="text-muted-foreground">
{{ formatCurrency(provider.used_usd) }} / {{ formatCurrency(provider.quota_usd) }}
</span>
</div>
<div class="h-2 rounded-full bg-muted">
<div
class="h-2 rounded-full bg-primary"
:style="{ width: `${Math.min(provider.usage_percent, 100)}%` }"
/>
</div>
<div class="flex items-center justify-between text-[11px] text-muted-foreground">
<span>剩余 {{ formatCurrency(provider.remaining_usd) }}</span>
<span v-if="provider.estimated_exhaust_at">
预计耗尽 {{ formatDate(provider.estimated_exhaust_at) }}
</span>
</div>
</div>
</div>
</Card>
</template>
<script setup lang="ts">
import { Card } from '@/components/ui'
import { EmptyState, LoadingState } from '@/components/common'
import { formatCurrency } from '@/utils/format'
import type { QuotaUsageProvider } from '@/api/admin'
interface Props {
title: string
subtitle?: string
providers: QuotaUsageProvider[]
loading?: boolean
}
withDefaults(defineProps<Props>(), {
subtitle: undefined,
loading: false
})
function formatDate(value: string) {
return new Date(value).toLocaleDateString()
}
</script>
-3
View File
@@ -1,7 +1,4 @@
export { default as ActivityHeatmap } from './ActivityHeatmap.vue'
export { default as LeaderboardTable } from './LeaderboardTable.vue'
export { default as CostForecastChart } from './CostForecastChart.vue'
export { default as QuotaProgressCard } from './QuotaProgressCard.vue'
export { default as PercentileChart } from './PercentileChart.vue'
export { default as ErrorDistributionChart } from './ErrorDistributionChart.vue'
export { default as ComparisonCard } from './ComparisonCard.vue'
+2 -1
View File
@@ -109,6 +109,7 @@ const props = defineProps<{
zIndex?: number // Custom z-index for nested dialogs (default: 60)
noPadding?: boolean // Disable default content padding
persistent?: boolean // Prevent closing on backdrop click
closeOnBackdrop?: boolean // Allow closing on backdrop click (default: true)
}>()
// Emits 定义
@@ -145,7 +146,7 @@ function handleClose() {
// 处理背景点击
function handleBackdropClick() {
if (!props.persistent) {
if (!props.persistent && props.closeOnBackdrop !== false) {
handleClose()
}
}
+14 -37
View File
@@ -6,7 +6,7 @@
<input
ref="inputRef"
:class="inputClass"
:style="inputStyle"
:style="maskStyle"
:value="modelValue"
:type="effectiveType"
:autocomplete="autocompleteAttr"
@@ -42,7 +42,7 @@
v-else
ref="inputRef"
:class="inputClass"
:style="inputStyle"
:style="maskStyle"
:value="modelValue"
:type="effectiveType"
:autocomplete="autocompleteAttr"
@@ -69,24 +69,6 @@ const emit = defineEmits<{
'update:modelValue': [value: string]
}>()
// 开发环境警告:type="password" 已被弃用
const warnPasswordType = import.meta.env.DEV
? (() => {
let warned = false
return () => {
if (!warned) {
warned = true
// eslint-disable-next-line no-console
console.warn(
'[Input] type="password" 已被弃用,请使用 masked 属性代替。\n' +
'示例:<Input v-model="apiKey" masked />\n' +
'masked 属性使用 CSS 遮蔽而非 password 类型,不会触发浏览器密码管理器。'
)
}
}
})()
: () => {}
interface Props {
modelValue?: string | number
class?: string
@@ -99,9 +81,8 @@ interface Props {
size?: 'default' | 'sm'
/**
* 遮蔽显示内容(用于 API Key 等敏感信息)
* 使用 CSS -webkit-text-security 实现,不会触发浏览器密码管理器
* 始终使用 text 输入框,隐藏态通过样式进行遮蔽
* 同时会显示一个小眼睛按钮用于切换显示/隐藏
* 注意:Firefox 不支持 -webkit-text-security,会显示明文(但仍可通过按钮切换)
*/
masked?: boolean
/**
@@ -134,17 +115,23 @@ const shouldDisableAutofill = computed(() => {
return props.disableAutofill ?? false
})
// 始终使用 text 类型,永远不用 password
const effectiveType = computed(() => {
const attrType = attrs.type as string | undefined
// 如果传入 password,强制转为 text(配合 masked 使用)
if (attrType === 'password') {
warnPasswordType()
const attrType = (attrs.type as string | undefined) ?? 'text'
if (props.masked) {
return 'text'
}
return attrType
})
const maskStyle = computed(() => {
if (!props.masked || isVisible.value) {
return undefined
}
return {
WebkitTextSecurity: 'disc'
}
})
// 过滤掉 type 和 class 属性,因为我们会单独处理
const filteredAttrs = computed(() => {
const { type: _type, class: _class, ...rest } = attrs
@@ -182,16 +169,6 @@ const inputClass = computed(() =>
)
)
// 当 masked 为 true 且未显示时,用 CSS 遮蔽文字
const inputStyle = computed(() => {
if (props.masked && !isVisible.value) {
// 使用 -webkit-text-security(Chrome, Safari, Edge 支持)
// Firefox 不支持此属性,会显示明文,但仍可通过小眼睛按钮切换
return { '-webkit-text-security': 'disc' }
}
return undefined
})
function handleInput(event: Event) {
const target = event.target as HTMLInputElement
emit('update:modelValue', target.value)
+11 -3
View File
@@ -59,7 +59,7 @@ const isHidden = computed(
)
const itemClass = computed(() =>
cn(
'relative flex w-full cursor-pointer select-none items-center rounded-lg py-1.5 pl-8 pr-2 text-sm outline-none',
'relative flex min-w-0 w-full cursor-pointer select-none items-center rounded-lg py-1.5 pl-8 pr-2 text-sm outline-none',
'data-[highlighted]:bg-accent focus:bg-accent text-foreground',
'transition-colors data-[disabled]:pointer-events-none data-[disabled]:opacity-50',
isHidden.value && 'hidden',
@@ -101,8 +101,16 @@ onBeforeUnmount(() => {
<Check class="h-4 w-4" />
</SelectItemIndicator>
</span>
<SelectItemText>
<slot />
<SelectItemText
as="div"
class="min-w-0 flex-1 overflow-hidden"
>
<span
class="block w-full min-w-0 truncate"
:title="normalizedText"
>
<slot />
</span>
</SelectItemText>
</SelectItemPrimitive>
</template>
@@ -25,7 +25,7 @@ const triggerClass = computed(() =>
:class="triggerClass"
:disabled="disabled"
>
<span class="truncate">
<span class="min-w-0 flex-1 truncate">
<slot />
</span>
<ChevronDown class="h-4 w-4 opacity-50 pointer-events-none flex-shrink-0" />
@@ -0,0 +1,3 @@
export { useLeaderboardData } from './useLeaderboardData'
export type { LeaderboardDimension } from './useLeaderboardData'
export { usePerformanceData } from './usePerformanceData'
@@ -0,0 +1,23 @@
import type { AnalyticsLeaderboardItem } from '@/api/analytics'
export interface LeaderboardTableRow {
rank: number
id: string
name: string
requests: number
tokens: number
cost: number
actualCost: number
}
export function mapAnalyticsLeaderboardItem(item: AnalyticsLeaderboardItem): LeaderboardTableRow {
return {
rank: item.rank,
id: item.id,
name: item.label,
requests: item.requests_total,
tokens: item.total_tokens,
cost: item.total_cost_usd,
actualCost: item.actual_total_cost_usd,
}
}
@@ -0,0 +1,272 @@
import { computed, onMounted, onUnmounted, ref, watch, type Ref } from 'vue'
import { analyticsApi, type AnalyticsFilterOption, type AnalyticsLeaderboardItem, type AnalyticsTimeseriesBucket } from '@/api/analytics'
import { buildTimeRangeParams, createLoader, createRequestGuard, useInjectedAnalyticsFilters } from '@/composables/useAnalyticsFilters'
import { fillMissingTimeseriesBuckets } from '@/utils/analyticsTimeseries'
import { isAnalyticsSingleDayRange, resolveAnalyticsAutoGranularity, type AnalyticsGranularityOption } from '@/utils/analyticsGranularity'
import { mapAnalyticsLeaderboardItem, type LeaderboardTableRow } from './mappers'
export type LeaderboardDimension = 'users' | 'api_keys'
export interface LeaderboardSummary {
total_requests: number
total_tokens: number
total_cost: number
total_actual_cost: number
avg_response_time: number
}
function dimensionToEntity(dimension: LeaderboardDimension): 'user' | 'api_key' {
return dimension === 'users' ? 'user' : 'api_key'
}
export function useLeaderboardData(dimension: Ref<LeaderboardDimension>) {
const filters = useInjectedAnalyticsFilters()
const leaderboardItems = ref<AnalyticsLeaderboardItem[]>([])
const latestUserOptions = ref<AnalyticsFilterOption[]>([])
const latestApiKeyOptions = ref<AnalyticsFilterOption[]>([])
const leaderboardLoading = ref(false)
const leaderboardError = ref(false)
const leaderboardHasLoaded = ref(false)
const selectedId = ref<string>('')
const selectedSummary = ref<LeaderboardSummary | null>(null)
const selectedTrendBuckets = ref<AnalyticsTimeseriesBucket[]>([])
const panelLoading = ref(false)
const panelError = ref(false)
const panelHasLoaded = ref(false)
const leaderboardGuard = createRequestGuard()
const filterOptionsGuard = createRequestGuard()
const panelGuard = createRequestGuard()
const entityOptions = computed(() => (
dimension.value === 'users' ? latestUserOptions.value : latestApiKeyOptions.value
))
const entityLabelMap = computed(() => (
new Map(entityOptions.value.map(option => [option.value, option.label]))
))
const leaderboard = computed<LeaderboardTableRow[]>(() => (
leaderboardItems.value.map(item => ({
...mapAnalyticsLeaderboardItem(item),
name: entityLabelMap.value.get(item.id) ?? item.label,
}))
))
const selectedItem = computed(() => (
leaderboard.value.find(item => item.id === selectedId.value) ?? null
))
const selectedLabel = computed(() => (
selectedItem.value?.name ?? entityLabelMap.value.get(selectedId.value) ?? null
))
const resolvedTrendGranularity = computed(() => resolveAnalyticsAutoGranularity(filters.timeRange.value))
const baseTimeRangePayload = computed(() => ({
start_date: filters.timeRange.value.start_date,
end_date: filters.timeRange.value.end_date,
preset: filters.timeRange.value.preset,
timezone: filters.timeRange.value.timezone,
tz_offset_minutes: filters.timeRange.value.tz_offset_minutes,
}))
const trendTimeRangePayload = computed(() => ({
...baseTimeRangePayload.value,
granularity: resolvedTrendGranularity.value,
}))
const currentTrendGranularitySelection = computed<AnalyticsGranularityOption>({
get: () => filters.timeRange.value.granularity || 'auto',
set: (value) => {
if (value === 'hour' && !isAnalyticsSingleDayRange(filters.timeRange.value)) {
value = 'auto'
}
filters.timeRange.value = {
...filters.timeRange.value,
granularity: value,
}
},
})
function resetPanelData(invalidate = false) {
if (invalidate) panelGuard.invalidate()
panelLoading.value = false
panelError.value = false
selectedSummary.value = null
selectedTrendBuckets.value = []
}
async function loadFilterOptions() {
const requestId = filterOptionsGuard.next()
const response = await analyticsApi.getFilterOptions({
scope: { kind: 'global' },
time_range: baseTimeRangePayload.value,
filters: {},
}).catch(() => null)
if (filterOptionsGuard.isStale(requestId) || !response) return
latestUserOptions.value = response.users ?? []
latestApiKeyOptions.value = response.api_keys ?? []
}
async function loadLeaderboard() {
const requestId = leaderboardGuard.next()
leaderboardLoading.value = true
leaderboardError.value = false
try {
const response = await analyticsApi.getLeaderboard({
scope: { kind: 'global' },
time_range: baseTimeRangePayload.value,
filters: {
statuses: ['completed', 'failed', 'cancelled'],
},
entity: dimensionToEntity(dimension.value),
metric: 'total_cost_usd',
limit: 20,
})
if (leaderboardGuard.isStale(requestId)) return
leaderboardItems.value = response.items
const nextSelectedId = response.items.some(item => item.id === selectedId.value)
? selectedId.value
: (response.items[0]?.id ?? '')
if (selectedId.value !== nextSelectedId) {
selectedId.value = nextSelectedId
}
leaderboardHasLoaded.value = true
} catch {
if (leaderboardGuard.isStale(requestId)) return
leaderboardError.value = true
} finally {
if (leaderboardGuard.isCurrent(requestId)) {
leaderboardLoading.value = false
}
}
}
async function loadPanel() {
if (!selectedId.value) {
resetPanelData(true)
return
}
const requestId = panelGuard.next()
panelLoading.value = true
panelError.value = false
try {
const scope = dimension.value === 'users'
? { kind: 'user' as const, user_id: selectedId.value }
: { kind: 'api_key' as const, api_key_id: selectedId.value }
const basePayload = {
scope,
time_range: baseTimeRangePayload.value,
filters: {
statuses: ['completed', 'failed', 'cancelled'],
},
}
const [overview, timeseries] = await Promise.all([
analyticsApi.getOverview(basePayload),
analyticsApi.getTimeseries({
...basePayload,
time_range: trendTimeRangePayload.value,
}),
])
if (panelGuard.isStale(requestId)) return
selectedSummary.value = {
total_requests: overview.summary.requests_total,
total_tokens: overview.summary.total_tokens,
total_cost: overview.summary.total_cost_usd,
total_actual_cost: overview.summary.actual_total_cost_usd,
avg_response_time: overview.summary.avg_response_time_ms / 1000,
}
selectedTrendBuckets.value = fillMissingTimeseriesBuckets(
timeseries.buckets,
trendTimeRangePayload.value,
)
panelHasLoaded.value = true
} catch {
if (panelGuard.isStale(requestId)) return
panelError.value = true
} finally {
if (panelGuard.isCurrent(requestId)) {
panelLoading.value = false
}
}
}
const filterOptionsLoader = createLoader(loadFilterOptions)
const leaderboardLoader = createLoader(loadLeaderboard)
const panelLoader = createLoader(loadPanel)
watch(
() => [filters.timeRange.value.preset, filters.timeRange.value.start_date, filters.timeRange.value.end_date],
() => {
if (filters.timeRange.value.granularity === 'hour' && !isAnalyticsSingleDayRange(filters.timeRange.value)) {
filters.timeRange.value = {
...filters.timeRange.value,
granularity: 'auto',
}
}
},
)
watch(
() => [baseTimeRangePayload.value, dimension.value],
() => {
filterOptionsLoader.schedule()
leaderboardLoader.schedule()
},
{ deep: true },
)
watch(
() => [
selectedId.value,
dimension.value,
baseTimeRangePayload.value,
trendTimeRangePayload.value,
],
() => panelLoader.schedule(),
{ deep: true },
)
onMounted(() => {
void (async () => {
await filterOptionsLoader.execute()
await leaderboardLoader.execute()
await panelLoader.execute()
})()
})
onUnmounted(() => {
filterOptionsLoader.cleanup()
leaderboardLoader.cleanup()
panelLoader.cleanup()
filterOptionsGuard.invalidate()
leaderboardGuard.invalidate()
panelGuard.invalidate()
})
return {
leaderboard,
leaderboardLoading,
leaderboardError,
leaderboardHasLoaded,
selectedId,
selectedItem,
selectedLabel,
selectedSummary,
selectedTrendBuckets,
panelLoading,
panelError,
panelHasLoaded,
resolvedTrendGranularity,
currentTrendGranularitySelection,
reloadLeaderboard: leaderboardLoader.schedule,
}
}
@@ -0,0 +1,98 @@
import { ref, watch, onMounted, onUnmounted } from 'vue'
import { analyticsApi, type AnalyticsErrorCategory, type AnalyticsErrorTrendItem, type AnalyticsPercentilePoint, type AnalyticsPerformanceResponse, type AnalyticsProviderHealthItem } from '@/api/analytics'
import { auditApi, type MonitoringSystemStatus } from '@/api/audit'
import { getHealthSummary, type HealthSummary } from '@/api/endpoints'
import { createLoader, createRequestGuard, useInjectedAnalyticsFilters } from '@/composables/useAnalyticsFilters'
const EMPTY_LATENCY: AnalyticsPerformanceResponse['latency'] = {
response_time_ms: { avg: 0, p50: null, p90: null, p99: null },
first_byte_time_ms: { avg: 0, p50: null, p90: null, p99: null },
}
export function usePerformanceData() {
const filters = useInjectedAnalyticsFilters()
const latency = ref<AnalyticsPerformanceResponse['latency']>(EMPTY_LATENCY)
const percentiles = ref<AnalyticsPercentilePoint[]>([])
const errorDistribution = ref<AnalyticsErrorCategory[]>([])
const errorTrend = ref<AnalyticsErrorTrendItem[]>([])
const errorTotal = ref(0)
const errorRate = ref(0)
const providerStatus = ref<AnalyticsProviderHealthItem[]>([])
const healthSummary = ref<HealthSummary | null>(null)
const systemStatus = ref<MonitoringSystemStatus | null>(null)
const loading = ref(false)
const loadError = ref(false)
const hasLoaded = ref(false)
const guard = createRequestGuard()
async function load() {
const requestId = guard.next()
loading.value = true
loadError.value = false
try {
const [response, summaryResponse, systemResponse] = await Promise.all([
analyticsApi.getPerformance({
scope: { kind: 'global' },
time_range: filters.getTimeRangeParams(),
filters: {
statuses: ['completed', 'failed', 'cancelled'],
},
}),
getHealthSummary().catch(() => null),
auditApi.getSystemStatus().catch(() => null),
])
if (guard.isStale(requestId)) return
latency.value = response.latency
percentiles.value = response.percentiles
errorTotal.value = response.errors.total
errorRate.value = response.errors.rate
errorDistribution.value = response.errors.categories
errorTrend.value = response.errors.trend
providerStatus.value = response.provider_health
if (summaryResponse) {
healthSummary.value = summaryResponse
}
if (systemResponse) {
systemStatus.value = systemResponse
}
hasLoaded.value = true
} catch {
if (guard.isStale(requestId)) return
loadError.value = true
} finally {
if (guard.isCurrent(requestId)) {
loading.value = false
}
}
}
const loader = createLoader(load)
watch(
() => [
filters.getTimeRangeParams(),
],
() => loader.schedule(),
{ deep: true },
)
onMounted(() => { void loader.execute() })
onUnmounted(() => { loader.cleanup(); guard.invalidate() })
return {
latency,
percentiles,
errorDistribution,
errorTrend,
errorTotal,
errorRate,
providerStatus,
healthSummary,
systemStatus,
loading,
loadError,
hasLoaded,
}
}
@@ -0,0 +1,144 @@
import { inject, ref, watch } from 'vue'
import { useRoute, useRouter } from 'vue-router'
import type { DateRangeParams, PeriodValue } from '@/features/usage/types'
import { getDateRangeFromPeriod } from '@/features/usage/composables'
/**
* 构建时间范围 API 参数(从各页面提取的公共逻辑)
*/
export function buildTimeRangeParams(timeRange: DateRangeParams) {
return {
start_date: timeRange.start_date,
end_date: timeRange.end_date,
preset: timeRange.preset,
timezone: timeRange.timezone,
tz_offset_minutes: timeRange.tz_offset_minutes,
granularity: timeRange.granularity && timeRange.granularity !== 'auto'
? timeRange.granularity
: 'day',
}
}
/**
* 通用防抖加载器工厂
* 封装 120ms debounce + in-flight promise 去重 + requestId 过期检查
*/
export function createLoader(loadFn: () => Promise<void>) {
let loadPromise: Promise<void> | null = null
let hasPending = false
let debounceTimer: ReturnType<typeof setTimeout> | null = null
function schedule() {
if (debounceTimer) clearTimeout(debounceTimer)
debounceTimer = setTimeout(() => {
debounceTimer = null
void execute()
}, 120)
}
async function execute() {
if (loadPromise) {
hasPending = true
return loadPromise
}
loadPromise = loadFn()
.finally(() => {
loadPromise = null
if (hasPending) {
hasPending = false
void execute()
}
})
return loadPromise
}
function cleanup() {
if (debounceTimer) {
clearTimeout(debounceTimer)
debounceTimer = null
}
hasPending = false
loadPromise = null
}
return { schedule, execute, cleanup }
}
/**
* 请求 ID 守卫:用于确保异步回调中的数据不被过期请求覆盖
*/
export function createRequestGuard() {
let id = 0
function next() { return ++id }
function isStale(requestId: number) { return requestId !== id }
function isCurrent(requestId: number) { return requestId === id }
function invalidate() { id++ }
return { next, isStale, isCurrent, invalidate }
}
export interface AnalyticsFiltersOptions {
defaultPreset?: PeriodValue
syncToUrl?: boolean
}
/**
* 统一的分析页面筛选状态
* - timeRange: 时间范围
* - modelFilter / providerFilter / apiKeyFilter / userFilter: 多维度筛选
* - activeTab: 当前 Tab
* - URL query 双向同步
*/
export function useAnalyticsFilters(options: AnalyticsFiltersOptions = {}) {
const { defaultPreset = 'last30days', syncToUrl = true } = options
const route = useRoute()
const router = useRouter()
// 从 URL query 初始化,无则用默认值
const initialPreset = (syncToUrl && route.query.preset as PeriodValue) || defaultPreset
const initialTab = (syncToUrl && route.query.tab as string) || 'detail'
const timeRange = ref<DateRangeParams>(getDateRangeFromPeriod(initialPreset))
const activeTab = ref(initialTab)
const modelFilter = ref<string[]>([])
const providerFilter = ref<string[]>([])
const apiKeyFilter = ref<string[]>([])
const userFilter = ref<string[]>([])
// URL 同步:仅同步 preset 和 tab(各类多选筛选太冗长不入 URL)
if (syncToUrl) {
watch([() => timeRange.value.preset, activeTab], ([preset, tab]) => {
const query: Record<string, string> = {}
if (preset && preset !== defaultPreset) query.preset = preset
if (tab && tab !== 'detail') query.tab = tab
// 保留其他 query 参数
const currentQuery = { ...route.query }
delete currentQuery.preset
delete currentQuery.tab
router.replace({ query: { ...currentQuery, ...query } })
}, { flush: 'post' })
}
function getTimeRangeParams() {
return buildTimeRangeParams(timeRange.value)
}
return {
timeRange,
activeTab,
modelFilter,
providerFilter,
apiKeyFilter,
userFilter,
getTimeRangeParams,
}
}
export type AnalyticsFilters = ReturnType<typeof useAnalyticsFilters>
export function useInjectedAnalyticsFilters(): AnalyticsFilters {
const filters = inject<AnalyticsFilters>('analyticsFilters', null)
if (!filters) {
throw new Error('analyticsFilters provider is missing')
}
return filters
}
+85 -2
View File
@@ -68,6 +68,86 @@ export function getProbeCountdown(nextProbeAt: string | null | undefined, _tick:
return '探测中'
}
/**
* Codex 配额重置倒计时状态
*/
export interface CodexResetStatus {
text: string
isUrgent: boolean
isCritical: boolean
isExpired: boolean
}
/**
* 计算 Codex 配额重置倒计时
* @param resetAt 绝对重置时间(Unix 秒)
* @param resetSecs 相对剩余秒数(用于 fallback)
* @param updatedAt 元数据更新时间(Unix 秒)
* @param _tick 响应式触发器(传入 tick.value 以触发响应式更新)
* @param remainingPercent 当前窗口剩余额度百分比(0-100,100 表示满额不启动倒计时)
*/
export function getCodexResetCountdown(
resetAt: number | null | undefined,
resetSecs: number | null | undefined,
updatedAt: number | null | undefined,
_tick: number,
remainingPercent?: number | null
): CodexResetStatus | null {
void _tick
if (remainingPercent != null) {
const normalizedRemaining = Number(remainingPercent)
if (Number.isFinite(normalizedRemaining) && normalizedRemaining >= 100) {
return null
}
}
const nowSec = Math.floor(Date.now() / 1000)
let remaining: number
if (resetAt != null && resetAt > 0) {
// 优先绝对时间戳,避免相对秒数快照漂移。
remaining = resetAt - nowSec
} else if (resetSecs != null && resetSecs >= 0) {
if (updatedAt != null && updatedAt > 0) {
// 时钟偏移下 updatedAt 可能晚于当前时间,elapsed 需要下限钳制到 0。
const elapsedSec = Math.max(nowSec - updatedAt, 0)
remaining = resetSecs - elapsedSec
} else {
remaining = resetSecs
}
} else {
return null
}
if (remaining <= 0) {
return { text: '已重置', isUrgent: false, isCritical: false, isExpired: true }
}
const total = Math.floor(remaining)
const days = Math.floor(total / 86400)
const hours = Math.floor((total % 86400) / 3600)
const minutes = Math.floor((total % 3600) / 60)
const seconds = total % 60
const pad = (n: number) => n.toString().padStart(2, '0')
let text: string
if (days > 0) {
text = `${days}天 ${hours}:${pad(minutes)}:${pad(seconds)}`
} else if (hours > 0) {
text = `${hours}:${pad(minutes)}:${pad(seconds)}`
} else {
text = `${minutes}:${pad(seconds)}`
}
return {
text,
isUrgent: total < 3600,
isCritical: total < 300,
isExpired: false,
}
}
/**
* OAuth Token 状态信息
*/
@@ -94,15 +174,18 @@ export function getOAuthExpiresCountdown(
invalidReason?: string | null
): OAuthStatusInfo | null {
void _tick
const normalizedInvalidReason = typeof invalidReason === 'string'
? invalidReason.trim()
: ''
// 优先检查失效状态(失效比过期更严重)
if (invalidAt != null) {
if (invalidAt != null || normalizedInvalidReason) {
return {
text: '已失效',
isExpired: false,
isExpiringSoon: false,
isInvalid: true,
invalidReason: invalidReason || undefined
invalidReason: normalizedInvalidReason || undefined
}
}
+23 -2
View File
@@ -15,6 +15,8 @@ export interface StartTestParams {
apiFormat?: string
endpointId?: string
message?: string
requestHeaders?: Record<string, unknown>
requestBody?: Record<string, unknown>
concurrency?: number
onSuccess?: (result: TestModelFailoverResponse) => void
/** Return `true` to indicate the failure has been handled; otherwise the composable sets `testResult`. */
@@ -61,6 +63,16 @@ export function useModelTest(options: UseModelTestOptions) {
}
}
async function refreshTraceSnapshot(reqId: string) {
try {
const trace = await requestTraceApi.getRequestTrace(reqId, { attemptedOnly: false })
if (requestId.value !== reqId) return
testTrace.value = trace
} catch (err: unknown) {
if (isAxiosError(err) && err.response?.status === 404) return
}
}
function stopPolling(opts: { clearState?: boolean } = {}) {
tracePollToken += 1
if (tracePollTimer) {
@@ -117,13 +129,19 @@ export function useModelTest(options: UseModelTestOptions) {
startPolling(reqId)
try {
const normalizedMessage = typeof params.message === 'string' && params.message.trim()
? params.message.trim()
: undefined
const result = await testModelFailover({
provider_id: providerId(),
mode: params.mode,
model_name: params.modelName,
api_format: params.apiFormat,
endpoint_id: params.endpointId,
message: params.message ?? 'hello',
...(normalizedMessage ? { message: normalizedMessage } : {}),
...(params.requestHeaders ? { request_headers: params.requestHeaders } : {}),
...(params.requestBody ? { request_body: params.requestBody } : {}),
request_id: reqId,
concurrency: params.concurrency,
}, {
@@ -131,6 +149,9 @@ export function useModelTest(options: UseModelTestOptions) {
})
if (result.success) {
await refreshTraceSnapshot(reqId)
stopPolling({ clearState: false })
testResult.value = result
const successAttempt = result.attempts.find(a => a.status === 'success')
const latency = successAttempt?.latency_ms != null ? ` (${successAttempt.latency_ms}ms)` : ''
const mapped = successAttempt?.effective_model && successAttempt.effective_model !== params.modelName
@@ -138,10 +159,10 @@ export function useModelTest(options: UseModelTestOptions) {
: ''
params.onSuccess?.(result)
showSuccess(`${params.displayLabel}${mapped} 测试成功${latency}`)
resetState()
return
}
await refreshTraceSnapshot(reqId)
stopPolling({ clearState: false })
const handled = params.onFailure?.(result)
if (!handled) {
+287
View File
@@ -0,0 +1,287 @@
import { computed, ref, watch, onMounted, onUnmounted, type Ref } from 'vue'
import {
analyticsApi,
type AnalyticsBreakdownRow,
type AnalyticsFilterOption,
type AnalyticsScope,
type AnalyticsSummary,
type AnalyticsTimeseriesBucket,
} from '@/api/analytics'
import type { DateRangeParams, PeriodValue } from '@/features/usage/types'
import { getDateRangeFromPeriod } from '@/features/usage/composables'
import { createLoader, createRequestGuard, buildTimeRangeParams } from '@/composables/useAnalyticsFilters'
import { fillMissingTimeseriesBuckets } from '@/utils/analyticsTimeseries'
import { getAnalyticsRangeDaysInclusive, resolveAnalyticsAutoGranularity, type ResolvedAnalyticsGranularity } from '@/utils/analyticsGranularity'
import { type DailyUsageBreakdown } from '@/utils/usageBreakdown'
export interface ModelSummary {
model: string
requests: number
tokens: number
cost: number
inputTokens: number
outputTokens: number
cacheCreationTokens: number
cacheReadTokens: number
cacheHitRate: number
avg_response_time?: number
avg_first_byte_time?: number
}
export interface ReportBucketStat {
bucket_start: string
bucket_end: string
requests: number
tokens: number
cost: number
avg_response_time: number
avg_first_byte_time: number
models_used_count: number
}
function mapBucketToReportStat(bucket: AnalyticsTimeseriesBucket): ReportBucketStat {
return {
bucket_start: bucket.bucket_start,
bucket_end: bucket.bucket_end,
requests: bucket.requests_total,
tokens: bucket.total_tokens,
cost: bucket.total_cost_usd,
avg_response_time: bucket.avg_response_time_ms / 1000,
avg_first_byte_time: bucket.avg_first_byte_time_ms / 1000,
models_used_count: bucket.models_used_count,
}
}
function mapBucketToBreakdown(bucket: AnalyticsTimeseriesBucket): DailyUsageBreakdown {
return {
date: bucket.bucket_start,
inputTokens: bucket.input_tokens,
outputTokens: bucket.output_tokens,
cacheCreationTokens: bucket.cache_creation_input_tokens,
cacheReadTokens: bucket.cache_read_input_tokens,
cacheCreationCost: bucket.cache_creation_cost_usd,
cacheReadCost: bucket.cache_read_cost_usd,
cacheHitRate: bucket.cache_hit_rate,
totalCacheCost: bucket.cache_cost_usd,
totalCacheTokens: bucket.cache_creation_input_tokens + bucket.cache_read_input_tokens,
totalTrackedTokens: bucket.total_tokens,
baseTokens: bucket.input_output_total_tokens,
}
}
function mapBreakdownRowToModelSummary(row: AnalyticsBreakdownRow): ModelSummary {
const requests = row.requests_total
const tokens = row.total_tokens
const cost = row.total_cost_usd
return {
model: row.label,
requests,
tokens,
cost,
inputTokens: row.input_tokens,
outputTokens: row.output_tokens,
cacheCreationTokens: row.cache_creation_input_tokens,
cacheReadTokens: row.cache_read_input_tokens,
cacheHitRate: row.cache_hit_rate,
avg_response_time: row.avg_response_time_ms / 1000,
avg_first_byte_time: row.avg_first_byte_time_ms / 1000,
}
}
type ReportGranularity = ResolvedAnalyticsGranularity
function resolveReportGranularity(range: DateRangeParams): ReportGranularity {
return resolveAnalyticsAutoGranularity(range)
}
function areStringArraysEqual(left: string[], right: string[]): boolean {
if (left.length !== right.length) return false
return left.every((value, index) => value === right[index])
}
export interface UseReportsDataOptions {
scope?: AnalyticsScope
defaultPreset?: PeriodValue
timeRange?: Ref<DateRangeParams>
userFilter?: Ref<string[]>
apiKeyFilter?: Ref<string[]>
loadApiKeyOptions?: boolean
}
export function useReportsData(options: UseReportsDataOptions = {}) {
const {
scope = { kind: 'me' as const },
defaultPreset = 'today',
timeRange: externalTimeRange,
userFilter: externalUserFilter,
apiKeyFilter: externalApiKeyFilter,
loadApiKeyOptions = scope.kind === 'me',
} = options
const timeRange = externalTimeRange ?? ref<DateRangeParams>({
...getDateRangeFromPeriod(defaultPreset),
granularity: 'auto',
})
const userFilter = externalUserFilter ?? ref<string[]>([])
const apiKeyFilter = externalApiKeyFilter ?? ref<string[]>([])
const apiKeyOptions = ref<AnalyticsFilterOption[]>([])
const summary = ref<AnalyticsSummary | null>(null)
const bucketStats = ref<ReportBucketStat[]>([])
const bucketBreakdowns = ref<Record<string, DailyUsageBreakdown>>({})
const modelSummary = ref<ModelSummary[]>([])
const loading = ref(false)
const loadError = ref(false)
const hasLoaded = ref(false)
const guard = createRequestGuard()
const apiKeysGuard = createRequestGuard()
let suppressNextApiKeyFilterLoad = false
const resolvedGranularity = computed(() => resolveReportGranularity(timeRange.value))
const timeseriesTimeRange = computed(() => ({
...buildTimeRangeParams(timeRange.value),
granularity: resolvedGranularity.value,
}))
watch(
() => [timeRange.value.preset, timeRange.value.start_date, timeRange.value.end_date],
() => {
if (timeRange.value.granularity === 'hour' && getAnalyticsRangeDaysInclusive(timeRange.value) !== 1) {
timeRange.value = {
...timeRange.value,
granularity: 'auto',
}
}
},
)
async function load() {
const requestId = guard.next()
loading.value = true
loadError.value = false
try {
const params = buildTimeRangeParams(timeRange.value)
const timeseriesParams = timeseriesTimeRange.value
const basePayload = {
scope,
time_range: params,
filters: {
user_ids: userFilter.value,
api_key_ids: apiKeyFilter.value,
},
}
const [overviewData, timeseriesData, modelBreakdownData] = await Promise.all([
analyticsApi.getOverview(basePayload),
analyticsApi.getTimeseries({
...basePayload,
time_range: timeseriesParams,
}),
analyticsApi.getBreakdown({
...basePayload,
dimension: 'model',
limit: 50,
}),
])
if (guard.isStale(requestId)) return
const filledBuckets = fillMissingTimeseriesBuckets(timeseriesData.buckets, timeseriesParams)
summary.value = overviewData.summary
bucketStats.value = filledBuckets.map(mapBucketToReportStat)
bucketBreakdowns.value = Object.fromEntries(
filledBuckets.map(bucket => {
const breakdown = mapBucketToBreakdown(bucket)
return [breakdown.date, breakdown]
}),
)
modelSummary.value = modelBreakdownData.rows.map(mapBreakdownRowToModelSummary)
hasLoaded.value = true
} catch {
if (guard.isStale(requestId)) return
loadError.value = true
} finally {
if (guard.isCurrent(requestId)) {
loading.value = false
}
}
}
async function loadApiKeys() {
if (!loadApiKeyOptions) return
const requestId = apiKeysGuard.next()
const filterOptions = await analyticsApi.getFilterOptions({
scope,
time_range: buildTimeRangeParams(timeRange.value),
filters: {},
}).catch(() => null)
if (apiKeysGuard.isStale(requestId)) return
const nextOptions = filterOptions?.api_keys ?? []
apiKeyOptions.value = nextOptions
const validOptionValues = new Set(nextOptions.map(option => option.value))
const nextFilter = apiKeyFilter.value.filter(value => validOptionValues.has(value))
if (!areStringArraysEqual(nextFilter, apiKeyFilter.value)) {
suppressNextApiKeyFilterLoad = true
apiKeyFilter.value = nextFilter
}
}
const loader = createLoader(load)
const apiKeysLoader = createLoader(loadApiKeys)
watch(
timeRange,
() => {
void (async () => {
if (loadApiKeyOptions) {
await apiKeysLoader.execute()
}
loader.schedule()
})()
},
{ deep: true },
)
watch(
[userFilter, apiKeyFilter],
() => {
if (suppressNextApiKeyFilterLoad) {
suppressNextApiKeyFilterLoad = false
return
}
loader.schedule()
},
{ deep: true },
)
onMounted(() => {
void (async () => {
if (loadApiKeyOptions) {
await apiKeysLoader.execute()
}
await loader.execute()
})()
})
onUnmounted(() => {
loader.cleanup()
apiKeysLoader.cleanup()
guard.invalidate()
apiKeysGuard.invalidate()
})
return {
timeRange,
apiKeyFilter,
apiKeyOptions,
summary,
bucketStats,
bucketBreakdowns,
modelSummary,
loading,
loadError,
hasLoaded,
resolvedGranularity,
}
}
+44
View File
@@ -0,0 +1,44 @@
import { useRoute, useRouter, type LocationQueryValue, type LocationQuery } from 'vue-router'
type QueryValue = LocationQueryValue | LocationQueryValue[]
function normalizeQueryValue(value: QueryValue): string | undefined {
if (Array.isArray(value)) {
return value.length > 0 ? (value[value.length - 1] ?? undefined) : undefined
}
return typeof value === 'string' ? value : undefined
}
function queriesEqual(left: LocationQuery, right: LocationQuery): boolean {
const keys = new Set([...Object.keys(left), ...Object.keys(right)])
for (const key of keys) {
if (normalizeQueryValue(left[key]) !== normalizeQueryValue(right[key])) {
return false
}
}
return true
}
export function useRouteQuery() {
const route = useRoute()
const router = useRouter()
function getQueryValue(key: string): string | undefined {
return normalizeQueryValue(route.query[key])
}
function patchQuery(patch: Record<string, string | undefined | null>) {
const next: LocationQuery = { ...route.query }
for (const [key, value] of Object.entries(patch)) {
if (value == null || value.trim() === '') {
delete next[key]
} else {
next[key] = value
}
}
if (queriesEqual(route.query, next)) return
void router.replace({ query: next }).catch(() => {})
}
return { route, router, getQueryValue, patchQuery }
}
@@ -96,23 +96,6 @@
{{ form.expires_at ? '到期后' + (form.auto_delete_on_expiry ? '自动删除' : '仅禁用') + '(当天 23:59 失效)' : '留空表示永不过期' }}
</p>
</div>
<div class="space-y-2">
<Label
for="form-rate-limit"
class="text-sm font-medium"
>速率限制 (请求/分钟)</Label>
<Input
id="form-rate-limit"
:model-value="form.rate_limit ?? ''"
type="number"
min="1"
max="10000"
placeholder="留空不限制"
class="h-10"
@update:model-value="(v) => form.rate_limit = parseNumberInput(v, { min: 1, max: 10000 })"
/>
</div>
</div>
<!-- 右侧:访问限制 -->
@@ -190,6 +173,36 @@
</div>
</div>
<div class="space-y-2">
<Label
for="form-rate-limit"
class="text-sm font-medium"
>速率限制 (请求/分钟)</Label>
<div class="flex items-center gap-3">
<div class="flex-1 min-w-0">
<Input
v-if="!form.rate_limit_inherited"
id="form-rate-limit"
:model-value="form.rate_limit ?? ''"
type="number"
min="0"
max="10000"
placeholder="0 = 不限速"
class="h-10"
@update:model-value="(v) => form.rate_limit = parseNumberInput(v, { min: 0, max: 10000 })"
/>
<span
v-else
class="flex h-10 w-full items-center rounded-lg border bg-background px-3 text-sm text-muted-foreground opacity-60"
>跟随系统</span>
</div>
<Switch
v-model="form.rate_limit_inherited"
class="shrink-0"
/>
</div>
</div>
<!-- 额度 -->
<div class="space-y-2">
<Label class="text-sm font-medium">额度</Label>
@@ -207,9 +220,9 @@
@update:model-value="(v) => form.initial_balance_usd = parseNumberInput(v, { allowFloat: true, min: 0.01 })"
/>
<span
v-else-if="form.unlimited_balance"
v-else
class="flex h-10 w-full items-center rounded-lg border bg-background px-3 text-sm text-muted-foreground opacity-60"
>无限制</span>
>{{ form.unlimited_balance ? '无限制' : '按钱包余额限制' }}</span>
</div>
<Switch
:model-value="form.unlimited_balance ?? false"
@@ -267,7 +280,7 @@ export interface StandaloneKeyFormData {
initial_balance_usd?: number
unlimited_balance?: boolean
expires_at?: string // ISO 日期字符串,如 "2025-12-31",undefined = 永不过期
rate_limit?: number
rate_limit?: number | null
auto_delete_on_expiry: boolean
allowed_providers?: string[] | null
allowed_api_formats?: string[] | null
@@ -280,6 +293,7 @@ interface StandaloneKeyFormState {
initial_balance_usd?: number
unlimited_balance?: boolean
expires_at?: string
rate_limit_inherited: boolean
rate_limit?: number
auto_delete_on_expiry: boolean
provider_unrestricted: boolean
@@ -333,6 +347,7 @@ const form = ref<StandaloneKeyFormState>({
initial_balance_usd: 10,
unlimited_balance: false,
expires_at: undefined,
rate_limit_inherited: true,
rate_limit: undefined,
auto_delete_on_expiry: false,
provider_unrestricted: true,
@@ -356,6 +371,7 @@ function resetForm() {
initial_balance_usd: 10,
unlimited_balance: false,
expires_at: undefined,
rate_limit_inherited: true,
rate_limit: undefined,
auto_delete_on_expiry: false,
provider_unrestricted: true,
@@ -375,7 +391,8 @@ function loadKeyData() {
initial_balance_usd: props.apiKey.initial_balance_usd,
unlimited_balance: props.apiKey.initial_balance_usd == null,
expires_at: props.apiKey.expires_at,
rate_limit: props.apiKey.rate_limit,
rate_limit_inherited: props.apiKey.rate_limit == null,
rate_limit: props.apiKey.rate_limit ?? undefined,
auto_delete_on_expiry: props.apiKey.auto_delete_on_expiry,
provider_unrestricted: props.apiKey.allowed_providers == null,
api_format_unrestricted: props.apiKey.allowed_api_formats == null,
@@ -425,7 +442,7 @@ function handleSubmit() {
initial_balance_usd: form.value.initial_balance_usd,
unlimited_balance: form.value.unlimited_balance,
expires_at: form.value.expires_at,
rate_limit: form.value.rate_limit,
rate_limit: form.value.rate_limit_inherited ? null : (form.value.rate_limit ?? 0),
auto_delete_on_expiry: form.value.auto_delete_on_expiry,
allowed_providers: form.value.provider_unrestricted ? null : [...form.value.allowed_providers],
allowed_api_formats: form.value.api_format_unrestricted ? null : [...form.value.allowed_api_formats],
@@ -179,7 +179,8 @@
<Input
id="login-password"
v-model="form.password"
type="password"
type="text"
masked
required
placeholder="输入密码"
autocomplete="off"
@@ -247,6 +248,7 @@ import { isDemoMode, DEMO_ACCOUNTS } from '@/config/demo'
import RegisterDialog from './RegisterDialog.vue'
import { authApi } from '@/api/auth'
import { oauthApi, type OAuthProviderInfo } from '@/api/oauth'
import { getClientDeviceId } from '@/utils/deviceId'
import { getApiUrl } from '@/utils/url'
import { getOAuthIcon } from '@/utils/oauth-icons'
@@ -350,7 +352,12 @@ async function handleLogin() {
function handleOAuthLogin(providerType: string) {
// 如果 sessionStorage 中没有 redirectPath(用户直接点击登录而非被守卫拦截),
// 则不设置,让 AuthCallback 使用默认跳转逻辑
window.location.href = getApiUrl(`/api/oauth/${providerType}/authorize`)
const authorizeUrl = new URL(
getApiUrl(`/api/oauth/${providerType}/authorize`),
window.location.origin,
)
authorizeUrl.searchParams.set('client_device_id', getClientDeviceId())
window.location.href = authorizeUrl.toString()
}
function handleSwitchToRegister() {
@@ -151,15 +151,12 @@
<Input
:id="`pwd-${formNonce}`"
v-model="formData.password"
type="text"
autocomplete="one-time-code"
data-form-type="other"
data-lpignore="true"
data-1p-ignore="true"
masked
autocomplete="new-password"
disable-autofill
:name="`pwd-${formNonce}`"
:placeholder="getPasswordPolicyPlaceholder(props.passwordPolicyLevel)"
required
class="-webkit-text-security-disc"
:disabled="isLoading"
/>
<p
@@ -182,17 +179,20 @@
<Input
:id="`pwd-confirm-${formNonce}`"
v-model="formData.confirmPassword"
type="text"
autocomplete="one-time-code"
data-form-type="other"
data-lpignore="true"
data-1p-ignore="true"
masked
autocomplete="new-password"
disable-autofill
:name="`pwd-confirm-${formNonce}`"
placeholder="再次输入密码"
required
class="-webkit-text-security-disc"
:disabled="isLoading"
/>
<p
v-if="formData.confirmPassword && formData.password !== formData.confirmPassword"
class="text-xs text-destructive"
>
两次输入的密码不一致
</p>
</div>
</form>
@@ -457,42 +457,41 @@ function confirmCustomInput(index: number) {
}
function updateCacheCreation(index: number, value: string | number) {
const numValue = parseFloatInput(value)
if (numValue > 0) {
cacheManuallySet[index] = { ...cacheManuallySet[index], creation: true }
localTiers.value[index].cache_creation_price_per_1m = numValue
} else {
if (value === '' || value === null || value === undefined) {
// 清空时恢复自动计算
cacheManuallySet[index] = { ...cacheManuallySet[index], creation: false }
localTiers.value[index].cache_creation_price_per_1m = undefined
} else {
const numValue = parseFloatInput(value)
cacheManuallySet[index] = { ...cacheManuallySet[index], creation: true }
localTiers.value[index].cache_creation_price_per_1m = numValue
}
syncToParent()
}
function updateCacheRead(index: number, value: string | number) {
const numValue = parseFloatInput(value)
if (numValue > 0) {
cacheManuallySet[index] = { ...cacheManuallySet[index], read: true }
localTiers.value[index].cache_read_price_per_1m = numValue
} else {
if (value === '' || value === null || value === undefined) {
// 清空时恢复自动计算
cacheManuallySet[index] = { ...cacheManuallySet[index], read: false }
localTiers.value[index].cache_read_price_per_1m = undefined
} else {
const numValue = parseFloatInput(value)
cacheManuallySet[index] = { ...cacheManuallySet[index], read: true }
localTiers.value[index].cache_read_price_per_1m = numValue
}
syncToParent()
}
function updateCache1h(index: number, value: string | number) {
const numValue = parseFloatInput(value)
const tier = localTiers.value[index]
if (numValue > 0) {
// 手动设置 1 小时缓存创建价格
cacheManuallySet[index] = { ...cacheManuallySet[index], cache1h: true }
tier.cache_ttl_pricing = [{ ttl_minutes: 60, cache_creation_price_per_1m: numValue }]
} else {
if (value === '' || value === null || value === undefined) {
// 清空时恢复自动计算
cacheManuallySet[index] = { ...cacheManuallySet[index], cache1h: false }
tier.cache_ttl_pricing = undefined
} else {
const numValue = parseFloatInput(value)
cacheManuallySet[index] = { ...cacheManuallySet[index], cache1h: true }
tier.cache_ttl_pricing = [{ ttl_minutes: 60, cache_creation_price_per_1m: numValue }]
}
syncToParent()
}
@@ -107,15 +107,15 @@
<div class="flex items-center gap-1.5">
<span class="text-xs font-medium truncate">{{ key.key_name || '未命名' }}</span>
<Badge
v-if="isOAuthInvalid(key)"
variant="destructive"
class="text-[10px] px-1 py-0 h-4 shrink-0"
>OAuth失效</Badge>
<Badge
v-else
variant="outline"
class="text-[10px] px-1 py-0 h-4 shrink-0"
>{{ normalizeAuthTypeLabel(key.auth_type) }}</Badge>
<Badge
v-if="getStatusBadgeLabel(key)"
variant="destructive"
class="text-[10px] px-1 py-0 h-4 shrink-0"
:title="getStatusBadgeTitle(key)"
>{{ getStatusBadgeLabel(key) }}</Badge>
<Badge
v-if="key.oauth_plan_type"
variant="outline"
@@ -125,13 +125,8 @@
v-if="getOAuthOrgBadge(key)"
variant="secondary"
class="text-[10px] px-1 py-0 h-4 shrink-0"
:title="getOAuthOrgBadge(key)?.id"
:title="getOAuthOrgBadge(key)?.title"
>{{ getOAuthOrgBadge(key)?.label }}</Badge>
<Badge
v-if="isBannedKey(key)"
variant="destructive"
class="text-[10px] px-1 py-0 h-4 shrink-0"
>封号</Badge>
</div>
<div class="flex items-center gap-1.5 mt-0.5 text-[11px] text-muted-foreground flex-wrap">
<span :class="key.is_active ? '' : 'text-destructive'">{{ key.is_active ? '启用' : '禁用' }}</span>
@@ -284,6 +279,12 @@ import { exportKey, refreshProviderQuota } from '@/api/endpoints/keys'
import { refreshProviderOAuth } from '@/api/endpoints/provider_oauth'
import { useProxyNodesStore } from '@/stores/proxy-nodes'
import { getOAuthOrgBadge } from '@/utils/oauthIdentity'
import {
getAccountStatusDisplay,
getAccountStatusTitle,
getOAuthStatusDisplay,
getOAuthStatusTitle,
} from '@/utils/providerKeyStatus'
type QuickSelectorValue =
| 'banned'
@@ -321,12 +322,12 @@ const emit = defineEmits<{
}>()
const QUICK_SELECT_OPTIONS: Array<{ value: QuickSelectorValue; label: string }> = [
{ value: 'banned', label: '已封号' },
{ value: 'banned', label: '账号异常' },
{ value: 'no_5h_limit', label: '无5H限额' },
{ value: 'no_weekly_limit', label: '无周限额' },
{ value: 'plan_free', label: '全部 Free' },
{ value: 'plan_team', label: '全部 Team' },
{ value: 'oauth_invalid', label: 'OAuth 失效' },
{ value: 'oauth_invalid', label: 'Token 异常' },
{ value: 'proxy_unset', label: '未配置代理' },
{ value: 'proxy_set', label: '已配置独立代理' },
{ value: 'disabled', label: '已禁用' },
@@ -426,25 +427,25 @@ function normalizeAuthTypeLabel(authType: string): string {
return 'API Key'
}
function isBannedKey(key: PoolKeyDetail): boolean {
const reason = normalizeText(key.oauth_invalid_reason)
if (reason && /(banned|forbidden|blocked|suspend|封|禁|受限)/.test(reason)) return true
if (Array.isArray(key.scheduling_reasons)) {
return key.scheduling_reasons.some((item) => {
const code = normalizeText(item.code)
return code === 'account_banned' || code === 'account_forbidden' || code === 'account_blocked'
})
}
return false
function getStatusBadgeLabel(key: PoolKeyDetail): string | null {
const account = getAccountStatusDisplay(key)
if (account.blocked && account.label) return account.label
const oauth = getOAuthStatusDisplay(key, 0)
if (oauth?.isInvalid) return 'Token 失效'
if (oauth?.isExpired) return 'Token 过期'
return null
}
function isOAuthInvalid(key: PoolKeyDetail): boolean {
if (normalizeText(key.auth_type) !== 'oauth') return false
if (key.oauth_invalid_at != null || normalizeText(key.oauth_invalid_reason)) return true
if (typeof key.oauth_expires_at === 'number' && key.oauth_expires_at > 0) {
return key.oauth_expires_at * 1000 <= Date.now()
}
return false
function getStatusBadgeTitle(key: PoolKeyDetail): string {
const label = getStatusBadgeLabel(key)
if (!label) return ''
const accountTitle = getAccountStatusTitle(key)
if (accountTitle) return accountTitle
const oauthTitle = getOAuthStatusTitle(key, 0)
return oauthTitle || label
}
function formatRelativeTime(value: string): string {
@@ -28,7 +28,7 @@
<div class="space-y-0.5">
<span class="text-sm font-medium">主动探测</span>
<p class="text-xs text-muted-foreground">
定期检查 Key 可用性,提前发现异常
按固定间隔主动刷新 Key 的账号状态与额度
</p>
</div>
<Switch
@@ -57,9 +57,9 @@
</div>
<div class="flex items-center justify-between p-3 border rounded-lg bg-muted/50">
<div class="space-y-0.5">
<span class="text-sm font-medium">封号自动清除</span>
<span class="text-sm font-medium">异常自动清除</span>
<p class="text-xs text-muted-foreground">
检测到账号被封禁时自动从号池中移除
仅在检测到不可恢复的账号异常时自动从号池中移除,不处理纯 Token 失效
</p>
</div>
<Switch
@@ -192,7 +192,6 @@ async function handleTestModel(modelName: string) {
model_name: modelName,
api_key_id: props.keyId,
api_format: 'gemini:chat',
message: 'hello',
})
if (result.success) {
@@ -0,0 +1,299 @@
<template>
<div>
<!-- Group node -->
<div
v-if="modelValue.kind === 'group'"
class="rounded-md border p-2"
:class="[
nested ? 'bg-muted/30 border-dashed' : 'bg-muted/10 border-border',
]"
>
<!-- Group header: mode selector + actions -->
<div class="flex items-center gap-1.5 mb-2">
<Button
v-if="!nested"
variant="ghost"
size="icon"
class="h-6 w-6 shrink-0 text-muted-foreground"
title="转回单条件"
@click="convertToLeaf"
>
<ListFilter class="w-3.5 h-3.5" />
</Button>
<Select
:model-value="modelValue.mode"
@update:model-value="(value: string) => updateGroupMode(value as ConditionGroupMode)"
>
<SelectTrigger class="w-[86px] h-6 text-xs font-medium shrink-0">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="all">
AND
</SelectItem>
<SelectItem value="any">
OR
</SelectItem>
</SelectContent>
</Select>
<div class="flex-1" />
<Button
variant="ghost"
size="sm"
class="h-6 px-1.5 text-xs text-muted-foreground"
title="添加条件"
@click="addLeafChild"
>
<Plus class="w-3 h-3 mr-0.5" />
条件
</Button>
<Button
v-if="!nested"
variant="ghost"
size="sm"
class="h-6 px-1.5 text-xs text-muted-foreground"
@click="addGroupChild(modelValue.mode === 'all' ? 'any' : 'all')"
>
<Plus class="w-3 h-3 mr-0.5" />
子组
</Button>
<Button
v-if="removable && nested"
variant="ghost"
size="icon"
class="h-6 w-6 shrink-0 text-muted-foreground hover:text-destructive"
@click="emit('remove')"
>
<X class="w-3 h-3" />
</Button>
</div>
<!-- Children with logic connector labels between them -->
<div class="space-y-0">
<template
v-for="(child, index) in modelValue.children"
:key="index"
>
<!-- Logic connector label between children -->
<div
v-if="index > 0"
class="flex items-center gap-2 py-0.5 pl-2"
>
<span
class="text-[10px] font-semibold px-1.5 py-0.5 rounded"
:class="modelValue.mode === 'all'
? 'bg-blue-500/15 text-blue-600 dark:text-blue-400'
: 'bg-amber-500/15 text-amber-600 dark:text-amber-400'"
>
{{ modelValue.mode === 'all' ? 'AND' : 'OR' }}
</span>
<div class="flex-1 border-t border-dashed border-muted-foreground/20" />
</div>
<EndpointConditionEditor
:model-value="child"
:path-hint="pathHint"
nested
removable
@update:model-value="(next) => updateChild(index, next)"
@remove="removeChild(index)"
/>
</template>
</div>
</div>
<!-- Leaf node -->
<div
v-else
class="flex flex-wrap items-center gap-1.5 rounded-md px-2 py-1.5"
:class="nested ? 'bg-background/60' : 'bg-muted/10 border border-border'"
>
<!-- Toggle to group mode (icon button at the start, top-level only) -->
<Button
v-if="!nested"
variant="ghost"
size="icon"
class="h-7 w-7 shrink-0 text-muted-foreground"
title="转为组合条件 (AND/OR)"
@click="convertToGroup('all')"
>
<ListFilter class="w-3.5 h-3.5" />
</Button>
<Select
:model-value="modelValue.source"
@update:model-value="(value: string) => updateLeafField('source', value)"
>
<SelectTrigger class="w-[96px] h-7 text-xs shrink-0">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="current">
Current
</SelectItem>
<SelectItem value="original">
Original
</SelectItem>
</SelectContent>
</Select>
<Input
:model-value="modelValue.path"
:placeholder="pathHint || '字段路径'"
size="sm"
class="flex-1 min-w-[120px] h-7 text-xs"
@update:model-value="(value) => updateLeafField('path', value)"
/>
<Select
:model-value="modelValue.op"
@update:model-value="(value: string) => updateLeafField('op', value)"
>
<SelectTrigger class="w-[110px] h-7 text-xs shrink-0">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem
v-for="option in CONDITION_OP_OPTIONS"
:key="option.value"
:value="option.value"
>
{{ option.label }}
</SelectItem>
</SelectContent>
</Select>
<Input
v-if="isConditionValueRequired(modelValue.op)"
:model-value="modelValue.value"
:placeholder="getConditionValuePlaceholder(modelValue.op)"
size="sm"
class="flex-1 min-w-[120px] h-7 text-xs"
@update:model-value="(value) => updateLeafField('value', value)"
/>
<Button
v-if="removable"
variant="ghost"
size="icon"
class="h-7 w-7 shrink-0 text-muted-foreground hover:text-destructive"
@click="emit('remove')"
>
<X class="w-3 h-3" />
</Button>
</div>
</div>
</template>
<script setup lang="ts">
import { Button, Input, Select, SelectContent, SelectItem, SelectTrigger, SelectValue } from '@/components/ui'
import { ListFilter, Plus, X } from 'lucide-vue-next'
import type { BodyRuleConditionOp } from '@/api/endpoints'
import {
CONDITION_OP_OPTIONS,
cloneEditableCondition,
createConditionGroup,
createEmptyConditionLeaf,
getConditionValuePlaceholder,
isConditionValueRequired,
type ConditionGroupMode,
type ConditionSource,
type EditableConditionGroup,
type EditableConditionLeaf,
type EditableConditionNode,
} from './endpoint-rule-condition'
const props = withDefaults(defineProps<{
modelValue: EditableConditionNode
pathHint?: string
nested?: boolean
removable?: boolean
}>(), {
pathHint: '',
nested: false,
removable: false,
})
const emit = defineEmits<{
'update:modelValue': [value: EditableConditionNode]
remove: []
}>()
defineOptions({ name: 'EndpointConditionEditor' })
function updateLeafField(field: keyof EditableConditionLeaf, rawValue: string): void {
if (props.modelValue.kind !== 'leaf') return
const next = cloneEditableCondition(props.modelValue) as EditableConditionLeaf
if (field === 'op') {
next.op = rawValue as BodyRuleConditionOp
if (!isConditionValueRequired(next.op)) next.value = ''
} else if (field === 'source') {
next.source = rawValue as ConditionSource
} else if (field === 'path' || field === 'value') {
next[field] = rawValue
}
emit('update:modelValue', next)
}
function updateGroupMode(mode: ConditionGroupMode): void {
if (props.modelValue.kind !== 'group') return
const next = cloneEditableCondition(props.modelValue) as EditableConditionGroup
next.mode = mode
emit('update:modelValue', next)
}
function convertToLeaf(): void {
emit('update:modelValue', createEmptyConditionLeaf())
}
function convertToGroup(mode: ConditionGroupMode): void {
const seed = props.modelValue.kind === 'leaf'
? cloneEditableCondition(props.modelValue)
: createEmptyConditionLeaf()
emit('update:modelValue', createConditionGroup(mode, [seed]))
}
function addLeafChild(): void {
if (props.modelValue.kind !== 'group') return
const next = cloneEditableCondition(props.modelValue) as EditableConditionGroup
next.children.push(createEmptyConditionLeaf())
emit('update:modelValue', next)
}
function addGroupChild(mode: ConditionGroupMode): void {
if (props.modelValue.kind !== 'group') return
const next = cloneEditableCondition(props.modelValue) as EditableConditionGroup
next.children.push(createConditionGroup(mode))
emit('update:modelValue', next)
}
function updateChild(index: number, child: EditableConditionNode): void {
if (props.modelValue.kind !== 'group') return
const next = cloneEditableCondition(props.modelValue) as EditableConditionGroup
next.children[index] = child
emit('update:modelValue', next)
}
function removeChild(index: number): void {
if (props.modelValue.kind !== 'group') return
const next = cloneEditableCondition(props.modelValue) as EditableConditionGroup
next.children.splice(index, 1)
// When all children are removed, if top-level group, emit remove to let parent clear the condition
if (next.children.length === 0) {
if (!props.nested) {
emit('remove')
} else {
next.children.push(createEmptyConditionLeaf())
emit('update:modelValue', next)
}
return
}
emit('update:modelValue', next)
}
</script>
@@ -105,7 +105,7 @@
<div class="space-y-1.5">
<Label class="text-xs text-muted-foreground">自定义路径</Label>
<Input
:model-value="getEndpointEditState(endpoint.id)?.path ?? (endpoint.custom_path || '')"
:model-value="getDisplayedPath(endpoint)"
:placeholder="getDefaultPath(endpoint.api_format, endpoint.base_url) || '留空使用默认'"
:disabled="isFixedProvider"
@update:model-value="(v) => updateEndpointField(endpoint.id, 'path', v)"
@@ -228,105 +228,126 @@
<span>拖拽左侧手柄可调整规则执行顺序</span>
</div>
<!-- 请求头规则列表 - 主题色边框 -->
<div
<template
v-for="(rule, index) in getEndpointEditRules(endpoint.id)"
:key="`header-${index}`"
class="flex items-center gap-1.5 px-2 py-1.5 rounded-md border-l-4 border-primary/60 bg-muted/30"
:class="[
isHeaderRuleDragging(endpoint.id, index) ? 'opacity-60 border-primary bg-primary/5' : '',
isHeaderRuleDragOver(endpoint.id, index) ? 'ring-1 ring-primary/40 bg-primary/10' : ''
]"
@dragover.prevent="handleHeaderRuleDragOver(endpoint.id, index)"
@dragleave="handleHeaderRuleDragLeave(endpoint.id, index)"
@drop.prevent="handleHeaderRuleDrop(endpoint.id, index)"
>
<button
type="button"
class="h-7 w-6 shrink-0 inline-flex items-center justify-center rounded-sm text-muted-foreground/60 hover:text-muted-foreground hover:bg-muted cursor-grab active:cursor-grabbing"
title="拖拽排序"
draggable="true"
@dragstart="(e) => handleHeaderRuleDragStart(endpoint.id, index, e)"
@dragend="() => handleHeaderRuleDragEnd(endpoint.id)"
<div
class="flex items-center gap-1.5 px-2 py-1.5 rounded-md border-l-4 border-primary/60 bg-muted/30"
:class="[
isHeaderRuleDragging(endpoint.id, index) ? 'opacity-60 border-primary bg-primary/5' : '',
isHeaderRuleDragOver(endpoint.id, index) ? 'ring-1 ring-primary/40 bg-primary/10' : ''
]"
@dragover.prevent="handleHeaderRuleDragOver(endpoint.id, index)"
@dragleave="handleHeaderRuleDragLeave(endpoint.id, index)"
@drop.prevent="handleHeaderRuleDrop(endpoint.id, index)"
>
<GripVertical class="w-3.5 h-3.5" />
</button>
<span
class="text-[10px] font-semibold text-primary shrink-0"
title="请求头"
>H</span>
<Select
:model-value="rule.action"
:open="ruleSelectOpen[`${endpoint.id}-${index}`]"
@update:model-value="(v) => updateEndpointRuleAction(endpoint.id, index, v as 'set' | 'drop' | 'rename')"
@update:open="(v) => handleRuleSelectOpen(endpoint.id, index, v)"
>
<SelectTrigger class="w-[88px] h-7 text-xs shrink-0">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="set">
覆写
</SelectItem>
<SelectItem value="drop">
删除
</SelectItem>
<SelectItem value="rename">
重命名
</SelectItem>
</SelectContent>
</Select>
<template v-if="rule.action === 'set'">
<Input
:model-value="rule.key"
placeholder="名称"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'key', v)"
/>
<span class="text-muted-foreground text-xs">=</span>
<Input
:model-value="rule.value"
placeholder="值"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'value', v)"
/>
</template>
<template v-else-if="rule.action === 'drop'">
<Input
:model-value="rule.key"
placeholder="要删除的名称"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'key', v)"
/>
</template>
<template v-else-if="rule.action === 'rename'">
<Input
:model-value="rule.from"
placeholder="原名"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'from', v)"
/>
<span class="text-muted-foreground text-xs">→</span>
<Input
:model-value="rule.to"
placeholder="新名"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'to', v)"
/>
</template>
<Button
variant="ghost"
size="icon"
class="h-7 w-7 shrink-0"
@click="removeEndpointRule(endpoint.id, index)"
>
<X class="w-3 h-3" />
</Button>
</div>
<button
type="button"
class="h-7 w-6 shrink-0 inline-flex items-center justify-center rounded-sm text-muted-foreground/60 hover:text-muted-foreground hover:bg-muted cursor-grab active:cursor-grabbing"
title="拖拽排序"
draggable="true"
@dragstart="(e) => handleHeaderRuleDragStart(endpoint.id, index, e)"
@dragend="() => handleHeaderRuleDragEnd(endpoint.id)"
>
<GripVertical class="w-3.5 h-3.5" />
</button>
<span
class="text-[10px] font-semibold text-primary shrink-0"
title="请求头"
>H</span>
<Select
:model-value="rule.action"
:open="ruleSelectOpen[`${endpoint.id}-${index}`]"
@update:model-value="(v) => updateEndpointRuleAction(endpoint.id, index, v as 'set' | 'drop' | 'rename')"
@update:open="(v) => handleRuleSelectOpen(endpoint.id, index, v)"
>
<SelectTrigger class="w-[88px] h-7 text-xs shrink-0">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="set">
覆写
</SelectItem>
<SelectItem value="drop">
删除
</SelectItem>
<SelectItem value="rename">
重命名
</SelectItem>
</SelectContent>
</Select>
<Button
variant="ghost"
size="icon"
class="h-7 w-7 shrink-0"
:class="rule.condition ? 'text-primary' : ''"
title="条件触发"
@click="toggleEndpointRuleCondition(endpoint.id, index)"
>
<Filter class="w-3 h-3" />
</Button>
<template v-if="rule.action === 'set'">
<Input
:model-value="rule.key"
placeholder="名称"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'key', v)"
/>
<span class="text-muted-foreground text-xs">=</span>
<Input
:model-value="rule.value"
placeholder="值"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'value', v)"
/>
</template>
<template v-else-if="rule.action === 'drop'">
<Input
:model-value="rule.key"
placeholder="要删除的名称"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'key', v)"
/>
</template>
<template v-else-if="rule.action === 'rename'">
<Input
:model-value="rule.from"
placeholder="原名"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'from', v)"
/>
<span class="text-muted-foreground text-xs">→</span>
<Input
:model-value="rule.to"
placeholder="新名"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointRuleField(endpoint.id, index, 'to', v)"
/>
</template>
<Button
variant="ghost"
size="icon"
class="h-7 w-7 shrink-0"
@click="removeEndpointRule(endpoint.id, index)"
>
<X class="w-3 h-3" />
</Button>
</div>
<EndpointConditionEditor
v-if="rule.condition"
:model-value="rule.condition"
path-hint="请求体字段路径"
removable
@update:model-value="(condition) => updateEndpointRuleCondition(endpoint.id, index, condition)"
@remove="clearEndpointRuleCondition(endpoint.id, index)"
/>
</template>
<div
v-if="getEndpointEditBodyRules(endpoint.id).length > 0"
@@ -397,14 +418,15 @@
<code>exists</code> <code>not_exists</code> 字段存在性<br>
<code>in</code> 在列表中(值填 <code>["a","b"]</code>)<br>
<code>type_is</code> 类型判断(string/number/boolean/array/object/null)<br>
条件路径支持 <code>$item.xxx</code> 引用通配符当前元素
条件路径支持 <code>$item.xxx</code> 引用通配符当前元素<br>
可切换 <code>Current</code>/<code>Original</code> 数据源,并支持 <code>ALL</code>/<code>ANY</code> 组合条件
</div>
</div>
<div class="text-muted-foreground">
规则按顺序执行,前面的修改对后续规则可见。
</div>
<div class="text-muted-foreground">
规则在格式转换之后执行,路径需按目标提供商的请求体结构填写。
规则默认在格式转换后按目标提供商结构匹配;条件切到 <code>Original</code> 时则按客户端原始请求体匹配。
</div>
</div>
</PopoverContent>
@@ -474,7 +496,7 @@
variant="ghost"
size="icon"
class="h-7 w-7 shrink-0"
:class="rule.conditionEnabled ? 'text-primary' : ''"
:class="rule.condition ? 'text-primary' : ''"
title="条件触发"
@click="toggleBodyRuleCondition(endpoint.id, index)"
>
@@ -638,80 +660,14 @@
<X class="w-3 h-3" />
</Button>
</div>
<!-- 条件编辑行 -->
<div
v-if="rule.conditionEnabled"
class="flex items-center gap-1.5 px-2 py-1 ml-6 rounded-md bg-muted/20"
>
<span class="text-[10px] font-semibold text-muted-foreground shrink-0">IF</span>
<Input
:model-value="rule.conditionPath"
:placeholder="rule.path?.includes('[*]') || rule.path?.match(/\[\d+-\d+\]/) ? '$item.字段名' : '字段路径'"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointBodyRuleField(endpoint.id, index, 'conditionPath', v)"
/>
<Select
:model-value="rule.conditionOp"
@update:model-value="(v: string) => updateEndpointBodyRuleField(endpoint.id, index, 'conditionOp', v)"
>
<SelectTrigger class="w-[100px] h-7 text-xs shrink-0">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="eq">
等于
</SelectItem>
<SelectItem value="neq">
不等于
</SelectItem>
<SelectItem value="gt">
大于
</SelectItem>
<SelectItem value="lt">
小于
</SelectItem>
<SelectItem value="gte">
大于等于
</SelectItem>
<SelectItem value="lte">
小于等于
</SelectItem>
<SelectItem value="starts_with">
开头匹配
</SelectItem>
<SelectItem value="ends_with">
结尾匹配
</SelectItem>
<SelectItem value="contains">
包含
</SelectItem>
<SelectItem value="matches">
正则匹配
</SelectItem>
<SelectItem value="exists">
存在
</SelectItem>
<SelectItem value="not_exists">
不存在
</SelectItem>
<SelectItem value="in">
在列表中
</SelectItem>
<SelectItem value="type_is">
类型是
</SelectItem>
</SelectContent>
</Select>
<Input
v-if="rule.conditionOp !== 'exists' && rule.conditionOp !== 'not_exists'"
:model-value="rule.conditionValue"
:placeholder="rule.conditionOp === 'in' ? '[&quot;a&quot;, &quot;b&quot;]' : rule.conditionOp === 'type_is' ? 'string/number/boolean/...' : '值'"
size="sm"
class="flex-1 min-w-0 h-7 text-xs"
@update:model-value="(v) => updateEndpointBodyRuleField(endpoint.id, index, 'conditionValue', v)"
/>
</div>
<EndpointConditionEditor
v-if="rule.condition"
:model-value="rule.condition"
:path-hint="getBodyRuleConditionPathPlaceholder(rule.path)"
removable
@update:model-value="(condition) => updateEndpointBodyRuleCondition(endpoint.id, index, condition)"
@remove="clearEndpointBodyRuleCondition(endpoint.id, index)"
/>
</template>
</div>
</CollapsibleContent>
@@ -839,6 +795,7 @@ import { useToast } from '@/composables/useToast'
import { parseApiError } from '@/utils/errorParser'
import { log } from '@/utils/logger'
import AlertDialog from '@/components/common/AlertDialog.vue'
import EndpointConditionEditor from './EndpointConditionEditor.vue'
import {
createEndpoint,
getDefaultBodyRules,
@@ -850,11 +807,18 @@ import {
type BodyRule,
type BodyRuleRegexReplace,
type BodyRuleNameStyle,
type BodyRuleCondition,
type BodyRuleConditionOp,
} from '@/api/endpoints'
import { adminApi } from '@/api/admin'
import { formatApiFormat } from '@/api/endpoints/types/api-format'
import {
conditionEquals,
conditionToEditable,
createEmptyConditionLeaf,
editableConditionToApi,
getBodyRuleConditionPathPlaceholder,
type EditableConditionNode,
validateEditableCondition,
} from './endpoint-rule-condition'
// 编辑用的规则类型(统一的可编辑结构)
interface EditableRule {
@@ -863,6 +827,7 @@ interface EditableRule {
value: string // set 用
from: string // rename 用
to: string // rename 用
condition: EditableConditionNode | null
}
// 编辑用的请求体规则类型
@@ -879,10 +844,7 @@ interface EditableBodyRule {
replacement: string // regex_replace 用
flags: string // regex_replace 用(i/m/s)
style: string // name_style 用(snake_case/camelCase/PascalCase/kebab-case/capitalize)
conditionEnabled: boolean // 是否启用条件
conditionPath: string
conditionOp: string
conditionValue: string // JSON 格式字符串(保存时 parse)
condition: EditableConditionNode | null
}
// 端点编辑状态(仅 URL、路径、规则,格式转换是直接保存的)
@@ -1197,7 +1159,7 @@ function restoreOriginalPlaceholder(value: unknown): unknown {
}
return value
}
if (Array.isArray(value)) return value.map(restoreOriginalPlaceholder)
if (Array.isArray(value)) return value.map(item => restoreOriginalPlaceholder(item))
if (value !== null && typeof value === 'object') {
const result: Record<string, unknown> = {}
for (const [k, v] of Object.entries(value as Record<string, unknown>)) result[k] = restoreOriginalPlaceholder(v)
@@ -1300,7 +1262,7 @@ async function loadDefaultBodyRulesForFormat(apiFormat: string, force = false):
} catch (error: unknown) {
defaultBodyRulesByFormat.value[cacheKey] = []
defaultBodyRulesLoaded.value[cacheKey] = true
log.warn('加载默认请求体规则失败', apiFormat, error)
log.warn('加载默认请求体规则失败', { apiFormat, error })
return []
} finally {
loadingDefaultBodyRulesByFormat.value[cacheKey] = false
@@ -1314,10 +1276,19 @@ async function preloadDefaultBodyRules(endpoints: ProviderEndpoint[]): Promise<v
// 获取指定 API 格式的默认路径
function getDefaultPath(apiFormat: string, baseUrl?: string): string {
const providerType = (props.provider?.provider_type || '').toLowerCase()
if (providerType === 'vertex_ai') {
if (apiFormat === 'gemini:chat') {
return '/v1/publishers/google/models/{model}:{action}'
}
if (apiFormat === 'claude:chat') {
return '/v1/projects/{project_id}/locations/{region}/publishers/anthropic/models/{model}:{action}'
}
}
const format = apiFormats.value.find(f => f.value === apiFormat)
const defaultPath = format?.default_path || ''
// Codex 端点使用 /responses 而非 /v1/responses
const providerType = (props.provider?.provider_type || '').toLowerCase()
const isCodex = providerType
? providerType === 'codex'
: (!!baseUrl && isCodexUrl(baseUrl))
@@ -1327,6 +1298,13 @@ function getDefaultPath(apiFormat: string, baseUrl?: string): string {
return defaultPath
}
function getDisplayedPath(endpoint: ProviderEndpoint): string {
if (isFixedProvider.value) {
return getDefaultPath(endpoint.api_format, endpoint.base_url)
}
return getEndpointEditState(endpoint.id)?.path ?? (endpoint.custom_path || '')
}
// 判断是否是 Codex OAuth 端点
function isCodexUrl(baseUrl: string): boolean {
const url = baseUrl.replace(/\/+$/, '')
@@ -1346,57 +1324,62 @@ function getEndpointUpstreamStreamPolicy(endpoint: ProviderEndpoint): string {
return 'auto'
}
function emptyHeaderRule(): EditableRule {
return { action: 'set', key: '', value: '', from: '', to: '', condition: null }
}
function emptyBodyRule(action: BodyRuleAction = 'set'): EditableBodyRule {
return {
action,
path: '',
value: '',
from: '',
to: '',
index: '',
pattern: '',
replacement: '',
flags: '',
style: '',
condition: null,
}
}
// 初始化端点的编辑状态
function initEndpointEditState(endpoint: ProviderEndpoint): EndpointEditState {
const rules: EditableRule[] = []
if (endpoint.header_rules && endpoint.header_rules.length > 0) {
for (const rule of endpoint.header_rules) {
if (rule.action === 'set') {
rules.push({ action: 'set', key: rule.key, value: rule.value || '', from: '', to: '' })
rules.push({ ...emptyHeaderRule(), action: 'set', key: rule.key, value: rule.value || '', condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'drop') {
rules.push({ action: 'drop', key: rule.key, value: '', from: '', to: '' })
rules.push({ ...emptyHeaderRule(), action: 'drop', key: rule.key, condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'rename') {
rules.push({ action: 'rename', key: '', value: '', from: rule.from, to: rule.to })
rules.push({ ...emptyHeaderRule(), action: 'rename', from: rule.from, to: rule.to, condition: conditionToEditable(rule.condition) })
}
}
}
const emptyBodyRule = (): Omit<EditableBodyRule, 'action'> => ({
path: '', value: '', from: '', to: '', index: '', pattern: '', replacement: '', flags: '', style: '',
conditionEnabled: false, conditionPath: '', conditionOp: 'eq', conditionValue: '',
})
const bodyRules: EditableBodyRule[] = []
if (endpoint.body_rules && endpoint.body_rules.length > 0) {
for (const rule of endpoint.body_rules) {
// 提取 condition 信息
const conditionFields = rule.condition ? {
conditionEnabled: true,
conditionPath: rule.condition.path || '',
conditionOp: rule.condition.op || 'eq',
conditionValue: rule.condition.value !== undefined
? (typeof rule.condition.value === 'string' ? rule.condition.value : JSON.stringify(rule.condition.value))
: '',
} : {}
if (rule.action === 'set') {
const { value } = initBodyRuleSetValueForEditor(rule.value)
bodyRules.push({ ...emptyBodyRule(), action: 'set', path: rule.path, value, ...conditionFields })
bodyRules.push({ ...emptyBodyRule('set'), path: rule.path, value, condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'drop') {
bodyRules.push({ ...emptyBodyRule(), action: 'drop', path: rule.path, ...conditionFields })
bodyRules.push({ ...emptyBodyRule('drop'), path: rule.path, condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'rename') {
bodyRules.push({ ...emptyBodyRule(), action: 'rename', from: rule.from, to: rule.to, ...conditionFields })
bodyRules.push({ ...emptyBodyRule('rename'), from: rule.from, to: rule.to, condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'append') {
// 前端将 append 统一展示为 insert(index 留空),保存时再根据 index 是否为空转回 append
const { value } = initBodyRuleSetValueForEditor(rule.value)
bodyRules.push({ ...emptyBodyRule(), action: 'insert', path: rule.path || '', value, index: '', ...conditionFields })
bodyRules.push({ ...emptyBodyRule('insert'), path: rule.path || '', value, index: '', condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'insert') {
const { value } = initBodyRuleSetValueForEditor(rule.value)
bodyRules.push({ ...emptyBodyRule(), action: 'insert', path: rule.path || '', value, index: String(rule.index ?? ''), ...conditionFields })
bodyRules.push({ ...emptyBodyRule('insert'), path: rule.path || '', value, index: String(rule.index ?? ''), condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'regex_replace') {
bodyRules.push({ ...emptyBodyRule(), action: 'regex_replace', path: rule.path || '', pattern: rule.pattern || '', replacement: rule.replacement || '', flags: rule.flags || '', ...conditionFields })
bodyRules.push({ ...emptyBodyRule('regex_replace'), path: rule.path || '', pattern: rule.pattern || '', replacement: rule.replacement || '', flags: rule.flags || '', condition: conditionToEditable(rule.condition) })
} else if (rule.action === 'name_style') {
bodyRules.push({ ...emptyBodyRule(), action: 'name_style', path: rule.path || '', style: rule.style || 'capitalize', ...conditionFields })
bodyRules.push({ ...emptyBodyRule('name_style'), path: rule.path || '', style: rule.style || 'capitalize', condition: conditionToEditable(rule.condition) })
}
}
}
@@ -1447,7 +1430,7 @@ function getEndpointEditRules(endpointId: string): EditableRule[] {
// 添加规则(同时自动展开折叠)
function handleAddEndpointRule(endpointId: string) {
const rules = getEndpointEditRules(endpointId)
rules.push({ action: 'set', key: '', value: '', from: '', to: '' })
rules.push(emptyHeaderRule())
// 自动展开折叠
endpointRulesExpanded.value[endpointId] = true
}
@@ -1464,11 +1447,8 @@ function removeEndpointRule(endpointId: string, index: number) {
function updateEndpointRuleAction(endpointId: string, index: number, action: 'set' | 'drop' | 'rename') {
const rules = getEndpointEditRules(endpointId)
if (rules[index]) {
rules[index].action = action
rules[index].key = ''
rules[index].value = ''
rules[index].from = ''
rules[index].to = ''
const currentCondition = rules[index].condition
rules[index] = { ...emptyHeaderRule(), action, condition: currentCondition }
}
}
@@ -1480,6 +1460,27 @@ function updateEndpointRuleField(endpointId: string, index: number, field: 'key'
}
}
function toggleEndpointRuleCondition(endpointId: string, index: number) {
const rules = getEndpointEditRules(endpointId)
if (rules[index]) {
rules[index].condition = rules[index].condition ? null : createEmptyConditionLeaf()
}
}
function updateEndpointRuleCondition(endpointId: string, index: number, condition: EditableConditionNode) {
const rules = getEndpointEditRules(endpointId)
if (rules[index]) {
rules[index].condition = condition
}
}
function clearEndpointRuleCondition(endpointId: string, index: number) {
const rules = getEndpointEditRules(endpointId)
if (rules[index]) {
rules[index].condition = null
}
}
// 验证规则 key(针对特定端点)
function validateRuleKeyForEndpoint(endpointId: string, key: string, index: number): string | null {
const trimmedKey = key.trim().toLowerCase()
@@ -1587,7 +1588,7 @@ function getEndpointEditBodyRules(endpointId: string): EditableBodyRule[] {
// 添加请求体规则(同时自动展开折叠)
function handleAddEndpointBodyRule(endpointId: string) {
const rules = getEndpointEditBodyRules(endpointId)
rules.push({ action: 'set', path: '', value: '', from: '', to: '', index: '', pattern: '', replacement: '', flags: '', style: '', conditionEnabled: false, conditionPath: '', conditionOp: 'eq', conditionValue: '' })
rules.push(emptyBodyRule('set'))
// 自动展开折叠
endpointRulesExpanded.value[endpointId] = true
}
@@ -1604,32 +1605,37 @@ function removeEndpointBodyRule(endpointId: string, index: number) {
function updateEndpointBodyRuleAction(endpointId: string, index: number, action: BodyRuleAction) {
const rules = getEndpointEditBodyRules(endpointId)
if (rules[index]) {
rules[index].action = action
rules[index].path = ''
rules[index].value = ''
rules[index].from = ''
rules[index].to = ''
rules[index].index = ''
rules[index].pattern = ''
rules[index].replacement = ''
rules[index].flags = ''
rules[index].style = ''
const currentCondition = rules[index].condition
rules[index] = { ...emptyBodyRule(action), condition: currentCondition }
}
}
// 更新请求体规则字段
function updateEndpointBodyRuleField(endpointId: string, index: number, field: 'path' | 'value' | 'from' | 'to' | 'index' | 'pattern' | 'replacement' | 'flags' | 'style' | 'conditionPath' | 'conditionOp' | 'conditionValue', value: string) {
function updateEndpointBodyRuleField(endpointId: string, index: number, field: 'path' | 'value' | 'from' | 'to' | 'index' | 'pattern' | 'replacement' | 'flags' | 'style', value: string) {
const rules = getEndpointEditBodyRules(endpointId)
if (rules[index]) {
rules[index][field] = value
}
}
// 切换请求体规则的条件启用状态
function toggleBodyRuleCondition(endpointId: string, index: number) {
const rules = getEndpointEditBodyRules(endpointId)
if (rules[index]) {
rules[index].conditionEnabled = !rules[index].conditionEnabled
rules[index].condition = rules[index].condition ? null : createEmptyConditionLeaf()
}
}
function updateEndpointBodyRuleCondition(endpointId: string, index: number, condition: EditableConditionNode) {
const rules = getEndpointEditBodyRules(endpointId)
if (rules[index]) {
rules[index].condition = condition
}
}
function clearEndpointBodyRuleCondition(endpointId: string, index: number) {
const rules = getEndpointEditBodyRules(endpointId)
if (rules[index]) {
rules[index].condition = null
}
}
@@ -1657,7 +1663,7 @@ function validateBodyRulePathForEndpoint(endpointId: string, path: string, index
const currentRule = rules[index]
// 任意一方启用了条件,则不视为冲突(条件可能互斥,真正冲突在运行时处理)
const duplicate = rules.findIndex(
(r, i) => i !== index && !currentRule.conditionEnabled && !r.conditionEnabled && (
(r, i) => i !== index && !currentRule.condition && !r.condition && (
((r.action === 'set' || r.action === 'drop') && r.path.trim().toLowerCase() === normalizedPath) ||
(r.action === 'rename' && r.to.trim().toLowerCase() === normalizedPath)
)
@@ -1689,7 +1695,7 @@ function validateBodyRenameFromForEndpoint(endpointId: string, from: string, ind
const rules = getEndpointEditBodyRules(endpointId)
const currentRule = rules[index]
const duplicate = rules.findIndex(
(r, i) => i !== index && !currentRule.conditionEnabled && !r.conditionEnabled &&
(r, i) => i !== index && !currentRule.condition && !r.condition &&
((r.action === 'set' && r.path.trim().toLowerCase() === normalizedFrom) ||
(r.action === 'drop' && r.path.trim().toLowerCase() === normalizedFrom) ||
(r.action === 'rename' && r.from.trim().toLowerCase() === normalizedFrom))
@@ -1721,7 +1727,7 @@ function validateBodyRenameToForEndpoint(endpointId: string, to: string, index:
const rules = getEndpointEditBodyRules(endpointId)
const currentRule = rules[index]
const duplicate = rules.findIndex(
(r, i) => i !== index && !currentRule.conditionEnabled && !r.conditionEnabled &&
(r, i) => i !== index && !currentRule.condition && !r.condition &&
((r.action === 'set' && r.path.trim().toLowerCase() === normalizedTo) ||
(r.action === 'rename' && r.to.trim().toLowerCase() === normalizedTo))
)
@@ -1932,17 +1938,7 @@ function hasBodyRulesChanges(endpoint: ProviderEndpoint): boolean {
if (edited.path !== (original.path ?? '')) return true
if (edited.style !== (original.style ?? '')) return true
}
// 条件变更检测
const origCond = original.condition
if (edited.conditionEnabled !== !!origCond) return true
if (edited.conditionEnabled && origCond) {
if (edited.conditionPath !== origCond.path) return true
if (edited.conditionOp !== origCond.op) return true
const origVal = origCond.value !== undefined
? (typeof origCond.value === 'string' ? origCond.value : JSON.stringify(origCond.value))
: ''
if (edited.conditionValue !== origVal) return true
}
if (!conditionEquals(edited.condition, conditionToEditable(original.condition))) return true
}
return false
}
@@ -1951,21 +1947,8 @@ function hasBodyRulesChanges(endpoint: ProviderEndpoint): boolean {
function rulesToBodyRules(rules: EditableBodyRule[]): BodyRule[] | null {
const result: BodyRule[] = []
// 构建 condition 对象(如果启用且有效)
function buildCondition(rule: EditableBodyRule): BodyRuleCondition | undefined {
if (!rule.conditionEnabled || !rule.conditionPath.trim() || !rule.conditionOp.trim()) return undefined
const op = rule.conditionOp as BodyRuleConditionOp
if (op === 'exists' || op === 'not_exists') {
return { path: rule.conditionPath.trim(), op }
}
const raw = rule.conditionValue.trim()
let val: unknown = raw
try { val = JSON.parse(raw) } catch { /* 保留原字符串 */ }
return { path: rule.conditionPath.trim(), op, value: val }
}
for (const rule of rules) {
const condition = buildCondition(rule)
const condition = editableConditionToApi(rule.condition)
if (rule.action === 'set' && rule.path.trim()) {
let value: unknown = rule.value
try { value = restoreOriginalPlaceholder(JSON.parse(prepareValueForJsonParse(rule.value.trim()))) } catch { value = rule.value }
@@ -2049,6 +2032,9 @@ function getBodyValidationErrorForEndpoint(endpointId: string): string | null {
const validStyles = new Set(['snake_case', 'camelCase', 'PascalCase', 'kebab-case', 'capitalize'])
if (!rule.style.trim() || !validStyles.has(rule.style.trim())) return `${prefix}请选择有效的命名风格`
}
const conditionErr = validateEditableCondition(rule.condition)
if (conditionErr) return `${prefix}${conditionErr}`
}
return null
}
@@ -2092,6 +2078,7 @@ function hasRulesChanges(endpoint: ProviderEndpoint): boolean {
} else if (edited.action === 'rename' && original.action === 'rename') {
if (edited.from !== original.from || edited.to !== original.to) return true
}
if (!conditionEquals(edited.condition, conditionToEditable(original.condition))) return true
}
return false
}
@@ -2143,31 +2130,37 @@ function rulesToHeaderRules(rules: EditableRule[]): HeaderRule[] | null {
const result: HeaderRule[] = []
for (const rule of rules) {
const condition = editableConditionToApi(rule.condition)
if (rule.action === 'set' && rule.key.trim()) {
result.push({ action: 'set', key: rule.key.trim(), value: rule.value })
result.push({ action: 'set', key: rule.key.trim(), value: rule.value, ...(condition ? { condition } : {}) })
} else if (rule.action === 'drop' && rule.key.trim()) {
result.push({ action: 'drop', key: rule.key.trim() })
result.push({ action: 'drop', key: rule.key.trim(), ...(condition ? { condition } : {}) })
} else if (rule.action === 'rename' && rule.from.trim() && rule.to.trim()) {
result.push({ action: 'rename', from: rule.from.trim(), to: rule.to.trim() })
result.push({ action: 'rename', from: rule.from.trim(), to: rule.to.trim(), ...(condition ? { condition } : {}) })
}
}
return result.length > 0 ? result : null
}
// 检查规则是否有验证错误
function hasValidationErrorsForEndpoint(endpointId: string): boolean {
function getHeaderValidationErrorForEndpoint(endpointId: string): string | null {
const rules = getEndpointEditRules(endpointId)
for (let i = 0; i < rules.length; i++) {
const rule = rules[i]
const prefix = `第 ${i + 1} 条请求头规则:`
if (rule.action === 'set' || rule.action === 'drop') {
if (validateRuleKeyForEndpoint(endpointId, rule.key, i)) return true
const err = validateRuleKeyForEndpoint(endpointId, rule.key, i)
if (err) return `${prefix}${err}`
} else if (rule.action === 'rename') {
if (validateRenameFromForEndpoint(endpointId, rule.from, i)) return true
if (validateRenameToForEndpoint(endpointId, rule.to, i)) return true
const fromErr = validateRenameFromForEndpoint(endpointId, rule.from, i)
if (fromErr) return `${prefix}${fromErr}`
const toErr = validateRenameToForEndpoint(endpointId, rule.to, i)
if (toErr) return `${prefix}${toErr}`
}
const conditionErr = validateEditableCondition(rule.condition)
if (conditionErr) return `${prefix}${conditionErr}`
}
return false
return null
}
// 新端点选择的格式的默认路径
@@ -2243,8 +2236,9 @@ async function saveEndpoint(endpoint: ProviderEndpoint) {
if (!state || !state.url) return
// 检查规则是否有验证错误
if (hasValidationErrorsForEndpoint(endpoint.id)) {
showError('请修正请求头规则中的错误')
const headerErr = getHeaderValidationErrorForEndpoint(endpoint.id)
if (headerErr) {
showError(headerErr)
return
}
@@ -3,50 +3,231 @@
variant="default"
class="overflow-hidden"
>
<!-- 标题和筛选器 -->
<div class="px-6 py-3.5 border-b border-border/60">
<div class="flex items-center justify-between gap-4">
<h3 class="text-base font-semibold">
{{ title }}
</h3>
<div class="flex items-center gap-3">
<Label class="text-xs text-muted-foreground">回溯时间:</Label>
<Select
v-model="lookbackHours"
<div class="border-b border-border/60 bg-muted/20 px-6 py-5">
<div class="flex flex-col gap-4">
<div class="flex flex-col gap-3 xl:flex-row xl:items-start xl:justify-between">
<div class="space-y-1">
<div class="flex items-center gap-2 flex-wrap">
<h3 class="text-base font-semibold">
{{ title }}
</h3>
<Badge
variant="outline"
class="text-[10px] uppercase tracking-[0.25em]"
>
{{ isAdmin ? 'Admin' : 'Public' }}
</Badge>
</div>
<p class="text-xs text-muted-foreground">
异常优先排序,支持状态筛选、详情诊断与自动刷新,首屏先看出问题的格式。
</p>
</div>
<div class="flex flex-wrap items-center gap-3">
<div class="text-xs text-muted-foreground">
最后刷新 {{ generatedAt ? formatRelativeTime(generatedAt) : '未加载' }}
</div>
<div class="flex items-center gap-2 rounded-xl border border-border/60 bg-card/70 px-3 py-1.5">
<Switch v-model="autoRefreshEnabled" />
<span class="text-xs font-medium">
{{ autoRefreshEnabled ? '自动刷新 15s' : '自动刷新已关闭' }}
</span>
</div>
<Button
v-if="isAdmin"
variant="outline"
size="sm"
class="h-9 px-3 text-xs"
:disabled="recoveringAllKeys || (healthSummary?.keys.circuit_open ?? 0) <= 0"
@click="recoverAllConfirmOpen = true"
>
<Loader2
v-if="recoveringAllKeys"
class="mr-2 h-3.5 w-3.5 animate-spin"
/>
恢复全部熔断 Key
</Button>
<RefreshButton
:loading="loadingMonitors"
@click="refreshData({ silent: false })"
/>
</div>
</div>
<div class="grid grid-cols-2 gap-3 xl:grid-cols-6">
<Card
v-for="item in summaryCards"
:key="item.label"
variant="inset"
class="p-3"
>
<SelectTrigger class="w-28 h-8 text-xs border-border/60">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="1">
1 小时
</SelectItem>
<SelectItem value="6">
6 小时
</SelectItem>
<SelectItem value="12">
12 小时
</SelectItem>
<SelectItem value="24">
24 小时
</SelectItem>
<SelectItem value="48">
48 小时
</SelectItem>
</SelectContent>
</Select>
<RefreshButton
:loading="loading"
@click="refreshData"
/>
<div class="flex items-start justify-between gap-3">
<div class="min-w-0">
<div class="text-[11px] font-semibold uppercase tracking-[0.18em] text-muted-foreground">
{{ item.label }}
</div>
<div
class="mt-2 text-xl font-semibold"
:class="item.valueClass"
>
{{ item.value }}
</div>
<div class="mt-1 text-[11px] text-muted-foreground">
{{ item.help }}
</div>
</div>
<div
class="rounded-xl border border-border/60 p-2"
:class="item.iconClass"
>
<component
:is="item.icon"
class="h-4 w-4"
/>
</div>
</div>
</Card>
</div>
<div class="flex flex-col gap-3 xl:flex-row xl:items-center xl:justify-between">
<div class="grid flex-1 grid-cols-1 gap-3 sm:grid-cols-2 xl:grid-cols-4">
<div>
<Label class="mb-1.5 block text-xs text-muted-foreground">回溯时间</Label>
<Select v-model="lookbackHours">
<SelectTrigger class="h-9 border-border/60 text-xs">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="1">
1 小时
</SelectItem>
<SelectItem value="6">
6 小时
</SelectItem>
<SelectItem value="12">
12 小时
</SelectItem>
<SelectItem value="24">
24 小时
</SelectItem>
<SelectItem value="48">
48 小时
</SelectItem>
</SelectContent>
</Select>
</div>
<div>
<Label class="mb-1.5 block text-xs text-muted-foreground">状态筛选</Label>
<Select v-model="statusFilter">
<SelectTrigger class="h-9 border-border/60 text-xs">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="all">
全部状态
</SelectItem>
<SelectItem value="unhealthy">
仅看异常
</SelectItem>
<SelectItem value="warning">
仅看预警
</SelectItem>
<SelectItem value="healthy">
仅看健康
</SelectItem>
<SelectItem value="unknown">
仅看无流量
</SelectItem>
</SelectContent>
</Select>
</div>
<div>
<Label class="mb-1.5 block text-xs text-muted-foreground">格式族</Label>
<Select v-model="familyFilter">
<SelectTrigger class="h-9 border-border/60 text-xs">
<SelectValue />
</SelectTrigger>
<SelectContent>
<SelectItem value="all">
全部格式
</SelectItem>
<SelectItem value="claude">
Claude
</SelectItem>
<SelectItem value="openai">
OpenAI
</SelectItem>
<SelectItem value="gemini">
Gemini
</SelectItem>
<SelectItem value="other">
其他
</SelectItem>
</SelectContent>
</Select>
</div>
<div>
<Label class="mb-1.5 block text-xs text-muted-foreground">显示范围</Label>
<div class="flex h-9 items-center justify-between rounded-xl border border-border/60 bg-card/70 px-3">
<span class="text-xs font-medium">仅看有流量</span>
<Switch v-model="onlyTraffic" />
</div>
</div>
</div>
<div class="flex items-center gap-2 text-xs text-muted-foreground">
<span>显示 {{ filteredMonitors.length }} / {{ monitors.length }} 个格式</span>
<Button
v-if="hasActiveFilters"
variant="ghost"
size="sm"
class="h-8 px-2 text-xs"
@click="resetFilters"
>
重置筛选
</Button>
</div>
</div>
<div class="flex flex-wrap items-center gap-2 text-xs">
<Badge
variant="success"
class="gap-1"
>
健康
</Badge>
<span class="text-muted-foreground">成功率 ≥ 95%</span>
<Badge
variant="warning"
class="gap-1"
>
预警
</Badge>
<span class="text-muted-foreground">70% - 95%</span>
<Badge
variant="destructive"
class="gap-1"
>
异常
</Badge>
<span class="text-muted-foreground">成功率 &lt; 70%</span>
<Badge
variant="outline"
class="gap-1"
>
无流量
</Badge>
<span class="text-muted-foreground">窗口内没有请求</span>
</div>
</div>
</div>
<!-- 内容区域 -->
<div class="p-6">
<div
v-if="loadingMonitors"
v-if="showInitialLoading"
class="flex items-center justify-center py-12"
>
<Loader2 class="w-6 h-6 animate-spin text-muted-foreground" />
@@ -57,81 +238,149 @@
v-else-if="monitors.length === 0"
class="flex flex-col items-center justify-center py-12 text-muted-foreground"
>
<Activity class="w-12 h-12 mb-3 opacity-30" />
<Activity class="mb-3 h-12 w-12 opacity-30" />
<p>暂无健康监控数据</p>
<p class="text-xs mt-1">
<p class="mt-1 text-xs">
端点尚未产生请求记录
</p>
</div>
<div
v-else-if="filteredMonitors.length === 0"
class="flex flex-col items-center justify-center py-12 text-muted-foreground"
>
<AlertTriangle class="mb-3 h-12 w-12 opacity-30" />
<p>当前筛选条件下没有匹配项</p>
<p class="mt-1 text-xs">
可以放宽状态、格式族或流量条件
</p>
<Button
variant="outline"
size="sm"
class="mt-4"
@click="resetFilters"
>
重置筛选
</Button>
</div>
<div
v-else
class="space-y-3"
>
<div
v-for="monitor in monitors"
v-for="monitor in filteredMonitors"
:key="monitor.api_format"
class="border border-border/60 rounded-lg p-4 hover:border-primary/50 transition-colors"
class="rounded-xl border p-4 transition-colors"
:class="getMonitorCardClass(monitor)"
>
<!-- 响应式布局:窄屏上下两行,宽屏左右结构 -->
<div class="flex flex-col sm:flex-row sm:gap-6 sm:items-center">
<!-- 第一行/左侧:信息区域 -->
<div class="sm:w-52 flex-shrink-0 space-y-1.5 mb-3 sm:mb-0">
<!-- API 格式标签和成功率 -->
<div class="flex flex-col gap-3 xl:flex-row xl:items-center xl:gap-6">
<div class="space-y-1.5 xl:w-64 xl:flex-shrink-0">
<div class="flex items-center gap-2 flex-wrap">
<Badge
variant="outline"
class="font-mono text-xs whitespace-nowrap"
class="whitespace-nowrap font-mono text-xs"
>
{{ formatApiFormat(monitor.api_format) }}
</Badge>
<Badge
:variant="getMonitorStatusVariant(monitor)"
class="text-xs whitespace-nowrap"
>
{{ getMonitorStatusLabel(monitor) }}
</Badge>
<Badge
v-if="monitor.total_attempts > 0"
:variant="getSuccessRateVariant(monitor.success_rate)"
class="text-xs whitespace-nowrap"
>
{{ (monitor.success_rate * 100).toFixed(0) }}%
{{ formatPercent(monitor.success_rate) }}
</Badge>
<!-- 提供商信息(仅管理员可见)- 窄屏时显示在同一行 -->
<span
v-if="showProviderInfo && 'provider_count' in monitor"
class="text-xs text-muted-foreground sm:hidden"
class="text-xs text-muted-foreground xl:hidden"
>
{{ monitor.provider_count }} 个提供商 / {{ monitor.key_count }} 个密钥
</span>
</div>
<!-- 提供商信息(仅管理员可见)- 宽屏时显示在下方 -->
<div
v-if="showProviderInfo && 'provider_count' in monitor"
class="text-xs text-muted-foreground hidden sm:block"
class="hidden text-xs text-muted-foreground xl:block"
>
{{ monitor.provider_count }} 个提供商 / {{ monitor.key_count }} 个密钥
</div>
<div class="text-xs text-muted-foreground">
{{ monitor.total_attempts }} 次请求
<span class="mx-1">·</span>
失败 {{ monitor.failed_count }}
<span class="mx-1">·</span>
跳过 {{ monitor.skipped_count }}
</div>
<div class="text-xs text-muted-foreground">
最近事件 {{ monitor.last_event_at ? formatRelativeTime(monitor.last_event_at) : '暂无' }}
</div>
<div class="pt-1">
<Button
variant="outline"
size="sm"
class="h-8 px-3 text-xs"
@click="openMonitorDetail(monitor)"
>
查看详情
</Button>
</div>
</div>
<!-- 第二行/右侧:时间线区域 -->
<div class="flex-1 min-w-0 sm:flex sm:justify-end">
<div class="w-full sm:max-w-5xl">
<EndpointHealthTimeline
:monitor="monitor"
:lookback-hours="parseInt(lookbackHours)"
/>
</div>
<div class="min-w-0 flex-1">
<EndpointHealthTimeline
:monitor="monitor"
:lookback-hours="parseInt(lookbackHours)"
/>
</div>
</div>
</div>
</div>
</div>
</Card>
<HealthMonitorDetailDialog
v-model="detailDialogOpen"
:monitor="selectedMonitor"
:is-admin="isAdmin"
:format-keys="selectedFormatKeys"
:loading-format-keys="loadingFormatKeys"
:recovering-key-id="recoveringKeyId"
@navigate-provider="goToProviderManagement"
@navigate-pool="goToPoolManagement"
@recover-key="recoverFormatKey"
/>
<AlertDialog
v-model="recoverAllConfirmOpen"
title="恢复全部熔断 Key"
:description="recoverAllDialogDescription"
type="warning"
confirm-text="立即恢复"
:loading="recoveringAllKeys"
@confirm="confirmRecoverAllKeys"
/>
</template>
<script setup lang="ts">
import { ref, onMounted, watch } from 'vue'
import { Activity, Loader2 } from 'lucide-vue-next'
import { computed, onMounted, onUnmounted, ref, watch, type Component } from 'vue'
import { useRouter } from 'vue-router'
import { Activity, AlertTriangle, Gauge, KeyRound, Loader2, Radio, ShieldAlert } from 'lucide-vue-next'
import AlertDialog from '@/components/common/AlertDialog.vue'
import HealthMonitorDetailDialog from './HealthMonitorDetailDialog.vue'
import Card from '@/components/ui/card.vue'
import Badge from '@/components/ui/badge.vue'
import Button from '@/components/ui/button.vue'
import Label from '@/components/ui/label.vue'
import Switch from '@/components/ui/switch.vue'
import Select from '@/components/ui/select.vue'
import SelectTrigger from '@/components/ui/select-trigger.vue'
import SelectValue from '@/components/ui/select-value.vue'
@@ -139,12 +388,24 @@ import SelectContent from '@/components/ui/select-content.vue'
import SelectItem from '@/components/ui/select-item.vue'
import RefreshButton from '@/components/ui/refresh-button.vue'
import EndpointHealthTimeline from './EndpointHealthTimeline.vue'
import { getEndpointStatusMonitor, getPublicEndpointStatusMonitor } from '@/api/endpoints/health'
import type { EndpointStatusMonitor, PublicEndpointStatusMonitor } from '@/api/endpoints/types'
import { getKeysGroupedByFormat, type GroupedFormatKey } from '@/api/endpoints/keys'
import { getEndpointStatusMonitor, getHealthSummary, getPublicEndpointStatusMonitor, recoverAllKeysHealth, recoverKeyHealth } from '@/api/endpoints/health'
import type { EndpointStatusMonitor, HealthSummary, PublicEndpointStatusMonitor } from '@/api/endpoints/types'
import { useToast } from '@/composables/useToast'
import { parseApiError } from '@/utils/errorParser'
import { formatApiFormat } from '@/api/endpoints/types/api-format'
type MonitorItem = EndpointStatusMonitor | PublicEndpointStatusMonitor
type MonitorStatus = 'healthy' | 'warning' | 'unhealthy' | 'unknown'
type SummaryCardItem = {
label: string
value: string
help: string
icon: Component
valueClass: string
iconClass: string
}
const props = withDefaults(defineProps<{
title?: string
isAdmin?: boolean
@@ -155,55 +416,510 @@ const props = withDefaults(defineProps<{
showProviderInfo: false
})
const { error: showError } = useToast()
const AUTO_REFRESH_INTERVAL_MS = 15_000
const STATUS_SORT_ORDER: Record<MonitorStatus, number> = {
unhealthy: 0,
warning: 1,
unknown: 2,
healthy: 3
}
const router = useRouter()
const { error: showError, success: showSuccess, warning: showWarning } = useToast()
const loading = ref(false)
const loadingMonitors = ref(false)
const monitors = ref<(EndpointStatusMonitor | PublicEndpointStatusMonitor)[]>([])
const monitors = ref<MonitorItem[]>([])
const generatedAt = ref<string | null>(null)
const healthSummary = ref<HealthSummary | null>(null)
const lookbackHours = ref('6')
const statusFilter = ref<'all' | MonitorStatus>('all')
const familyFilter = ref<'all' | 'claude' | 'openai' | 'gemini' | 'other'>('all')
const onlyTraffic = ref(false)
const autoRefreshEnabled = ref(true)
const detailDialogOpen = ref(false)
const selectedMonitorFormat = ref<string | null>(null)
const recoverAllConfirmOpen = ref(false)
const recoveringAllKeys = ref(false)
const keysByFormat = ref<Record<string, GroupedFormatKey[]>>({})
const loadingFormatKeys = ref(false)
const recoveringKeyId = ref<string | null>(null)
async function loadMonitors() {
let autoRefreshTimer: ReturnType<typeof setInterval> | null = null
let loadRequestId = 0
const isAdmin = computed(() => props.isAdmin)
const showInitialLoading = computed(() => loadingMonitors.value && monitors.value.length === 0)
const unhealthyCount = computed(() =>
monitors.value.filter(item => getMonitorStatus(item) === 'unhealthy').length
)
const warningCount = computed(() =>
monitors.value.filter(item => getMonitorStatus(item) === 'warning').length
)
const trafficCount = computed(() =>
monitors.value.filter(item => item.total_attempts > 0).length
)
const overallSuccessRate = computed(() => {
const totals = monitors.value.reduce((acc, item) => {
acc.success += item.success_count
acc.failed += item.failed_count
return acc
}, { success: 0, failed: 0 })
const denominator = totals.success + totals.failed
if (denominator <= 0) return null
return totals.success / denominator
})
const summaryCards = computed<SummaryCardItem[]>(() => {
const items: SummaryCardItem[] = [
{
label: '异常格式',
value: `${unhealthyCount.value}`,
help: `共 ${monitors.value.length} 个格式`,
icon: AlertTriangle,
valueClass: unhealthyCount.value > 0 ? 'text-destructive' : 'text-foreground',
iconClass: unhealthyCount.value > 0 ? 'bg-destructive/10 text-destructive' : 'bg-muted/60 text-muted-foreground'
},
{
label: '预警格式',
value: `${warningCount.value}`,
help: '建议优先观察',
icon: Radio,
valueClass: warningCount.value > 0 ? 'text-yellow-600 dark:text-yellow-400' : 'text-foreground',
iconClass: warningCount.value > 0 ? 'bg-yellow-500/10 text-yellow-600 dark:text-yellow-400' : 'bg-muted/60 text-muted-foreground'
},
{
label: '有流量格式',
value: `${trafficCount.value}`,
help: `最近 ${lookbackHours.value} 小时`,
icon: Activity,
valueClass: 'text-foreground',
iconClass: 'bg-primary/10 text-primary'
},
{
label: '整体成功率',
value: overallSuccessRate.value == null ? '--' : formatPercent(overallSuccessRate.value),
help: '按 success / (success + failed)',
icon: Gauge,
valueClass: overallSuccessRate.value != null && overallSuccessRate.value < 0.95
? 'text-yellow-600 dark:text-yellow-400'
: 'text-foreground',
iconClass: 'bg-primary/10 text-primary'
}
]
if (props.isAdmin) {
items.push(
{
label: '熔断 Key',
value: `${healthSummary.value?.keys.circuit_open ?? 0}`,
help: '已打开熔断器',
icon: ShieldAlert,
valueClass: (healthSummary.value?.keys.circuit_open ?? 0) > 0 ? 'text-destructive' : 'text-foreground',
iconClass: (healthSummary.value?.keys.circuit_open ?? 0) > 0 ? 'bg-destructive/10 text-destructive' : 'bg-muted/60 text-muted-foreground'
},
{
label: '活跃 Key',
value: `${healthSummary.value?.keys.active ?? 0}`,
help: '当前可调度密钥',
icon: KeyRound,
valueClass: 'text-foreground',
iconClass: 'bg-primary/10 text-primary'
}
)
}
return items
})
const filteredMonitors = computed(() => {
const filtered = monitors.value.filter((monitor) => {
const status = getMonitorStatus(monitor)
const family = getMonitorFamily(monitor.api_format)
if (statusFilter.value !== 'all' && status !== statusFilter.value) {
return false
}
if (familyFilter.value !== 'all' && family !== familyFilter.value) {
return false
}
if (onlyTraffic.value && monitor.total_attempts <= 0) {
return false
}
return true
})
return filtered.sort((left, right) => {
const statusGap = STATUS_SORT_ORDER[getMonitorStatus(left)] - STATUS_SORT_ORDER[getMonitorStatus(right)]
if (statusGap !== 0) return statusGap
const leftTime = left.last_event_at ? new Date(left.last_event_at).getTime() : 0
const rightTime = right.last_event_at ? new Date(right.last_event_at).getTime() : 0
if (leftTime !== rightTime) return rightTime - leftTime
if (left.total_attempts !== right.total_attempts) {
return right.total_attempts - left.total_attempts
}
return left.api_format.localeCompare(right.api_format)
})
})
const hasActiveFilters = computed(() =>
statusFilter.value !== 'all' || familyFilter.value !== 'all' || onlyTraffic.value
)
const selectedMonitor = computed<MonitorItem | null>(() => {
if (!selectedMonitorFormat.value) return null
return monitors.value.find(item => item.api_format === selectedMonitorFormat.value) ?? null
})
const selectedFormatKeys = computed<GroupedFormatKey[]>(() => {
if (!selectedMonitorFormat.value) return []
const keys = keysByFormat.value[selectedMonitorFormat.value] ?? []
return [...keys].sort((left, right) => {
const circuitGap = Number(right.circuit_breaker_open) - Number(left.circuit_breaker_open)
if (circuitGap !== 0) return circuitGap
const leftHealth = left.health_score ?? 1
const rightHealth = right.health_score ?? 1
if (leftHealth !== rightHealth) return leftHealth - rightHealth
if (left.request_count !== right.request_count) {
return right.request_count - left.request_count
}
return left.name.localeCompare(right.name)
})
})
const recoverAllDialogDescription = computed(() => {
const circuitOpen = healthSummary.value?.keys.circuit_open ?? 0
return [
'这会重置所有已熔断 Key 的健康状态并关闭熔断器。',
`当前检测到 **${circuitOpen}** 个熔断 Key。`,
'建议在确认上游服务恢复后执行。'
].join('\n')
})
async function loadMonitors(silent = false) {
const currentRequestId = ++loadRequestId
loadingMonitors.value = true
try {
const params = {
lookback_hours: parseInt(lookbackHours.value),
per_format_limit: 100
per_format_limit: 60
}
const monitorPromise = props.isAdmin
? getEndpointStatusMonitor(params)
: getPublicEndpointStatusMonitor(params)
const [monitorResult, summaryResult] = await Promise.allSettled([
monitorPromise,
props.isAdmin ? getHealthSummary() : Promise.resolve(null)
])
if (currentRequestId !== loadRequestId) return
if (monitorResult.status === 'fulfilled') {
monitors.value = monitorResult.value.formats || []
generatedAt.value = monitorResult.value.generated_at ?? null
} else if (!silent) {
showError(parseApiError(monitorResult.reason, '加载健康监控数据失败'), '错误')
}
if (props.isAdmin) {
const data = await getEndpointStatusMonitor(params)
monitors.value = data.formats || []
if (summaryResult.status === 'fulfilled') {
healthSummary.value = summaryResult.value
}
} else {
const data = await getPublicEndpointStatusMonitor(params)
monitors.value = data.formats || []
healthSummary.value = null
}
if (selectedMonitorFormat.value && !monitors.value.some(item => item.api_format === selectedMonitorFormat.value)) {
detailDialogOpen.value = false
selectedMonitorFormat.value = null
}
} catch (err: unknown) {
showError(parseApiError(err, '加载健康监控数据失败'), '错误')
if (!silent) {
showError(parseApiError(err, '加载健康监控数据失败'), '错误')
}
} finally {
loadingMonitors.value = false
if (currentRequestId === loadRequestId) {
loadingMonitors.value = false
}
}
}
async function refreshData() {
loading.value = true
async function refreshData(options: { silent?: boolean } = {}) {
await loadMonitors(Boolean(options.silent))
}
async function loadKeysByFormat(force = false) {
if (!props.isAdmin) return
if (!force && Object.keys(keysByFormat.value).length > 0) return
loadingFormatKeys.value = true
try {
await loadMonitors()
keysByFormat.value = await getKeysGroupedByFormat()
} catch (err: unknown) {
showError(parseApiError(err, '加载 Key 诊断数据失败'), '错误')
} finally {
loading.value = false
loadingFormatKeys.value = false
}
}
function getSuccessRateVariant(rate: number): 'default' | 'secondary' | 'destructive' | 'outline' {
if (rate >= 0.95) return 'default'
if (rate >= 0.8) return 'secondary'
return 'destructive'
function resetFilters() {
statusFilter.value = 'all'
familyFilter.value = 'all'
onlyTraffic.value = false
}
function stopAutoRefresh() {
if (!autoRefreshTimer) return
clearInterval(autoRefreshTimer)
autoRefreshTimer = null
}
function startAutoRefresh() {
stopAutoRefresh()
if (!autoRefreshEnabled.value) return
autoRefreshTimer = setInterval(() => {
if (document.visibilityState === 'hidden') return
void refreshData({ silent: true })
}, AUTO_REFRESH_INTERVAL_MS)
}
function handleVisibilityChange() {
if (document.visibilityState === 'visible' && autoRefreshEnabled.value) {
void refreshData({ silent: true })
}
}
async function openMonitorDetail(monitor: MonitorItem) {
selectedMonitorFormat.value = monitor.api_format
detailDialogOpen.value = true
if (props.isAdmin) {
await loadKeysByFormat()
}
}
async function confirmRecoverAllKeys() {
if (recoveringAllKeys.value) return
recoveringAllKeys.value = true
try {
const result = await recoverAllKeysHealth()
recoverAllConfirmOpen.value = false
if (result.recovered_count > 0) {
showSuccess(result.message || `已恢复 ${result.recovered_count} 个熔断 Key`)
} else {
showWarning(result.message || '当前没有需要恢复的熔断 Key')
}
await refreshData({ silent: false })
} catch (err: unknown) {
showError(parseApiError(err, '批量恢复熔断 Key 失败'), '错误')
} finally {
recoveringAllKeys.value = false
}
}
async function recoverFormatKey(key: GroupedFormatKey) {
if (recoveringKeyId.value || !selectedMonitorFormat.value) return
recoveringKeyId.value = key.id
try {
const result = await recoverKeyHealth(key.id, selectedMonitorFormat.value)
showSuccess(result.message || '当前格式已恢复')
await Promise.all([
refreshData({ silent: true }),
loadKeysByFormat(true)
])
} catch (err: unknown) {
showError(parseApiError(err, '恢复 Key 健康状态失败'), '错误')
} finally {
recoveringKeyId.value = null
}
}
function goToProviderManagement(providerId: string) {
void router.push({
name: 'ProviderManagement',
query: { providerId }
})
}
function goToPoolManagement(providerId: string, search?: string) {
const query: Record<string, string> = { providerId }
const normalizedSearch = search?.trim()
if (normalizedSearch) {
query.search = normalizedSearch
}
void router.push({
name: 'PoolManagement',
query
})
}
function getMonitorFamily(apiFormat: string): 'claude' | 'openai' | 'gemini' | 'other' {
if (apiFormat.startsWith('claude:')) return 'claude'
if (apiFormat.startsWith('openai:')) return 'openai'
if (apiFormat.startsWith('gemini:')) return 'gemini'
return 'other'
}
function getLatestTimelineStatus(monitor: MonitorItem): MonitorStatus | null {
const timeline = monitor.timeline ?? []
for (let index = timeline.length - 1; index >= 0; index -= 1) {
const status = timeline[index]
if (status === 'healthy' || status === 'warning' || status === 'unhealthy') {
return status
}
}
return null
}
function getMonitorStatus(monitor: MonitorItem): MonitorStatus {
const timelineStatus = getLatestTimelineStatus(monitor)
if (timelineStatus === 'unhealthy') return 'unhealthy'
if (timelineStatus === 'warning') return 'warning'
if (monitor.total_attempts <= 0) {
return 'unknown'
}
if (monitor.success_rate < 0.7) return 'unhealthy'
if (monitor.success_rate < 0.95) return 'warning'
if (timelineStatus === 'healthy' || monitor.success_rate >= 0.95) return 'healthy'
return 'unknown'
}
function getMonitorStatusLabel(monitor: MonitorItem): string {
switch (getMonitorStatus(monitor)) {
case 'unhealthy':
return '异常'
case 'warning':
return '预警'
case 'healthy':
return '健康'
default:
return '无流量'
}
}
function getMonitorStatusVariant(monitor: MonitorItem): 'success' | 'warning' | 'destructive' | 'outline' {
switch (getMonitorStatus(monitor)) {
case 'healthy':
return 'success'
case 'warning':
return 'warning'
case 'unhealthy':
return 'destructive'
default:
return 'outline'
}
}
function getMonitorCardClass(monitor: MonitorItem): string {
switch (getMonitorStatus(monitor)) {
case 'unhealthy':
return 'border-red-300/70 bg-red-500/5 hover:border-red-400/80'
case 'warning':
return 'border-yellow-300/70 bg-yellow-500/5 hover:border-yellow-400/80'
case 'healthy':
return 'border-border/60 hover:border-primary/50'
default:
return 'border-border/40 bg-muted/20 hover:border-border/70'
}
}
function getSuccessRateVariant(rate: number): 'success' | 'warning' | 'destructive' | 'outline' {
if (rate >= 0.95) return 'success'
if (rate >= 0.8) return 'warning'
if (rate >= 0) return 'destructive'
return 'outline'
}
function formatPercent(rate: number): string {
return `${(rate * 100).toFixed(0)}%`
}
function formatRelativeTime(dateString: string): string {
const diff = Date.now() - new Date(dateString).getTime()
if (Number.isNaN(diff)) return '未知'
if (diff < 60_000) return '刚刚'
const minutes = Math.floor(diff / 60_000)
if (minutes < 60) return `${minutes} 分钟前`
const hours = Math.floor(minutes / 60)
if (hours < 24) return `${hours} 小时前`
const days = Math.floor(hours / 24)
if (days < 7) return `${days} 天前`
return formatAbsoluteTime(dateString)
}
function formatAbsoluteTime(dateString: string): string {
return new Date(dateString).toLocaleString('zh-CN', {
month: '2-digit',
day: '2-digit',
hour: '2-digit',
minute: '2-digit',
second: '2-digit'
})
}
watch(lookbackHours, () => {
loadMonitors()
void refreshData({ silent: false })
})
watch(() => props.isAdmin, () => {
keysByFormat.value = {}
void refreshData({ silent: false })
})
watch(autoRefreshEnabled, () => {
startAutoRefresh()
})
watch(detailDialogOpen, (open) => {
if (!open) {
selectedMonitorFormat.value = null
}
})
onMounted(() => {
refreshData()
document.addEventListener('visibilitychange', handleVisibilityChange)
startAutoRefresh()
void refreshData({ silent: false })
})
onUnmounted(() => {
document.removeEventListener('visibilitychange', handleVisibilityChange)
stopAutoRefresh()
})
</script>
@@ -0,0 +1,936 @@
<template>
<Dialog
v-model="open"
size="5xl"
close-on-backdrop
:title="monitor ? `${formatApiFormat(monitor.api_format)} 详情` : '监控详情'"
:description="dialogDescription"
>
<template #header-actions>
<Button
variant="ghost"
size="icon"
class="h-8 w-8"
aria-label="关闭详情"
@click="open = false"
>
<X class="h-4 w-4" />
</Button>
</template>
<div
v-if="monitor"
class="space-y-4"
>
<div class="grid grid-cols-2 gap-3 xl:grid-cols-5">
<Card
variant="inset"
class="p-3"
>
<div class="text-[11px] uppercase tracking-[0.18em] text-muted-foreground">
当前状态
</div>
<div class="mt-2">
<Badge :variant="getMonitorStatusVariant(monitor)">
{{ getMonitorStatusLabel(monitor) }}
</Badge>
</div>
</Card>
<Card
variant="inset"
class="p-3"
>
<div class="text-[11px] uppercase tracking-[0.18em] text-muted-foreground">
成功率
</div>
<div class="mt-2 text-xl font-semibold">
{{ formatPercent(monitor.success_rate) }}
</div>
</Card>
<Card
variant="inset"
class="p-3"
>
<div class="text-[11px] uppercase tracking-[0.18em] text-muted-foreground">
请求总数
</div>
<div class="mt-2 text-xl font-semibold">
{{ monitor.total_attempts }}
</div>
</Card>
<Card
variant="inset"
class="p-3"
>
<div class="text-[11px] uppercase tracking-[0.18em] text-muted-foreground">
最近失败
</div>
<div class="mt-2 text-xl font-semibold">
{{ failureEvents.length }}
</div>
</Card>
<Card
variant="inset"
class="p-3"
>
<div class="text-[11px] uppercase tracking-[0.18em] text-muted-foreground">
最近事件
</div>
<div class="mt-2 text-sm font-medium">
{{ monitor.last_event_at ? formatAbsoluteTime(monitor.last_event_at) : '暂无' }}
</div>
</Card>
</div>
<div
v-if="monitorApiPath"
class="rounded-xl border border-border/60 bg-muted/20 px-4 py-3 text-sm"
>
<div class="text-[11px] uppercase tracking-[0.18em] text-muted-foreground">
请求入口
</div>
<div class="mt-2 font-mono text-foreground">
{{ monitorApiPath }}
</div>
</div>
<div
v-if="isAdmin"
class="grid grid-cols-1 gap-4 xl:grid-cols-[minmax(0,0.85fr)_minmax(0,1.15fr)]"
>
<Card
variant="inset"
class="p-4"
>
<div class="mb-3 flex items-center justify-between">
<h4 class="text-sm font-semibold">
受影响 Provider
</h4>
<span class="text-xs text-muted-foreground">
{{ providerSummaries.length }} 个
</span>
</div>
<div
v-if="loadingFormatKeys"
class="flex items-center gap-2 py-6 text-sm text-muted-foreground"
>
<Loader2 class="h-4 w-4 animate-spin" />
正在加载 Key 诊断数据...
</div>
<div
v-else-if="providerSummaries.length === 0"
class="py-6 text-sm text-muted-foreground"
>
当前格式下没有可诊断的 provider/key 数据。
</div>
<div
v-else
class="max-h-[420px] space-y-2 overflow-y-auto pr-1"
>
<div
v-for="provider in providerSummaries"
:key="provider.provider_id"
class="rounded-xl border p-3 transition-colors"
:class="providerFilter === provider.provider_id
? 'border-primary/60 bg-primary/5'
: 'border-border/50 bg-card/70'"
>
<div class="flex items-center justify-between gap-3">
<div class="min-w-0">
<div class="flex items-center gap-2 flex-wrap">
<span class="text-sm font-medium">
{{ provider.provider_name }}
</span>
<Badge :variant="provider.circuitOpenKeys > 0 ? 'destructive' : provider.providerActive ? 'success' : 'outline'">
{{ provider.circuitOpenKeys > 0 ? `${provider.circuitOpenKeys} 熔断` : provider.providerActive ? '活跃' : '停用' }}
</Badge>
</div>
<div class="mt-1 text-xs text-muted-foreground">
{{ provider.totalKeys }} 个 Key
<span class="mx-1">·</span>
活跃 {{ provider.activeKeys }}
<span class="mx-1">·</span>
平均健康 {{ formatScore(provider.avgHealthScore) }}
</div>
</div>
<div class="text-right text-xs text-muted-foreground">
<div>{{ provider.requestCount }} 请求</div>
</div>
</div>
<div class="mt-3 flex flex-wrap gap-2">
<Button
variant="outline"
size="sm"
:class="FILTER_BUTTON_CLASS"
@click="toggleProviderFilter(provider.provider_id)"
>
{{ providerFilter === provider.provider_id ? '取消筛选' : '只看该 Provider' }}
</Button>
<Button
variant="outline"
size="sm"
:class="FILTER_BUTTON_CLASS"
@click="$emit('navigateProvider', provider.provider_id)"
>
前往 Provider
</Button>
<Button
v-if="provider.poolEnabled"
variant="outline"
size="sm"
:class="FILTER_BUTTON_CLASS"
@click="$emit('navigatePool', provider.provider_id)"
>
前往号池
</Button>
</div>
</div>
</div>
</Card>
<Card
variant="inset"
class="p-4"
>
<div class="mb-3 flex flex-col gap-3">
<div class="flex items-center justify-between gap-3">
<h4 class="text-sm font-semibold">
Key 诊断
</h4>
<span class="text-xs text-muted-foreground">
{{ filteredFormatKeys.length }} / {{ sortedFormatKeys.length }} 个 Key
</span>
</div>
<div class="flex flex-wrap gap-2">
<Button
variant="outline"
size="sm"
:class="[FILTER_BUTTON_CLASS, keyScopeFilter === 'all' ? 'border-primary/60 bg-primary/5 text-primary' : '']"
@click="keyScopeFilter = 'all'"
>
全部
</Button>
<Button
variant="outline"
size="sm"
:class="[FILTER_BUTTON_CLASS, keyScopeFilter === 'circuit-open' ? 'border-primary/60 bg-primary/5 text-primary' : '']"
@click="keyScopeFilter = 'circuit-open'"
>
仅看熔断
</Button>
<Button
variant="outline"
size="sm"
:class="[FILTER_BUTTON_CLASS, keyScopeFilter === 'recoverable' ? 'border-primary/60 bg-primary/5 text-primary' : '']"
@click="keyScopeFilter = 'recoverable'"
>
仅看可恢复项
</Button>
<Button
v-if="providerFilter !== 'all' || keyScopeFilter !== 'all'"
variant="outline"
size="sm"
:class="FILTER_BUTTON_CLASS"
@click="resetFilters"
>
清空局部筛选
</Button>
</div>
<div
v-if="providerFilter !== 'all' || keyScopeFilter !== 'all'"
class="text-xs text-muted-foreground"
>
<span v-if="providerFilter !== 'all'">
Provider:{{ providerFilterName }}
</span>
<span v-if="providerFilter !== 'all' && keyScopeFilter !== 'all'"> · </span>
<span v-if="keyScopeFilter === 'circuit-open'">范围:仅熔断项</span>
<span v-else-if="keyScopeFilter === 'recoverable'">范围:仅可恢复项</span>
</div>
</div>
<div
v-if="loadingFormatKeys"
class="flex items-center gap-2 py-6 text-sm text-muted-foreground"
>
<Loader2 class="h-4 w-4 animate-spin" />
正在加载 Key 诊断数据...
</div>
<div
v-else-if="sortedFormatKeys.length === 0"
class="py-6 text-sm text-muted-foreground"
>
当前格式下没有 Key 样本。
</div>
<div
v-else-if="filteredFormatKeys.length === 0"
class="py-6 text-sm text-muted-foreground"
>
当前局部筛选下没有匹配的 Key。
</div>
<div
v-else
class="max-h-[420px] space-y-2 overflow-y-auto pr-1"
>
<div
v-for="key in filteredFormatKeys"
:key="key.id"
class="rounded-xl border border-border/50 bg-card/70 p-3"
>
<div class="flex flex-col gap-3 sm:flex-row sm:items-start sm:justify-between">
<div class="min-w-0">
<div class="flex items-center gap-2 flex-wrap">
<span class="text-sm font-medium">
{{ key.name }}
</span>
<Badge variant="outline">
{{ key.provider_name }}
</Badge>
<Badge :variant="key.circuit_breaker_open ? 'destructive' : key.is_active && key.provider_active ? 'success' : 'outline'">
{{ key.circuit_breaker_open ? '已熔断' : key.is_active && key.provider_active ? '可用' : '停用' }}
</Badge>
<Badge
v-if="shouldShowRecoverAction(key)"
variant="warning"
>
建议恢复
</Badge>
</div>
<div class="mt-1 font-mono text-xs text-muted-foreground">
{{ key.api_key_masked }}
</div>
<div class="mt-2 flex flex-wrap gap-x-3 gap-y-1 text-xs text-muted-foreground">
<span>健康 {{ formatScore(key.health_score) }}</span>
<span v-if="key.success_rate != null">{{ formatPercent(key.success_rate) }}</span>
<span>{{ key.request_count }} 请求</span>
<span v-if="key.avg_response_time_ms != null">均延迟 {{ formatLatency(key.avg_response_time_ms) }}</span>
</div>
</div>
<div class="flex flex-wrap gap-2 sm:justify-end">
<Button
v-if="key.pool_enabled"
variant="outline"
size="sm"
:class="ACTION_BUTTON_CLASS"
@click="$emit('navigatePool', key.provider_id, key.name)"
>
号池内查看
</Button>
<Button
v-if="shouldShowRecoverAction(key)"
variant="outline"
size="sm"
:class="ACTION_BUTTON_CLASS"
:disabled="recoveringKeyId !== null"
@click="$emit('recoverKey', key)"
>
<Loader2
v-if="recoveringKeyId === key.id"
class="mr-2 h-3.5 w-3.5 animate-spin"
/>
恢复健康度
</Button>
</div>
</div>
</div>
</div>
</Card>
</div>
<div class="grid grid-cols-1 gap-4 xl:grid-cols-2">
<Card
variant="inset"
class="p-4"
>
<div class="mb-3 flex items-center justify-between">
<h4 class="text-sm font-semibold">
最近失败
</h4>
<span class="text-xs text-muted-foreground">
最近 {{ Math.min(failureEvents.length, 8) }} 条
</span>
</div>
<div
v-if="failureEvents.length === 0"
class="py-6 text-sm text-muted-foreground"
>
当前窗口内没有失败事件。
</div>
<div
v-else
class="space-y-2"
>
<div
v-for="event in failureEvents.slice(0, 8)"
:key="buildEventKey(event)"
class="rounded-xl border border-border/50 bg-card/70 p-3"
>
<div class="flex items-center justify-between gap-3">
<div class="flex items-center gap-2">
<Badge variant="destructive">
失败
</Badge>
<span
v-if="event.status_code"
class="text-xs font-medium text-foreground"
>
HTTP {{ event.status_code }}
</span>
</div>
<span class="text-xs text-muted-foreground">
{{ formatAbsoluteTime(event.timestamp) }}
</span>
</div>
<div class="mt-2 text-sm text-foreground">
{{ event.error_type || event.error_message || '未提供错误类型' }}
</div>
<div class="mt-1 text-xs text-muted-foreground">
{{ event.latency_ms != null ? `延迟 ${formatLatency(event.latency_ms)}` : '延迟未知' }}
</div>
</div>
</div>
</Card>
<Card
variant="inset"
class="p-4"
>
<div class="mb-3">
<h4 class="text-sm font-semibold">
错误分布
</h4>
</div>
<div class="space-y-4">
<div>
<div class="mb-2 text-xs uppercase tracking-[0.18em] text-muted-foreground">
Error Type
</div>
<div
v-if="errorTypes.length === 0"
class="text-sm text-muted-foreground"
>
当前没有可归类的错误类型。
</div>
<div
v-else
class="flex flex-wrap gap-2"
>
<Badge
v-for="item in errorTypes"
:key="item.label"
variant="outline"
>
{{ item.label }} × {{ item.count }}
</Badge>
</div>
</div>
<div>
<div class="mb-2 text-xs uppercase tracking-[0.18em] text-muted-foreground">
Status Code
</div>
<div
v-if="statusCodes.length === 0"
class="text-sm text-muted-foreground"
>
当前没有状态码样本。
</div>
<div
v-else
class="flex flex-wrap gap-2"
>
<Badge
v-for="item in statusCodes"
:key="item.label"
variant="outline"
>
{{ item.label }} × {{ item.count }}
</Badge>
</div>
</div>
<div>
<div class="mb-2 text-xs uppercase tracking-[0.18em] text-muted-foreground">
延迟样本
</div>
<div class="grid grid-cols-3 gap-2 text-sm">
<div class="rounded-lg border border-border/50 bg-card/70 px-3 py-2">
<div class="text-[11px] text-muted-foreground">
P50
</div>
<div class="mt-1 font-semibold">
{{ latencyStats.p50 }}
</div>
</div>
<div class="rounded-lg border border-border/50 bg-card/70 px-3 py-2">
<div class="text-[11px] text-muted-foreground">
P95
</div>
<div class="mt-1 font-semibold">
{{ latencyStats.p95 }}
</div>
</div>
<div class="rounded-lg border border-border/50 bg-card/70 px-3 py-2">
<div class="text-[11px] text-muted-foreground">
Max
</div>
<div class="mt-1 font-semibold">
{{ latencyStats.max }}
</div>
</div>
</div>
</div>
</div>
</Card>
</div>
<Card
variant="inset"
class="p-4"
>
<div class="mb-3 flex items-center justify-between">
<h4 class="text-sm font-semibold">
最近活动
</h4>
<span class="text-xs text-muted-foreground">
最近 {{ Math.min(sortedEvents.length, 12) }} 条事件
</span>
</div>
<div
v-if="sortedEvents.length === 0"
class="py-6 text-sm text-muted-foreground"
>
当前没有事件样本。
</div>
<div
v-else
class="space-y-2"
>
<div
v-for="event in sortedEvents.slice(0, 12)"
:key="buildEventKey(event)"
class="flex flex-col gap-2 rounded-xl border border-border/50 bg-card/70 p-3 sm:flex-row sm:items-center sm:justify-between"
>
<div class="flex min-w-0 items-center gap-2 flex-wrap">
<Badge :variant="getEventStatusVariant(event.status)">
{{ getEventStatusLabel(event.status) }}
</Badge>
<span
v-if="event.status_code"
class="text-xs text-foreground"
>
HTTP {{ event.status_code }}
</span>
<span
v-if="event.latency_ms != null"
class="text-xs text-muted-foreground"
>
{{ formatLatency(event.latency_ms) }}
</span>
<span
v-if="event.error_type"
class="truncate text-xs text-muted-foreground"
>
{{ event.error_type }}
</span>
</div>
<div class="text-xs text-muted-foreground">
{{ formatAbsoluteTime(event.timestamp) }}
</div>
</div>
</div>
</Card>
</div>
</Dialog>
</template>
<script setup lang="ts">
import { computed, ref, watch } from 'vue'
import { Loader2, X } from 'lucide-vue-next'
import Card from '@/components/ui/card.vue'
import Badge from '@/components/ui/badge.vue'
import Button from '@/components/ui/button.vue'
import Dialog from '@/components/ui/dialog/Dialog.vue'
import type { GroupedFormatKey } from '@/api/endpoints/keys'
import type { EndpointHealthEvent, EndpointStatusMonitor, PublicEndpointStatusMonitor, PublicHealthEvent } from '@/api/endpoints/types'
import { formatApiFormat } from '@/api/endpoints/types/api-format'
type MonitorItem = EndpointStatusMonitor | PublicEndpointStatusMonitor
type MonitorEvent = EndpointHealthEvent | PublicHealthEvent
type MonitorStatus = 'healthy' | 'warning' | 'unhealthy' | 'unknown'
type CountItem = { label: string; count: number }
type KeyScopeFilter = 'all' | 'circuit-open' | 'recoverable'
type ProviderImpactSummary = {
provider_id: string
provider_name: string
providerActive: boolean
poolEnabled: boolean
totalKeys: number
activeKeys: number
circuitOpenKeys: number
avgHealthScore: number | null
requestCount: number
}
const props = defineProps<{
monitor: MonitorItem | null
isAdmin: boolean
formatKeys: GroupedFormatKey[]
loadingFormatKeys: boolean
recoveringKeyId: string | null
}>()
defineEmits<{
navigateProvider: [providerId: string]
navigatePool: [providerId: string, search?: string]
recoverKey: [key: GroupedFormatKey]
}>()
const open = defineModel<boolean>({ default: false })
const FILTER_BUTTON_CLASS = 'h-8 px-3 text-xs'
const ACTION_BUTTON_CLASS = 'h-8 px-3 text-xs'
const providerFilter = ref<string>('all')
const keyScopeFilter = ref<KeyScopeFilter>('all')
watch(open, (value) => {
if (!value) {
providerFilter.value = 'all'
keyScopeFilter.value = 'all'
}
})
const monitorApiPath = computed(() =>
props.monitor && 'api_path' in props.monitor ? props.monitor.api_path : null
)
const sortedEvents = computed<MonitorEvent[]>(() => {
const events = props.monitor?.events ?? []
return [...events].sort((left, right) =>
new Date(right.timestamp).getTime() - new Date(left.timestamp).getTime()
)
})
const failureEvents = computed<MonitorEvent[]>(() =>
sortedEvents.value.filter(event => event.status === 'failed')
)
const sortedFormatKeys = computed<GroupedFormatKey[]>(() =>
[...props.formatKeys].sort((left, right) => {
const circuitGap = Number(right.circuit_breaker_open) - Number(left.circuit_breaker_open)
if (circuitGap !== 0) return circuitGap
const leftHealth = left.health_score ?? 1
const rightHealth = right.health_score ?? 1
if (leftHealth !== rightHealth) return leftHealth - rightHealth
if (left.request_count !== right.request_count) {
return right.request_count - left.request_count
}
return left.name.localeCompare(right.name)
})
)
const providerSummaries = computed<ProviderImpactSummary[]>(() => {
const grouped = new Map<string, ProviderImpactSummary & { healthValues: number[] }>()
for (const key of sortedFormatKeys.value) {
const current = grouped.get(key.provider_id) ?? {
provider_id: key.provider_id,
provider_name: key.provider_name,
providerActive: key.provider_active,
poolEnabled: key.pool_enabled,
totalKeys: 0,
activeKeys: 0,
circuitOpenKeys: 0,
avgHealthScore: null,
requestCount: 0,
healthValues: []
}
current.totalKeys += 1
current.requestCount += key.request_count
current.providerActive = current.providerActive || key.provider_active
current.poolEnabled = current.poolEnabled || key.pool_enabled
if (key.is_active) current.activeKeys += 1
if (key.circuit_breaker_open) current.circuitOpenKeys += 1
if (key.health_score != null) current.healthValues.push(key.health_score)
grouped.set(key.provider_id, current)
}
return Array.from(grouped.values())
.map((provider) => ({
provider_id: provider.provider_id,
provider_name: provider.provider_name,
providerActive: provider.providerActive,
poolEnabled: provider.poolEnabled,
totalKeys: provider.totalKeys,
activeKeys: provider.activeKeys,
circuitOpenKeys: provider.circuitOpenKeys,
avgHealthScore: provider.healthValues.length > 0
? provider.healthValues.reduce((sum, value) => sum + value, 0) / provider.healthValues.length
: null,
requestCount: provider.requestCount,
}))
.sort((left, right) => {
if (left.circuitOpenKeys !== right.circuitOpenKeys) {
return right.circuitOpenKeys - left.circuitOpenKeys
}
const leftHealth = left.avgHealthScore ?? 1
const rightHealth = right.avgHealthScore ?? 1
if (leftHealth !== rightHealth) return leftHealth - rightHealth
return right.requestCount - left.requestCount
})
})
const providerFilterName = computed(() =>
providerSummaries.value.find(p => p.provider_id === providerFilter.value)?.provider_name ?? '当前 Provider'
)
const filteredFormatKeys = computed<GroupedFormatKey[]>(() =>
sortedFormatKeys.value.filter((key) => {
if (providerFilter.value !== 'all' && key.provider_id !== providerFilter.value) {
return false
}
if (keyScopeFilter.value === 'circuit-open') {
return key.circuit_breaker_open
}
if (keyScopeFilter.value === 'recoverable') {
return shouldShowRecoverAction(key)
}
return true
})
)
const dialogDescription = computed(() => {
if (!props.monitor) return '查看最近失败、错误分布与事件样本'
const timeText = props.monitor.last_event_at
? `最近事件 ${formatRelativeTime(props.monitor.last_event_at)}`
: '当前窗口没有事件'
return `${getMonitorStatusLabel(props.monitor)} · ${props.monitor.total_attempts} 次请求 · ${timeText}`
})
const errorTypes = computed<CountItem[]>(() =>
buildCountItems(
failureEvents.value
.map(event => event.error_type?.trim())
.filter((value): value is string => Boolean(value))
)
)
const statusCodes = computed<CountItem[]>(() =>
buildCountItems(
failureEvents.value
.map(event => event.status_code != null ? String(event.status_code) : null)
.filter((value): value is string => Boolean(value))
)
)
const latencyStats = computed(() => {
const values = sortedEvents.value
.map(event => event.latency_ms)
.filter((value): value is number => value != null)
.sort((left, right) => left - right)
if (values.length === 0) {
return { p50: '--', p95: '--', max: '--' }
}
return {
p50: formatLatency(pickPercentile(values, 0.5)),
p95: formatLatency(pickPercentile(values, 0.95)),
max: formatLatency(values[values.length - 1])
}
})
watch(sortedFormatKeys, (keys) => {
if (providerFilter.value === 'all') return
if (!keys.some(key => key.provider_id === providerFilter.value)) {
providerFilter.value = 'all'
}
})
function resetFilters() {
providerFilter.value = 'all'
keyScopeFilter.value = 'all'
}
function toggleProviderFilter(providerId: string) {
providerFilter.value = providerFilter.value === providerId ? 'all' : providerId
}
function shouldShowRecoverAction(key: GroupedFormatKey): boolean {
if (key.health_score == null) return false
return key.health_score < 0.999
}
function getLatestTimelineStatus(monitor: MonitorItem): MonitorStatus | null {
const timeline = monitor.timeline ?? []
for (let index = timeline.length - 1; index >= 0; index -= 1) {
const status = timeline[index]
if (status === 'healthy' || status === 'warning' || status === 'unhealthy') {
return status
}
}
return null
}
function getMonitorStatus(monitor: MonitorItem): MonitorStatus {
const timelineStatus = getLatestTimelineStatus(monitor)
if (timelineStatus === 'unhealthy') return 'unhealthy'
if (timelineStatus === 'warning') return 'warning'
if (monitor.total_attempts <= 0) {
return 'unknown'
}
if (monitor.success_rate < 0.7) return 'unhealthy'
if (monitor.success_rate < 0.95) return 'warning'
if (timelineStatus === 'healthy' || monitor.success_rate >= 0.95) return 'healthy'
return 'unknown'
}
function getMonitorStatusLabel(monitor: MonitorItem): string {
switch (getMonitorStatus(monitor)) {
case 'unhealthy':
return '异常'
case 'warning':
return '预警'
case 'healthy':
return '健康'
default:
return '无流量'
}
}
function getMonitorStatusVariant(monitor: MonitorItem): 'success' | 'warning' | 'destructive' | 'outline' {
switch (getMonitorStatus(monitor)) {
case 'healthy':
return 'success'
case 'warning':
return 'warning'
case 'unhealthy':
return 'destructive'
default:
return 'outline'
}
}
function getEventStatusVariant(status: string): 'success' | 'warning' | 'destructive' | 'outline' {
switch (status) {
case 'success':
return 'success'
case 'failed':
return 'destructive'
case 'skipped':
return 'warning'
default:
return 'outline'
}
}
function getEventStatusLabel(status: string): string {
switch (status) {
case 'success':
return '成功'
case 'failed':
return '失败'
case 'skipped':
return '跳过'
case 'started':
return '执行中'
default:
return '未知'
}
}
function buildCountItems(values: string[]): CountItem[] {
const counter = new Map<string, number>()
for (const value of values) {
counter.set(value, (counter.get(value) ?? 0) + 1)
}
return Array.from(counter.entries())
.map(([label, count]) => ({ label, count }))
.sort((left, right) => right.count - left.count)
.slice(0, 8)
}
function pickPercentile(sortedValues: number[], percentile: number): number {
if (sortedValues.length === 0) return 0
const index = Math.min(sortedValues.length - 1, Math.max(0, Math.ceil(sortedValues.length * percentile) - 1))
return sortedValues[index]
}
function formatPercent(rate: number): string {
return `${(rate * 100).toFixed(0)}%`
}
function formatScore(score: number | null | undefined): string {
if (score == null) return '--'
return `${(score * 100).toFixed(0)}%`
}
function formatLatency(latencyMs: number): string {
if (latencyMs < 1000) return `${latencyMs}ms`
return `${(latencyMs / 1000).toFixed(2)}s`
}
function formatRelativeTime(dateString: string): string {
const diff = Date.now() - new Date(dateString).getTime()
if (Number.isNaN(diff)) return '未知'
if (diff < 60_000) return '刚刚'
const minutes = Math.floor(diff / 60_000)
if (minutes < 60) return `${minutes} 分钟前`
const hours = Math.floor(minutes / 60)
if (hours < 24) return `${hours} 小时前`
const days = Math.floor(hours / 24)
if (days < 7) return `${days} 天前`
return formatAbsoluteTime(dateString)
}
function formatAbsoluteTime(dateString: string): string {
return new Date(dateString).toLocaleString('zh-CN', {
month: '2-digit',
day: '2-digit',
hour: '2-digit',
minute: '2-digit',
second: '2-digit'
})
}
function buildEventKey(event: MonitorEvent): string {
return `${event.timestamp}-${event.status}-${event.status_code ?? 'na'}-${event.error_type ?? 'none'}-${event.latency_ms ?? 'na'}`
}
</script>
@@ -79,18 +79,19 @@
{{ editingKey ? '' : '*' }}
</Label>
<template v-if="form.auth_type === 'service_account'">
<Textarea
:id="apiKeyInputId"
<JsonImportInput
v-model="form.auth_config_text"
:required="!editingKey"
:placeholder="editingKey ? '留空表示不修改' : '粘贴完整的 Service Account JSON'"
class="min-h-[120px] font-mono text-xs"
autocomplete="off"
spellcheck="false"
:disabled="saving"
:reset-key="formNonce"
accept=".json,.txt,application/json,text/plain"
:multiple="false"
drop-title="拖入 Service Account JSON 或点击选择"
drop-hint="支持 .json / .txt,单文件导入"
:manual-placeholder="editingKey ? '留空表示不修改,或粘贴完整的 Service Account JSON' : '粘贴完整的 Service Account JSON'"
:manual-description="serviceAccountDescription"
textarea-class="min-h-[160px] font-mono text-xs break-all !rounded-xl"
@error="handleServiceAccountImportError"
/>
<p class="text-xs text-muted-foreground mt-1">
JSON 格式,包含 project_id、private_key 等字段
</p>
</template>
<template v-else>
<Input
@@ -330,13 +331,14 @@
<script setup lang="ts">
import { ref, computed, onMounted, watch } from 'vue'
import { Dialog, Button, Input, Label, Switch, Select, SelectTrigger, SelectValue, SelectContent, SelectItem, Textarea } from '@/components/ui'
import { Dialog, Button, Input, Label, Switch, Select, SelectTrigger, SelectValue, SelectContent, SelectItem } from '@/components/ui'
import { Key, SquarePen } from 'lucide-vue-next'
import { useToast } from '@/composables/useToast'
import { useFormDialog } from '@/composables/useFormDialog'
import { parseApiError } from '@/utils/errorParser'
import { parseNumberInput, parseNullableNumberInput } from '@/utils/form'
import { log } from '@/utils/logger'
import JsonImportInput from '@/components/common/JsonImportInput.vue'
import {
addProviderKey,
updateProviderKey,
@@ -415,6 +417,12 @@ const visibleApiFormats = computed(() => {
const showAuthTypeSelector = computed(() => props.providerType === 'vertex_ai')
const serviceAccountDescription = computed(() => (
props.editingKey
? '留空表示不修改;JSON 格式,包含 project_id、private_key 等字段'
: 'JSON 格式,包含 project_id、private_key 等字段'
))
// 默认认证类型
const defaultAuthType = 'api_key' as const
@@ -676,6 +684,10 @@ function parseAuthConfig(): Record<string, unknown> | null {
}
}
function handleServiceAccountImportError(payload: { message: string, title?: string }) {
showError(payload.message, payload.title || '错误')
}
async function handleSave() {
// 必须有 providerId
if (!props.providerId) {
@@ -282,8 +282,6 @@ export interface AliasGroup {
const props = defineProps<{
open: boolean
providerId: string
/** @deprecated */
providerApiFormats?: string[]
models: Model[]
editingGroup?: AliasGroup | null
preselectedModelId?: string | null
@@ -410,8 +410,8 @@
:reset-key="importInputResetKey"
drop-title="拖入授权文件或点击选择"
drop-hint="支持 .json / .txt,可多选"
manual-placeholder="粘贴 Refresh Token 或 JSON 内容"
paste-toggle-text="或手动粘贴 Refresh Token"
manual-placeholder="粘贴 Refresh Token / Access Token 或 JSON 内容"
paste-toggle-text="或手动粘贴 Token"
file-toggle-text="或选择 JSON 文件导入"
textarea-class="min-h-[200px] text-xs font-mono break-all !rounded-xl"
@error="handleImportInputError"
@@ -893,7 +893,7 @@ function isBatchImport(text: string): boolean {
return lines.length > 1
}
function parseImportText(text: string): { refresh_token: string; name?: string } | null {
function parseImportText(text: string): { refresh_token?: string; access_token?: string; name?: string } | null {
const trimmed = text.trim()
if (!trimmed) return null
@@ -907,9 +907,19 @@ function parseImportText(text: string): { refresh_token: string; name?: string }
if (typeof parsed === 'object' && parsed !== null) {
const obj = parsed as Record<string, unknown>
const refreshToken = obj.refresh_token
if (typeof refreshToken === 'string' && refreshToken.trim()) {
const refreshTokenCamel = obj.refreshToken
const accessToken = obj.access_token
const accessTokenCamel = obj.accessToken
const normalizedRefreshToken = typeof refreshToken === 'string' && refreshToken.trim()
? refreshToken.trim()
: (typeof refreshTokenCamel === 'string' && refreshTokenCamel.trim() ? refreshTokenCamel.trim() : undefined)
const normalizedAccessToken = typeof accessToken === 'string' && accessToken.trim()
? accessToken.trim()
: (typeof accessTokenCamel === 'string' && accessTokenCamel.trim() ? accessTokenCamel.trim() : undefined)
if (normalizedRefreshToken || normalizedAccessToken) {
return {
refresh_token: refreshToken.trim(),
refresh_token: normalizedRefreshToken,
access_token: normalizedAccessToken,
name: (typeof obj.name === 'string' ? obj.name : undefined) || (typeof obj.oauth_email === 'string' ? obj.oauth_email : undefined),
}
}
@@ -919,9 +929,34 @@ function parseImportText(text: string): { refresh_token: string; name?: string }
// Not JSON: treat as raw token.
}
if (isLikelyJwtToken(trimmed)) {
return { access_token: trimmed }
}
return { refresh_token: trimmed }
}
function isLikelyJwtToken(token: string): boolean {
const parts = token.trim().split('.')
if (parts.length !== 3 || parts.some(part => !part)) return false
try {
const header = JSON.parse(decodeBase64Url(parts[0])) as Record<string, unknown>
const payload = JSON.parse(decodeBase64Url(parts[1])) as Record<string, unknown>
const tokenType = typeof header.typ === 'string' ? header.typ.toLowerCase() : ''
if (tokenType && tokenType !== 'jwt' && tokenType !== 'at+jwt') return false
return ['exp', 'aud', 'iss', 'scope', 'scp'].some(key => key in payload)
} catch {
return false
}
}
function decodeBase64Url(value: string): string {
const normalized = value.replace(/-/g, '+').replace(/_/g, '/')
const padded = normalized.padEnd(normalized.length + ((4 - (normalized.length % 4)) % 4), '=')
return atob(padded)
}
function handleImportInputError(payload: { message: string; title?: string }) {
showError(payload.message, payload.title)
}
@@ -228,17 +228,17 @@
? 'border-primary/30 bg-primary/5'
: 'border-border/50 bg-background hover:border-border hover:bg-muted/30'
]"
:draggable="!key.is_pool_aggregate && key.is_active && key.provider_active"
@dragstart="(!key.is_pool_aggregate && key.is_active && key.provider_active) && handleKeyDragStart(format, key.id, $event)"
:draggable="key.is_active && key.provider_active"
@dragstart="(key.is_active && key.provider_active) && handleKeyDragStart(format, key.id, $event)"
@dragend="handleKeyDragEnd(format)"
@dragover.prevent="!key.is_pool_aggregate && handleKeyDragOver(format, key.id)"
@dragover.prevent="handleKeyDragOver(format, key.id)"
@dragleave="handleKeyDragLeave(format)"
@drop="!key.is_pool_aggregate && handleKeyDrop(format, key.id)"
@drop="handleKeyDrop(format, key.id)"
>
<!-- 拖拽手柄 -->
<div
class="p-0.5 rounded transition-colors shrink-0"
:class="(!key.is_pool_aggregate && key.is_active && key.provider_active)
:class="(key.is_active && key.provider_active)
? 'cursor-grab active:cursor-grabbing text-muted-foreground/30 group-hover:text-muted-foreground'
: 'text-muted-foreground/15 cursor-default'"
>
@@ -248,7 +248,7 @@
<!-- 可编辑序号 -->
<div class="shrink-0">
<input
v-if="!key.is_pool_aggregate && editingKeyPriority[format] === key.id"
v-if="editingKeyPriority[format] === key.id"
type="number"
min="1"
:value="key.priority"
@@ -260,12 +260,9 @@
>
<div
v-else
class="w-5 h-5 rounded bg-muted/50 flex items-center justify-center text-[11px] font-medium transition-colors"
:class="key.is_pool_aggregate
? 'text-muted-foreground/60 cursor-not-allowed'
: 'text-muted-foreground cursor-pointer hover:bg-primary/10 hover:text-primary'"
:title="key.is_pool_aggregate ? '号池优先级请在号池调度中调整' : '点击编辑优先级'"
@click.stop="!key.is_pool_aggregate && startEditKeyPriority(format, key)"
class="w-5 h-5 rounded bg-muted/50 flex items-center justify-center text-[11px] font-medium transition-colors text-muted-foreground cursor-pointer hover:bg-primary/10 hover:text-primary"
title="点击编辑优先级"
@click.stop="startEditKeyPriority(format, key)"
>
{{ key.priority }}
</div>
@@ -540,6 +537,12 @@ const dragOverKey = ref<Record<string, string | null>>({})
const loadingKeys = ref(false)
const saving = ref(false)
const SAVE_CONCURRENCY = 6
let originalProviderPriorityById = new Map<string, number>()
let originalPoolPriorityByProviderId = new Map<string, number | null>()
let originalKeyPriorityById = new Map<string, Record<string, number>>()
// Key 优先级编辑状态
const editingKeyPriority = ref<Record<string, string | null>>({}) // format -> keyId
@@ -645,6 +648,77 @@ function normalizePriorityMap(
return normalized
}
function normalizeOptionalPriority(value: unknown): number | null {
const num = Number(value)
if (!Number.isFinite(num)) return null
return Math.trunc(num)
}
function normalizeRequiredPriority(value: unknown, fallback: number): number {
return normalizeOptionalPriority(value) ?? fallback
}
function normalizeProvidersForEditing(
providers: ProviderWithEndpointsSummary[]
): ProviderWithEndpointsSummary[] {
const normalized = providers.map((provider, index) => ({
...provider,
provider_priority: normalizeRequiredPriority(provider.provider_priority, 100 + index),
pool_advanced: provider.pool_advanced
? {
...provider.pool_advanced,
global_priority: normalizeOptionalPriority(provider.pool_advanced.global_priority),
}
: provider.pool_advanced,
}))
const minProviderPriority = normalized.reduce(
(min, provider) => Math.min(min, provider.provider_priority),
Number.POSITIVE_INFINITY,
)
const providerOffset = Number.isFinite(minProviderPriority) && minProviderPriority < 1
? 1 - minProviderPriority
: 0
const explicitPoolPriorities = normalized
.map((provider) => normalizeOptionalPriority(provider.pool_advanced?.global_priority))
.filter((priority): priority is number => priority != null)
const minPoolPriority = explicitPoolPriorities.length > 0
? Math.min(...explicitPoolPriorities)
: null
const poolOffset = minPoolPriority != null && minPoolPriority < 1
? 1 - minPoolPriority
: 0
return normalized.map((provider) => ({
...provider,
provider_priority: provider.provider_priority + providerOffset,
pool_advanced: provider.pool_advanced
? {
...provider.pool_advanced,
global_priority: provider.pool_advanced.global_priority == null
? null
: provider.pool_advanced.global_priority + poolOffset,
}
: provider.pool_advanced,
}))
}
function snapshotProviderBaseline(providers: ProviderWithEndpointsSummary[]) {
originalProviderPriorityById = new Map(
providers.map((provider, index) => [
provider.id,
normalizeRequiredPriority(provider.provider_priority, 100 + index),
])
)
originalPoolPriorityByProviderId = new Map(
providers.map((provider) => [
provider.id,
normalizeOptionalPriority(provider.pool_advanced?.global_priority),
])
)
}
function normalizeRateMultipliers(
value: Record<string, unknown> | null | undefined
): Record<string, number> | null {
@@ -660,6 +734,69 @@ function normalizeRateMultipliers(
return Object.keys(normalized).length > 0 ? normalized : null
}
function buildEditableKeyPriorityMap(): Map<string, Record<string, number>> {
const priorityMapByKeyId = new Map<string, Record<string, number>>()
for (const format of Object.keys(keysByFormat.value)) {
const normalizedFormat = normalizeApiFormatKey(format)
if (!normalizedFormat) continue
const keys = keysByFormat.value[format].filter((key) => !isPoolManagedKey(key))
keys.forEach((key) => {
const existing = priorityMapByKeyId.get(key.id) || normalizePriorityMap(key.global_priority_by_format)
existing[normalizedFormat] = Math.max(0, Math.trunc(key.priority))
priorityMapByKeyId.set(key.id, existing)
})
}
return priorityMapByKeyId
}
function snapshotKeyBaseline() {
originalKeyPriorityById = new Map(
Array.from(buildEditableKeyPriorityMap().entries()).map(([keyId, priorityMap]) => [
keyId,
{ ...priorityMap },
])
)
}
function arePriorityMapsEqual(
left: Record<string, number> | undefined,
right: Record<string, number> | undefined,
): boolean {
const leftEntries = Object.entries(left || {}).sort(([a], [b]) => a.localeCompare(b))
const rightEntries = Object.entries(right || {}).sort(([a], [b]) => a.localeCompare(b))
if (leftEntries.length !== rightEntries.length) return false
return leftEntries.every(([format, priority], index) => {
const [rightFormat, rightPriority] = rightEntries[index] || []
return format === rightFormat && priority === rightPriority
})
}
async function runTasksWithConcurrency(
tasks: Array<() => Promise<unknown>>,
concurrency: number = SAVE_CONCURRENCY,
) {
if (tasks.length === 0) return
let cursor = 0
const runNext = async (): Promise<void> => {
while (cursor < tasks.length) {
const index = cursor++
await tasks[index]()
}
}
const workers = Array.from(
{ length: Math.min(concurrency, tasks.length) },
() => runNext(),
)
await Promise.all(workers)
}
const providerById = computed(() => {
const map = new Map<string, ProviderWithEndpointsSummary>()
sortedProviders.value.forEach((provider) => {
@@ -720,6 +857,17 @@ function isPoolAggregateItem(key: KeyWithMeta): boolean {
return key.is_pool_aggregate === true
}
// 将号池聚合项的优先级写回 provider 的 pool_advanced.global_priority(本地状态)
function updatePoolGlobalPriority(providerId: string, priority: number) {
const provider = sortedProviders.value.find((p) => p.id === providerId)
if (!provider) return
if (provider.pool_advanced) {
provider.pool_advanced.global_priority = priority
} else {
provider.pool_advanced = { global_priority: priority }
}
}
function toNumberOrNull(value: unknown): number | null {
const num = Number(value)
return Number.isFinite(num) ? num : null
@@ -849,8 +997,13 @@ watch(internalOpen, async (open) => {
async function loadAllProviders() {
try {
const response = await getProvidersSummary({ page: 1, page_size: 9999 })
sortedProviders.value = sortProvidersByActiveAndPriority(response.items)
snapshotProviderBaseline(response.items)
sortedProviders.value = sortProvidersByActiveAndPriority(
normalizeProvidersForEditing(response.items)
)
} catch {
originalProviderPriorityById = new Map()
originalPoolPriorityByProviderId = new Map()
sortedProviders.value = []
}
}
@@ -987,12 +1140,14 @@ async function loadKeysByFormat() {
data[format] = sortKeysByActiveAndPriority(data[format])
}
keysByFormat.value = data
snapshotKeyBaseline()
const formats = sortApiFormats(Object.keys(data))
if (formats.length > 0 && !formats.includes(activeFormatTab.value)) {
activeFormatTab.value = formats[0]
}
} catch (err: unknown) {
originalKeyPriorityById = new Map()
showError(parseApiError(err, '加载 Key 列表失败'), '错误')
} finally {
loadingKeys.value = false
@@ -1025,7 +1180,6 @@ async function toggleKeyActive(format: string, key: KeyWithMeta) {
// Key 优先级编辑
function startEditKeyPriority(format: string, key: KeyWithMeta) {
if (isPoolAggregateItem(key)) return
editingKeyPriority.value[format] = key.id
}
@@ -1034,16 +1188,17 @@ function cancelEditKeyPriority(format: string) {
}
function finishEditKeyPriority(format: string, key: KeyWithMeta, event: FocusEvent) {
if (isPoolAggregateItem(key)) {
editingKeyPriority.value[format] = null
return
}
const input = event.target as HTMLInputElement
const newPriority = parseInt(input.value, 10)
if (!isNaN(newPriority) && newPriority >= 1) {
// 每个格式独立管理优先级,只更新当前格式
key.priority = newPriority
if (isPoolAggregateItem(key) && key.provider_id) {
// 号池聚合项:写回 provider 的 pool_advanced.global_priority
updatePoolGlobalPriority(key.provider_id, newPriority)
} else {
// 普通 key:直接更新优先级
key.priority = newPriority
}
// 重新排序当前格式
keysByFormat.value[format] = sortKeysByActiveAndPriority(keysByFormat.value[format])
}
@@ -1166,7 +1321,7 @@ function handleProviderDrop(dropIndex: number) {
// Key 拖拽处理
function getEditableKeysForFormat(format: string): KeyWithMeta[] {
return (keysByFormat.value[format] || []).filter((key) => !isPoolManagedKey(key))
return displayKeysByFormat.value[format] || []
}
function handleKeyDragStart(format: string, keyId: string, event: DragEvent) {
@@ -1267,7 +1422,7 @@ function handleKeyDrop(format: string, dropKeyId: string) {
updatedPriorityById.set(key.id, key.priority)
})
// 将新的优先级写回原始数据(含号池 key),但只修改非号池条目
// 将新的优先级写回原始数据(普通 key)
keysByFormat.value[format] = sortKeysByActiveAndPriority(
(keysByFormat.value[format] || []).map((key) => {
const nextPriority = updatedPriorityById.get(key.id)
@@ -1278,6 +1433,14 @@ function handleKeyDrop(format: string, dropKeyId: string) {
}
})
)
// 号池聚合项的优先级写回 provider 的 pool_advanced.global_priority
for (const item of items) {
if (item.is_pool_aggregate && item.provider_id) {
updatePoolGlobalPriority(item.provider_id, item.priority)
}
}
draggedKey.value[format] = null
dragOverKey.value[format] = null
}
@@ -1291,28 +1454,43 @@ async function save() {
// 第一步:先保存所有 Provider 和 Key 的优先级数据
// 确保优先级数据全部到位后,再切换调度模式,避免瞬态不一致
const providerUpdates = sortedProviders.value.map((provider) =>
updateProvider(provider.id, { provider_priority: provider.provider_priority })
)
const providerTasks: Array<() => Promise<unknown>> = []
sortedProviders.value.forEach((provider) => {
const payload: Parameters<typeof updateProvider>[1] = {}
const currentProviderPriority = Math.max(0, Math.trunc(provider.provider_priority))
const originalProviderPriority = originalProviderPriorityById.get(provider.id)
if (originalProviderPriority == null || originalProviderPriority !== currentProviderPriority) {
payload.provider_priority = currentProviderPriority
}
const currentPoolPriority = normalizeOptionalPriority(provider.pool_advanced?.global_priority)
const originalPoolPriority = originalPoolPriorityByProviderId.get(provider.id) ?? null
if (currentPoolPriority !== originalPoolPriority) {
payload.pool_advanced = provider.pool_advanced
? {
...provider.pool_advanced,
global_priority: currentPoolPriority,
}
: null
}
if (Object.keys(payload).length > 0) {
providerTasks.push(() => updateProvider(provider.id, payload))
}
})
// 收集每个 Key 的按格式优先级(保留原有其他格式的配置)
const keyPriorityByFormatMap = new Map<string, Record<string, number>>()
for (const format of Object.keys(keysByFormat.value)) {
const keys = keysByFormat.value[format].filter((key) => !isPoolManagedKey(key))
keys.forEach((key) => {
// 合并原有配置,避免丢失未显示格式的优先级
const existing = keyPriorityByFormatMap.get(key.id)
|| normalizePriorityMap(key.global_priority_by_format)
existing[normalizeApiFormatKey(format)] = key.priority
keyPriorityByFormatMap.set(key.id, existing)
})
}
const keyPriorityByFormatMap = buildEditableKeyPriorityMap()
const keyTasks = Array.from(keyPriorityByFormatMap.entries())
.filter(([keyId, priorityByFormat]) => !arePriorityMapsEqual(
originalKeyPriorityById.get(keyId),
priorityByFormat,
))
.map(([keyId, priorityByFormat]) => () =>
updateProviderKey(keyId, { global_priority_by_format: priorityByFormat })
)
const keyUpdates = Array.from(keyPriorityByFormatMap.entries()).map(([keyId, priorityByFormat]) =>
updateProviderKey(keyId, { global_priority_by_format: priorityByFormat })
)
await Promise.all([...providerUpdates, ...keyUpdates])
await runTasksWithConcurrency([...providerTasks, ...keyTasks])
// 第二步:优先级数据全部就绪后,顺序保存调度配置
// 先保存优先级模式,再保存调度模式,确保 Scheduler 状态完整切换
@@ -1327,6 +1505,7 @@ async function save() {
'调度模式:cache_affinity(缓存亲和模式) 或 load_balance(负载均衡模式) 或 fixed_order(固定顺序模式)'
)
await loadAllProviders()
await loadKeysByFormat()
success('优先级已保存')
@@ -293,7 +293,7 @@
v-if="getOAuthOrgBadge(key)"
variant="secondary"
class="text-[10px] px-1.5 py-0 shrink-0"
:title="getOAuthOrgBadge(key)?.id"
:title="getOAuthOrgBadge(key)?.title"
>
{{ getOAuthOrgBadge(key)?.label }}
</Badge>
@@ -334,7 +334,7 @@
<!-- OAuth 状态(失效/过期/倒计时)和刷新按钮 -->
<template v-if="getKeyOAuthExpires(key)">
<!-- 账号级别异常:醒目提示 + 清除按钮 -->
<template v-if="getKeyOAuthExpires(key)?.isInvalid && isAccountLevelBlock(key)">
<template v-if="isAccountLevelBlock(key)">
<Badge
variant="destructive"
class="text-[10px] px-1.5 py-0 shrink-0 gap-0.5"
@@ -370,12 +370,21 @@
>
{{ getKeyOAuthExpires(key)?.text }}
</span>
<Badge
v-if="key.oauth_temporary"
variant="outline"
class="text-[10px] px-1.5 py-0 shrink-0"
title="仅通过 Access Token 导入,无法自动刷新,到期后需要重新导入"
>
临时
</Badge>
<Button
v-else
variant="ghost"
size="icon"
class="h-4 w-4 shrink-0"
:disabled="refreshingOAuthKeyId === key.id"
:title="getKeyOAuthExpires(key)?.isInvalid ? '重新授权' : '刷新 Token'"
:title="getOAuthRefreshButtonTitle(key)"
@click.stop="handleRefreshOAuth(key)"
>
<RefreshCw
@@ -544,12 +553,12 @@
v-if="refreshingQuota"
class="w-3 h-3 text-muted-foreground/70 animate-spin"
/>
<span
v-if="key.upstream_metadata.codex?.updated_at"
class="text-[9px] text-muted-foreground/70"
>
{{ formatCodexUpdatedAt(key.upstream_metadata.codex.updated_at) }}
</span>
<span
v-if="key.upstream_metadata.codex?.updated_at"
class="text-[9px] text-muted-foreground/70"
>
{{ formatUpdatedAt(key.upstream_metadata.codex.updated_at) }}
</span>
</div>
</div>
<!-- 限额并排显示:Team/Plus/Enterprise 账号 2列, Free 账号 1列 -->
@@ -573,10 +582,21 @@
/>
</div>
<div
v-if="key.upstream_metadata.codex.primary_reset_seconds"
class="text-[9px] text-muted-foreground/70 mt-0.5"
v-if="(key.upstream_metadata.codex.primary_reset_at || key.upstream_metadata.codex.primary_reset_seconds) && shouldStartCodexResetCountdown(key.upstream_metadata.codex.primary_used_percent)"
class="text-[9px] mt-0.5 tabular-nums"
:class="getResetCountdownClass(
key.upstream_metadata.codex.primary_reset_at,
key.upstream_metadata.codex.primary_reset_seconds,
key.upstream_metadata.codex.updated_at,
key.upstream_metadata.codex.primary_used_percent
)"
>
{{ formatResetTime(key.upstream_metadata.codex.primary_reset_seconds) }}后重置
{{ getResetCountdownText(
key.upstream_metadata.codex.primary_reset_at,
key.upstream_metadata.codex.primary_reset_seconds,
key.upstream_metadata.codex.updated_at,
key.upstream_metadata.codex.primary_used_percent
) }}
</div>
</div>
<!-- 5H限额(仅 Team/Plus/Enterprise 显示) -->
@@ -594,9 +614,23 @@
:style="{ width: `${Math.max(100 - key.upstream_metadata.codex.secondary_used_percent, 0)}%` }"
/>
</div>
<div class="text-[9px] text-muted-foreground/70 mt-0.5">
<template v-if="key.upstream_metadata.codex.secondary_reset_seconds">
{{ formatResetTime(key.upstream_metadata.codex.secondary_reset_seconds) }}后重置
<div
v-if="shouldStartCodexResetCountdown(key.upstream_metadata.codex.secondary_used_percent)"
class="text-[9px] mt-0.5 tabular-nums"
:class="getResetCountdownClass(
key.upstream_metadata.codex.secondary_reset_at,
key.upstream_metadata.codex.secondary_reset_seconds,
key.upstream_metadata.codex.updated_at,
key.upstream_metadata.codex.secondary_used_percent
)"
>
<template v-if="key.upstream_metadata.codex.secondary_reset_at || key.upstream_metadata.codex.secondary_reset_seconds">
{{ getResetCountdownText(
key.upstream_metadata.codex.secondary_reset_at,
key.upstream_metadata.codex.secondary_reset_seconds,
key.upstream_metadata.codex.updated_at,
key.upstream_metadata.codex.secondary_used_percent
) }}
</template>
<template v-else>
已重置
@@ -649,7 +683,7 @@
v-if="key.upstream_metadata.antigravity?.updated_at"
class="text-[9px] text-muted-foreground/70"
>
{{ formatAntigravityUpdatedAt(key.upstream_metadata.antigravity.updated_at) }}
{{ formatUpdatedAt(key.upstream_metadata.antigravity.updated_at) }}
</span>
</div>
</div>
@@ -912,6 +946,7 @@
ref="modelMappingTabRef"
:key="`mapping-${provider.id}`"
:provider="provider"
:endpoints="endpoints"
:provider-keys="providerKeys"
:models="providerModels"
:mapping-preview="providerMappingPreview"
@@ -1063,7 +1098,7 @@ import { Popover, PopoverTrigger, PopoverContent } from '@/components/ui'
import { useToast } from '@/composables/useToast'
import { useConfirm } from '@/composables/useConfirm'
import { useClipboard } from '@/composables/useClipboard'
import { useCountdownTimer, formatCountdown, getOAuthExpiresCountdown } from '@/composables/useCountdownTimer'
import { useCountdownTimer, formatCountdown, getCodexResetCountdown } from '@/composables/useCountdownTimer'
import {
getProvider,
getProviderEndpoints,
@@ -1110,8 +1145,15 @@ import {
import type { UpstreamMetadata, AntigravityModelQuota } from '@/api/endpoints/types'
import { formatApiFormat } from '@/api/endpoints/types/api-format'
import { isOAuthAccountProviderType, isKeyManagedProviderType } from '../utils/providerTypeUtils'
import { isAccountLevelBlockReason, cleanAccountBlockReason } from '@/utils/accountBlock'
import { getOAuthOrgBadge } from '@/utils/oauthIdentity'
import { getOAuthRefreshFeedback } from '@/utils/oauthRefreshFeedback'
import {
getAccountStatusDisplay,
getAccountStatusTitle,
getOAuthRefreshButtonTitle as resolveOAuthRefreshButtonTitle,
getOAuthStatusDisplay,
getOAuthStatusTitle as resolveOAuthStatusTitle,
} from '@/utils/providerKeyStatus'
// 扩展端点类型,包含密钥列表
interface ProviderEndpointWithKeys extends ProviderEndpoint {
@@ -1132,7 +1174,7 @@ const emit = defineEmits<{
(e: 'refresh'): void
}>()
const { error: showError, success: showSuccess } = useToast()
const { error: showError, success: showSuccess, warning: showWarning } = useToast()
const { confirm } = useConfirm()
const { copyToClipboard } = useClipboard()
const { tick: countdownTick, start: startCountdownTimer, stop: stopCountdownTimer } = useCountdownTimer()
@@ -1617,7 +1659,7 @@ async function handleRefreshOAuth(key: EndpointAPIKey) {
refreshingOAuthKeyId.value = key.id
try {
const result = await refreshProviderOAuth(key.id)
showSuccess('Token 刷新成功')
let refreshedKey: EndpointAPIKey | null = null
// 更新本地数据
const keyInList = providerKeys.value.find(k => k.id === key.id)
if (keyInList) {
@@ -1629,8 +1671,19 @@ async function handleRefreshOAuth(key: EndpointAPIKey) {
if (freshKeys) {
providerKeys.value = freshKeys
syncCurrentSelections(endpoints.value, freshKeys)
refreshedKey = freshKeys.find(item => item.id === key.id) ?? null
}
}
const feedback = getOAuthRefreshFeedback({
accountStateRecheckAttempted: result.account_state_recheck_attempted,
accountStateRecheckError: result.account_state_recheck_error,
snapshot: refreshedKey,
})
if (feedback.tone === 'warning') {
showWarning(feedback.message)
} else {
showSuccess(feedback.message)
}
// Antigravity:token 刷新后可能完成了账号激活,触发配额获取
// (不 emit('refresh'),避免触发全局 provider 余额刷新)
void autoRefreshQuotaInBackground()
@@ -1643,7 +1696,9 @@ async function handleRefreshOAuth(key: EndpointAPIKey) {
// 判断是否为账号级别的封禁(刷新 token 无法修复)
function isAccountLevelBlock(key: EndpointAPIKey): boolean {
return isAccountLevelBlockReason(key.oauth_invalid_reason)
const account = getAccountStatusDisplay(key)
const oauth = getOAuthStatusDisplay(key, countdownTick.value)
return account.blocked && !oauth?.isInvalid
}
// 清除 OAuth 失效标记
@@ -1652,7 +1707,7 @@ async function handleClearOAuthInvalid(key: EndpointAPIKey) {
const confirmed = await confirm({
title: '清除账号异常标记',
message: `确认账号 "${key.name || key.id.slice(0, 8)}" 已手动完成验证?清除后该 Key 将恢复正常调度。`,
message: `确认账号 "${key.name || key.id.slice(0, 8)}" 已手动完成验证?清除后系统会按当前手动开关和调度状态重新评估该 Key。`,
confirmText: '确认清除',
variant: 'default',
})
@@ -1661,13 +1716,33 @@ async function handleClearOAuthInvalid(key: EndpointAPIKey) {
clearingOAuthInvalidKeyId.value = key.id
try {
await clearOAuthInvalid(key.id)
showSuccess('已清除 OAuth 异常标记,Key 已自动启用')
showSuccess('已清除 OAuth 异常标记')
// 更新本地数据
const keyInList = providerKeys.value.find(k => k.id === key.id)
if (keyInList) {
keyInList.oauth_invalid_at = null
keyInList.oauth_invalid_reason = null
keyInList.is_active = true
if (keyInList.status_snapshot) {
keyInList.status_snapshot = {
...keyInList.status_snapshot,
oauth: {
...keyInList.status_snapshot.oauth,
code: 'none',
label: null,
reason: null,
invalid_at: null,
requires_reauth: false,
},
account: {
...keyInList.status_snapshot.account,
code: 'ok',
label: null,
reason: null,
blocked: false,
recoverable: false,
},
}
}
}
await loadEndpoints()
} catch (err: unknown) {
@@ -1799,8 +1874,11 @@ function shouldAutoRefreshCodexQuota(): boolean {
// 检查 OAuth Token 是否即将过期(Codex / Antigravity / Kiro)
function isTokenExpiringSoon(key: EndpointAPIKey, now: number): boolean {
return key.oauth_invalid_at == null
&& typeof key.oauth_expires_at === 'number'
const oauthCode = String(key.status_snapshot?.oauth?.code || '').trim().toLowerCase()
if (oauthCode && oauthCode !== 'valid' && oauthCode !== 'expiring') {
return false
}
return typeof key.oauth_expires_at === 'number'
&& (key.oauth_expires_at - now) <= AUTO_TOKEN_REFRESH_SKEW_SECONDS
}
@@ -2402,10 +2480,6 @@ function formatUpdatedAt(updatedAt: number): string {
return `${days}天前更新`
}
// 兼容旧函数名
const formatCodexUpdatedAt = formatUpdatedAt
const formatAntigravityUpdatedAt = formatUpdatedAt
function secondsUntilReset(resetTime: string): number | null {
if (!resetTime) return null
const ts = Date.parse(resetTime)
@@ -2526,6 +2600,55 @@ function getAntigravityQuotaSummary(metadata: UpstreamMetadata | null | undefine
return result
}
function getResetCountdownText(
resetAt: number | null | undefined,
resetSecs: number | null | undefined,
updatedAt: number | null | undefined,
usedPercent: number | null | undefined
): string {
const status = getCodexResetCountdown(
resetAt,
resetSecs,
updatedAt,
countdownTick.value,
toCodexRemainingPercent(usedPercent)
)
if (!status) return ''
return status.isExpired ? status.text : `${status.text} 后重置`
}
function getResetCountdownClass(
resetAt: number | null | undefined,
resetSecs: number | null | undefined,
updatedAt: number | null | undefined,
usedPercent: number | null | undefined
): string {
const status = getCodexResetCountdown(
resetAt,
resetSecs,
updatedAt,
countdownTick.value,
toCodexRemainingPercent(usedPercent)
)
if (!status || status.isExpired) return 'text-muted-foreground/70'
if (status.isCritical) return 'text-destructive font-medium animate-pulse'
if (status.isUrgent) return 'text-amber-500 dark:text-amber-400'
return 'text-muted-foreground/70'
}
function toCodexRemainingPercent(usedPercent: number | null | undefined): number | null {
const normalizedUsed = Number(usedPercent)
if (!Number.isFinite(normalizedUsed)) return null
const clampedUsed = Math.min(Math.max(normalizedUsed, 0), 100)
return Math.max(100 - clampedUsed, 0)
}
function shouldStartCodexResetCountdown(usedPercent: number | null | undefined): boolean {
const remainingPercent = toCodexRemainingPercent(usedPercent)
if (remainingPercent == null) return true
return remainingPercent < 100
}
// 格式化重置时间
function formatResetTime(seconds: number): string {
const days = Math.floor(seconds / 86400)
@@ -2559,31 +2682,20 @@ function getOAuthPlanTypeClass(planType: string): string {
// OAuth 状态信息(包括失效和过期)
function getKeyOAuthExpires(key: EndpointAPIKey) {
if (key.auth_type !== 'oauth') return null
// 即使没有 expires_at,也要检查 invalid_at
if (!key.oauth_expires_at && !key.oauth_invalid_at) return null
return getOAuthExpiresCountdown(
key.oauth_expires_at,
countdownTick.value,
key.oauth_invalid_at,
key.oauth_invalid_reason
)
return getOAuthStatusDisplay(key, countdownTick.value)
}
function getOAuthRefreshButtonTitle(key: EndpointAPIKey): string {
return resolveOAuthRefreshButtonTitle(key, countdownTick.value)
}
// OAuth 状态的 title 提示
function getOAuthStatusTitle(key: EndpointAPIKey): string {
const status = getKeyOAuthExpires(key)
if (!status) return ''
if (status.isInvalid) {
const cleaned = status.invalidReason && isAccountLevelBlockReason(status.invalidReason)
? cleanAccountBlockReason(status.invalidReason)
: status.invalidReason
return cleaned ? `Token 已失效: ${cleaned}` : 'Token 已失效'
const accountTitle = getAccountStatusTitle(key)
if (accountTitle && isAccountLevelBlock(key)) {
return accountTitle
}
if (status.isExpired) {
return 'Token 已过期,请重新授权'
}
return `Token 剩余有效期: ${status.text}`
return resolveOAuthStatusTitle(key, countdownTick.value)
}
// 健康度颜色
@@ -45,8 +45,11 @@
<SelectItem value="vertex_ai">
Vertex AI
</SelectItem>
<SelectItem value="claude_code">
ClaudeCode
<SelectItem
value="claude_code"
disabled
>
ClaudeCode(暂不可用)
</SelectItem>
<SelectItem value="codex">
Codex
@@ -432,6 +435,11 @@ watch(() => form.value.provider_type, () => {
// 提交表单
const handleSubmit = async () => {
if (!isEditMode.value && form.value.provider_type === 'claude_code') {
showError('ClaudeCode 提供商类型暂时禁用', '验证失败')
return
}
// 月卡类型必须设置周期开始时间
if (form.value.billing_type === 'monthly_quota' && !form.value.quota_last_reset_at) {
showError('月卡类型必须设置周期开始时间', '验证失败')
@@ -0,0 +1,233 @@
import type { BodyRuleCondition, BodyRuleConditionOp } from '@/api/endpoints'
export type ConditionSource = 'current' | 'original'
export type ConditionGroupMode = 'all' | 'any'
export interface EditableConditionLeaf {
kind: 'leaf'
path: string
op: BodyRuleConditionOp
value: string
source: ConditionSource
}
export interface EditableConditionGroup {
kind: 'group'
mode: ConditionGroupMode
children: EditableConditionNode[]
}
export type EditableConditionNode = EditableConditionLeaf | EditableConditionGroup
export const CONDITION_OP_OPTIONS: Array<{ value: BodyRuleConditionOp; label: string }> = [
{ value: 'eq', label: '等于' },
{ value: 'neq', label: '不等于' },
{ value: 'gt', label: '大于' },
{ value: 'lt', label: '小于' },
{ value: 'gte', label: '大于等于' },
{ value: 'lte', label: '小于等于' },
{ value: 'starts_with', label: '开头匹配' },
{ value: 'ends_with', label: '结尾匹配' },
{ value: 'contains', label: '包含' },
{ value: 'matches', label: '正则匹配' },
{ value: 'exists', label: '存在' },
{ value: 'not_exists', label: '不存在' },
{ value: 'in', label: '在列表中' },
{ value: 'type_is', label: '类型是' },
]
const NUMERIC_OPS = new Set(['gt', 'lt', 'gte', 'lte'])
const STRING_OPS = new Set(['starts_with', 'ends_with'])
const TYPE_IS_VALUES = new Set(['string', 'number', 'boolean', 'array', 'object', 'null'])
export function createEmptyConditionLeaf(): EditableConditionLeaf {
return {
kind: 'leaf',
path: '',
op: 'eq',
value: '',
source: 'current',
}
}
export function createConditionGroup(
mode: ConditionGroupMode = 'all',
children: EditableConditionNode[] = [createEmptyConditionLeaf()],
): EditableConditionGroup {
return {
kind: 'group',
mode,
children,
}
}
export function cloneEditableCondition(node: EditableConditionNode): EditableConditionNode {
if (node.kind === 'group') {
return {
kind: 'group',
mode: node.mode,
children: node.children.map(cloneEditableCondition),
}
}
return { ...node }
}
export function conditionToEditable(condition?: BodyRuleCondition | null): EditableConditionNode | null {
if (!condition) return null
if ('all' in condition) {
return createConditionGroup(
'all',
condition.all.map(child => conditionToEditable(child) || createEmptyConditionLeaf()),
)
}
if ('any' in condition) {
return createConditionGroup(
'any',
condition.any.map(child => conditionToEditable(child) || createEmptyConditionLeaf()),
)
}
return {
kind: 'leaf',
path: condition.path || '',
op: condition.op || 'eq',
value: condition.value !== undefined
? (typeof condition.value === 'string' ? condition.value : JSON.stringify(condition.value))
: '',
source: condition.source === 'original' ? 'original' : 'current',
}
}
export function editableConditionToApi(node: EditableConditionNode | null): BodyRuleCondition | undefined {
if (!node) return undefined
if (node.kind === 'group') {
const children = node.children
.map(child => editableConditionToApi(child))
.filter((child): child is BodyRuleCondition => !!child)
if (!children.length) return undefined
return node.mode === 'all' ? { all: children } : { any: children }
}
const path = node.path.trim()
if (!path) return undefined
const base = {
path,
op: node.op,
...(node.source === 'original' ? { source: 'original' as const } : {}),
}
if (node.op === 'exists' || node.op === 'not_exists') {
return base
}
const raw = node.value.trim()
if (!raw) {
return { ...base, value: '' }
}
try {
return { ...base, value: JSON.parse(raw) }
} catch {
return { ...base, value: raw }
}
}
export function isConditionValueRequired(op: BodyRuleConditionOp): boolean {
return op !== 'exists' && op !== 'not_exists'
}
export function getConditionValuePlaceholder(op: BodyRuleConditionOp): string {
if (op === 'in') return '["a", "b"]'
if (op === 'type_is') return 'string/number/boolean/...'
return '值'
}
export function getBodyRuleConditionPathPlaceholder(path: string): string {
return path.includes('[*]') || /\[\d+-\d+\]/.test(path) ? '$item.字段名' : '字段路径'
}
export function conditionEquals(
left: EditableConditionNode | null,
right: EditableConditionNode | null,
): boolean {
if (left === right) return true
if (!left || !right) return false
if (left.kind !== right.kind) return false
if (left.kind === 'group' && right.kind === 'group') {
if (left.mode !== right.mode) return false
if (left.children.length !== right.children.length) return false
return left.children.every((child, i) => conditionEquals(child, right.children[i]))
}
if (left.kind === 'leaf' && right.kind === 'leaf') {
return left.path === right.path
&& left.op === right.op
&& left.value === right.value
&& left.source === right.source
}
return false
}
export function validateEditableCondition(node: EditableConditionNode | null): string | null {
if (!node) return null
if (node.kind === 'group') {
if (!node.children.length) return '组合条件至少需要一个子条件'
for (let i = 0; i < node.children.length; i += 1) {
const err = validateEditableCondition(node.children[i])
if (err) return `子条件 ${i + 1}: ${err}`
}
return null
}
const path = node.path.trim()
if (!path) return '条件路径不能为空'
if (!isConditionValueRequired(node.op)) return null
const raw = node.value.trim()
let parsed: unknown = raw
if (raw) {
try {
parsed = JSON.parse(raw)
} catch {
parsed = raw
}
}
if (NUMERIC_OPS.has(node.op)) {
if (typeof parsed !== 'number' || Number.isNaN(parsed)) return '数值条件必须填写数字'
return null
}
if (STRING_OPS.has(node.op)) {
if (typeof parsed !== 'string') return '该条件值必须为字符串'
return null
}
if (node.op === 'matches') {
if (typeof parsed !== 'string' || !parsed) return '正则条件值不能为空'
try {
new RegExp(parsed)
return null
} catch (error: unknown) {
return `正则表达式无效:${error instanceof Error ? error.message : String(error)}`
}
}
if (node.op === 'in') {
if (!Array.isArray(parsed)) return 'in 条件值必须是 JSON 数组'
return null
}
if (node.op === 'type_is') {
if (typeof parsed !== 'string' || !TYPE_IS_VALUES.has(parsed)) {
return 'type_is 仅支持 string/number/boolean/array/object/null'
}
}
return null
}
@@ -165,7 +165,6 @@
<ModelMappingDialog
v-model:open="dialogOpen"
:provider-id="provider.id"
:provider-api-formats="providerApiFormats"
:models="models"
:editing-group="editingGroup"
:preselected-model-id="preselectedModelId"
@@ -195,7 +194,6 @@ import { useToast } from '@/composables/useToast'
import {
getProviderModels,
testModel,
API_FORMAT_LABELS,
type Model,
type ProviderModelAlias
} from '@/api/endpoints'
@@ -227,16 +225,6 @@ const preselectedModelId = ref<string | null>(null)
// 列表展开状态
const expandedAliasGroups = ref<Set<string>>(new Set())
// 获取 Provider 支持的 API 格式
const providerApiFormats = computed(() => {
const formats = props.provider?.api_formats
if (Array.isArray(formats) && formats.length > 0) {
const order = Object.keys(API_FORMAT_LABELS)
return [...formats].sort((a, b) => order.indexOf(a) - order.indexOf(b))
}
return []
})
// 生成作用域唯一键
function getApiFormatsKey(formats: string[] | undefined): string {
if (!formats || formats.length === 0) return ''
@@ -391,7 +379,6 @@ async function testMapping(group: AliasGroup, mapping: ProviderModelAlias) {
const result = await testModel({
provider_id: props.provider.id,
model_name: mapping.name, // 使用映射名称进行测试
message: "hello",
api_format: apiFormat
})
@@ -314,10 +314,24 @@
:result="modelTest.testResult.value"
mode="direct"
:selecting-model-name="testingModelName"
:endpoints="activeEndpoints"
:selected-endpoint="selectedTestEndpoint"
:testing="modelTest.testing.value"
:trace="modelTest.testTrace.value"
:request-id="modelTest.requestId.value"
:request-headers-draft="testRequestHeadersDraft"
:request-headers-reset-value="testRequestHeadersResetValue"
:request-headers-error="testRequestHeadersError"
:request-body-draft="testRequestBodyDraft"
:request-body-reset-value="testRequestBodyResetValue"
:request-body-error="testRequestBodyError"
:start-disabled="!selectedTestEndpoint || !!testRequestHeadersError || !!testRequestBodyError"
@close="handleTestDialogClose"
@back="handleTestDialogBack"
@select-endpoint="handleSelectTestEndpoint"
@start="handleStartMappingTest"
@update:request-headers-draft="testRequestHeadersDraft = $event"
@update:request-body-draft="testRequestBodyDraft = $event"
/>
</template>
@@ -335,6 +349,7 @@ import ModelTestDialog from './ModelTestDialog.vue'
import { useToast } from '@/composables/useToast'
import {
type Model,
type ProviderEndpoint,
type ProviderModelAlias,
type ProviderMappingPreviewResponse,
} from '@/api/endpoints'
@@ -342,6 +357,14 @@ import { type EndpointAPIKey } from '@/api/endpoints/keys'
import { updateModel } from '@/api/endpoints/models'
import { parseApiError } from '@/utils/errorParser'
import type { ProviderWithEndpointsSummary } from '@/api/endpoints'
import {
buildDefaultModelTestRequestHeaders,
buildDefaultModelTestRequestBody,
parseModelTestRequestHeadersDraft,
parseModelTestRequestBodyDraft,
POOL_TEST_CONCURRENCY,
SINGLE_TEST_CONCURRENCY,
} from './model-test-request'
interface MappingItem {
name: string
@@ -369,6 +392,7 @@ interface CombinedMapping {
const props = defineProps<{
provider: ProviderWithEndpointsSummary
endpoints?: ProviderEndpoint[]
providerKeys?: EndpointAPIKey[]
models?: Model[]
mappingPreview?: ProviderMappingPreviewResponse | null
@@ -390,8 +414,20 @@ const deleteConfirmOpen = ref(false)
const editingGroup = ref<AliasGroup | null>(null)
const deletingGroup = ref<AliasGroup | null>(null)
const testingMapping = ref<string | null>(null)
const pendingMappingKey = ref<string | null>(null)
const testingModelName = ref<string | null>(null)
const preselectedModelId = ref<string | null>(null)
const selectedTestEndpoint = ref<ProviderEndpoint | null>(null)
const testRequestHeadersDraft = ref('')
const testRequestHeadersResetValue = ref('')
const testRequestBodyDraft = ref('')
const testRequestBodyResetValue = ref('')
const isPoolManagedProvider = computed(() => Boolean(props.provider.pool_advanced))
const activeEndpoints = computed(() => (props.endpoints ?? []).filter(endpoint => endpoint.is_active))
const parsedTestRequestHeaders = computed(() => parseModelTestRequestHeadersDraft(testRequestHeadersDraft.value))
const testRequestHeadersError = computed(() => parsedTestRequestHeaders.value.error)
const parsedTestRequestBody = computed(() => parseModelTestRequestBodyDraft(testRequestBodyDraft.value))
const testRequestBodyError = computed(() => parsedTestRequestBody.value.error)
// 使用 props 传入的数据
const models = computed(() => props.models ?? [])
@@ -626,22 +662,81 @@ async function onDialogSaved() {
function handleTestDialogClose() {
modelTest.resetState()
pendingMappingKey.value = null
testingModelName.value = null
testingMapping.value = null
selectedTestEndpoint.value = null
testRequestHeadersDraft.value = ''
testRequestHeadersResetValue.value = ''
testRequestBodyDraft.value = ''
testRequestBodyResetValue.value = ''
}
function handleTestDialogBack() {
if (modelTest.testing.value) return
modelTest.testResult.value = null
modelTest.stopPolling()
}
function handleSelectTestEndpoint(endpointId: string) {
const endpoint = activeEndpoints.value.find(item => item.id === endpointId)
if (!endpoint) return
selectedTestEndpoint.value = endpoint
}
// 测试映射(直连测试,带故障转移和实时进度)
async function runMappingTest(testingKey: string, modelName: string) {
testingMapping.value = testingKey
function runMappingTest(testingKey: string, modelName: string) {
if (activeEndpoints.value.length === 0) {
showError('暂无可用于测试的活跃端点')
return
}
pendingMappingKey.value = testingKey
modelTest.testResult.value = null
modelTest.dialogOpen.value = true
testingMapping.value = null
testingModelName.value = modelName
selectedTestEndpoint.value = activeEndpoints.value[0] ?? null
testRequestHeadersResetValue.value = buildDefaultModelTestRequestHeaders()
testRequestHeadersDraft.value = testRequestHeadersResetValue.value
testRequestBodyResetValue.value = buildDefaultModelTestRequestBody(modelName)
testRequestBodyDraft.value = testRequestBodyResetValue.value
}
async function handleStartMappingTest() {
if (modelTest.testing.value || !testingModelName.value) return
const endpoint = selectedTestEndpoint.value || activeEndpoints.value[0]
if (!endpoint) {
showError('请选择要测试的端点')
return
}
const { value: requestHeaders, error: requestHeadersError } = parsedTestRequestHeaders.value
if (!requestHeaders || requestHeadersError) {
showError(`测试请求头无效: ${requestHeadersError || '无效 JSON'}`)
return
}
const { value: requestBody, error } = parsedTestRequestBody.value
if (!requestBody || error) {
showError(`测试请求体无效: ${error || '无效 JSON'}`)
return
}
const currentMappingKey = pendingMappingKey.value || testingModelName.value
testingMapping.value = pendingMappingKey.value ? currentMappingKey : null
await modelTest.startTest({
mode: 'direct',
modelName,
displayLabel: `映射 "${modelName}"`,
message: 'hello',
onSuccess: () => {
testingModelName.value = null
},
modelName: testingModelName.value,
displayLabel: `[${endpoint.api_format}] 映射 "${testingModelName.value}"`,
apiFormat: endpoint.api_format,
endpointId: endpoint.id,
requestHeaders,
requestBody,
concurrency: isPoolManagedProvider.value ? POOL_TEST_CONCURRENCY : SINGLE_TEST_CONCURRENCY,
})
if (pendingMappingKey.value === currentMappingKey) {
pendingMappingKey.value = null
}
testingMapping.value = null
}
File diff suppressed because it is too large Load Diff
@@ -221,10 +221,19 @@
:testing="modelTest.testing.value"
:trace="modelTest.testTrace.value"
:request-id="modelTest.requestId.value"
:show-endpoint-selector="activeEndpoints.length > 1"
:request-headers-draft="testRequestHeadersDraft"
:request-headers-reset-value="testRequestHeadersResetValue"
:request-headers-error="testRequestHeadersError"
:request-body-draft="testRequestBodyDraft"
:request-body-reset-value="testRequestBodyResetValue"
:request-body-error="testRequestBodyError"
:start-disabled="!selectedTestEndpoint || !!testRequestHeadersError || !!testRequestBodyError"
@close="handleTestDialogClose"
@back="handleTestDialogBack"
@start="handleStartPendingTest"
@select-endpoint="handleSelectTestEndpoint"
@update:request-headers-draft="testRequestHeadersDraft = $event"
@update:request-body-draft="testRequestBodyDraft = $event"
/>
</template>
@@ -247,6 +256,14 @@ import { parseApiError } from '@/utils/errorParser'
import { formatApiFormat } from '@/api/endpoints/types/api-format'
import type { ProviderWithEndpointsSummary } from '@/api/endpoints'
import ModelTestDialog from './ModelTestDialog.vue'
import {
buildDefaultModelTestRequestHeaders,
buildDefaultModelTestRequestBody,
parseModelTestRequestHeadersDraft,
parseModelTestRequestBodyDraft,
POOL_TEST_CONCURRENCY,
SINGLE_TEST_CONCURRENCY,
} from './model-test-request'
const props = defineProps<{
provider: ProviderWithEndpointsSummary
@@ -272,7 +289,16 @@ const localModels = ref<Model[]>([])
const togglingModelId = ref<string | null>(null)
const pendingTestModel = ref<Model | null>(null)
const selectedTestEndpoint = ref<ProviderEndpoint | null>(null)
const testRequestHeadersDraft = ref('')
const testRequestHeadersResetValue = ref('')
const testRequestBodyDraft = ref('')
const testRequestBodyResetValue = ref('')
const isPoolManagedProvider = computed(() => Boolean(props.provider.pool_advanced))
const activeEndpoints = computed(() => (props.endpoints ?? []).filter(endpoint => endpoint.is_active))
const parsedTestRequestHeaders = computed(() => parseModelTestRequestHeadersDraft(testRequestHeadersDraft.value))
const testRequestHeadersError = computed(() => parsedTestRequestHeaders.value.error)
const parsedTestRequestBody = computed(() => parseModelTestRequestBodyDraft(testRequestBodyDraft.value))
const testRequestBodyError = computed(() => parsedTestRequestBody.value.error)
const models = computed(() => props.models ?? localModels.value)
// 按名称排序的模型列表
const sortedModels = computed(() => {
@@ -438,19 +464,47 @@ function handleTestDialogClose() {
modelTest.resetState()
pendingTestModel.value = null
selectedTestEndpoint.value = null
testRequestHeadersDraft.value = ''
testRequestHeadersResetValue.value = ''
testRequestBodyDraft.value = ''
testRequestBodyResetValue.value = ''
}
function handleTestDialogBack() {
if (modelTest.testing.value) return
modelTest.testResult.value = null
selectedTestEndpoint.value = null
modelTest.stopPolling()
}
async function handleSelectTestEndpoint(endpointId: string) {
if (!pendingTestModel.value) return
function handleSelectTestEndpoint(endpointId: string) {
const endpoint = activeEndpoints.value.find(item => item.id === endpointId)
if (!endpoint) return
selectedTestEndpoint.value = endpoint
}
async function handleStartPendingTest() {
if (modelTest.testing.value) return
if (!pendingTestModel.value) return
const endpoint = selectedTestEndpoint.value || activeEndpoints.value[0]
if (!endpoint) {
showError('请选择要测试的端点')
return
}
const { value: requestHeaders, error: requestHeadersError } = parsedTestRequestHeaders.value
if (!requestHeaders || requestHeadersError) {
showError(`测试请求头无效: ${requestHeadersError || '无效 JSON'}`)
return
}
const { value: requestBody, error } = parsedTestRequestBody.value
if (!requestBody || error) {
showError(`测试请求体无效: ${error || '无效 JSON'}`)
return
}
selectedTestEndpoint.value = endpoint
const model = pendingTestModel.value
const modelName = model.global_model_name || model.provider_model_name
const endpointPrefix = `[${formatApiFormat(endpoint.api_format)}] `
@@ -460,15 +514,11 @@ async function handleSelectTestEndpoint(endpointId: string) {
displayLabel: `${endpointPrefix}${modelName}`,
apiFormat: endpoint.api_format,
endpointId: endpoint.id,
message: 'hello',
concurrency: 5,
onSuccess: () => {
pendingTestModel.value = null
selectedTestEndpoint.value = null
},
requestHeaders,
requestBody,
concurrency: isPoolManagedProvider.value ? POOL_TEST_CONCURRENCY : SINGLE_TEST_CONCURRENCY,
onError: () => {
if (activeEndpoints.value.length > 1) {
selectedTestEndpoint.value = null
return true
}
},
@@ -484,13 +534,15 @@ async function testModelConnection(model: Model) {
}
pendingTestModel.value = model
selectedTestEndpoint.value = null
selectedTestEndpoint.value = activeEndpoints.value[0] ?? null
testRequestHeadersResetValue.value = buildDefaultModelTestRequestHeaders()
testRequestHeadersDraft.value = testRequestHeadersResetValue.value
testRequestBodyResetValue.value = buildDefaultModelTestRequestBody(
model.global_model_name || model.provider_model_name,
)
testRequestBodyDraft.value = testRequestBodyResetValue.value
modelTest.testResult.value = null
modelTest.dialogOpen.value = true
if (activeEndpoints.value.length === 1) {
await handleSelectTestEndpoint(activeEndpoints.value[0].id)
}
}
// 暴露给父组件
@@ -0,0 +1,80 @@
const DEFAULT_MODEL_TEST_MESSAGE = 'Hello! This is a test message.'
/** Pool-managed provider runs concurrent checks; single-key provider does not. */
export const POOL_TEST_CONCURRENCY = 5
export const SINGLE_TEST_CONCURRENCY = 1
export function buildDefaultModelTestRequestBody(modelName: string): string {
return JSON.stringify({
model: modelName,
messages: [
{
role: 'user',
content: DEFAULT_MODEL_TEST_MESSAGE,
},
],
max_tokens: 30,
temperature: 0.7,
stream: true,
}, null, 2)
}
export function buildDefaultModelTestRequestHeaders(): string {
return JSON.stringify({}, null, 2)
}
function parseModelTestJsonObjectDraft(
draft: string,
options: {
emptyValue: Record<string, unknown> | null
emptyError: string | null
invalidTypeError: string
},
): { value: Record<string, unknown> | null; error: string | null } {
const normalized = draft.trim()
if (!normalized) {
return {
value: options.emptyValue,
error: options.emptyError,
}
}
try {
const parsed = JSON.parse(normalized)
if (!parsed || typeof parsed !== 'object' || Array.isArray(parsed)) {
return {
value: null,
error: options.invalidTypeError,
}
}
return {
value: parsed as Record<string, unknown>,
error: null,
}
} catch (error) {
return {
value: null,
error: error instanceof Error ? error.message : '无效的 JSON',
}
}
}
export function parseModelTestRequestBodyDraft(
draft: string,
): { value: Record<string, unknown> | null; error: string | null } {
return parseModelTestJsonObjectDraft(draft, {
emptyValue: null,
emptyError: '测试请求体不能为空',
invalidTypeError: '测试请求体必须是 JSON 对象',
})
}
export function parseModelTestRequestHeadersDraft(
draft: string,
): { value: Record<string, unknown> | null; error: string | null } {
return parseModelTestJsonObjectDraft(draft, {
emptyValue: {},
emptyError: null,
invalidTypeError: '测试请求头必须是 JSON 对象',
})
}
@@ -20,26 +20,31 @@
</div>
<div
v-if="isLoading"
class="h-full min-h-[160px] flex items-center justify-center text-sm text-muted-foreground"
class="flex items-center justify-center text-sm text-muted-foreground"
:style="{ minHeight: `${minHeight}px` }"
>
<Loader2 class="h-5 w-5 animate-spin mr-2" />
加载中...
</div>
<div
v-else-if="hasError"
class="h-full min-h-[160px] flex items-center justify-center text-sm text-destructive"
class="flex items-center justify-center text-sm text-destructive"
:style="{ minHeight: `${minHeight}px` }"
>
<AlertCircle class="h-4 w-4 mr-1.5" />
加载失败
数据暂不可用
</div>
<ActivityHeatmap
v-else-if="hasData"
:data="data"
:show-header="false"
:value-label="valueLabel"
:actual-value-label="actualValueLabel"
/>
<div
v-else
class="h-full min-h-[160px] flex items-center justify-center text-sm text-muted-foreground"
class="flex items-center justify-center text-sm text-muted-foreground"
:style="{ minHeight: `${minHeight}px` }"
>
暂无活跃数据
</div>
@@ -58,6 +63,9 @@ const props = defineProps<{
title: string
isLoading?: boolean
hasError?: boolean
minHeight?: number
valueLabel?: string
actualValueLabel?: string
}>()
const legendLevels = [0.08, 0.25, 0.45, 0.65, 0.85]
@@ -65,4 +73,6 @@ const legendLevels = [0.08, 0.25, 0.45, 0.65, 0.85]
const hasData = computed(() =>
props.data && props.data.days && props.data.days.length > 0
)
const minHeight = computed(() => props.minHeight ?? 160)
</script>
@@ -240,6 +240,10 @@
-->{{ proxyTimingBreakdown(currentAttempt.extra_data.proxy) }}<!--
-->)</span>
</span>
<code
v-if="typeof currentAttempt.extra_data.proxy.node_id === 'string' && currentAttempt.extra_data.proxy.node_id"
class="text-xs font-mono text-muted-foreground"
>节点 Key {{ currentAttempt.extra_data.proxy.node_id }}</code>
</span>
</div>
<div
@@ -462,7 +466,7 @@ interface UsageData {
}
const props = defineProps<{
requestId: string
requestId?: string | null
/** 外部传入的状态码,用于覆盖 trace.final_status 的判断 */
overrideStatusCode?: number
/** 请求侧 API 格式(客户端入口格式) */
@@ -471,6 +475,12 @@ const props = defineProps<{
usageData?: UsageData | null
/** 请求元数据(用于号池调度组装) */
requestMetadata?: Record<string, unknown> | null
/** 已获取的追踪数据;传入时不再内部拉取 */
traceData?: RequestTrace | null
}>()
const emit = defineEmits<{
selectAttempt: [attempt: CandidateRecord | null]
}>()
// 用量数据(从 props 获取)
@@ -525,7 +535,8 @@ const getFinalStatusBadgeVariant = (status: string): BadgeVariant => {
const loading = ref(false)
const error = ref<string | null>(null)
const trace = ref<RequestTrace | null>(null)
const internalTrace = ref<RequestTrace | null>(null)
const trace = computed(() => props.traceData ?? internalTrace.value)
const selectedGroupIndex = ref(0)
const selectedAttemptIndex = ref(0)
const hoveredGroupIndex = ref<number | null>(null)
@@ -769,11 +780,11 @@ const poolAttemptCandidates = computed<CandidateRecord[]>(() => {
}
merged.status = normalizeTimelineStatus(raw.status ?? merged.status)
if (typeof raw.provider_id === 'string') merged.provider_id = raw.provider_id
if (typeof raw.provider_name === 'string') merged.provider_name = raw.provider_name
if (typeof raw.endpoint_id === 'string') merged.endpoint_id = raw.endpoint_id
if (typeof raw.key_id === 'string') merged.key_id = raw.key_id
if (typeof raw.key_name === 'string') merged.key_name = raw.key_name
if (!merged.provider_id && typeof raw.provider_id === 'string') merged.provider_id = raw.provider_id
if (!merged.provider_name && typeof raw.provider_name === 'string') merged.provider_name = raw.provider_name
if (!merged.endpoint_id && typeof raw.endpoint_id === 'string') merged.endpoint_id = raw.endpoint_id
if (!merged.key_id && typeof raw.key_id === 'string') merged.key_id = raw.key_id
if (!merged.key_name && typeof raw.key_name === 'string') merged.key_name = raw.key_name
if (typeof raw.status_code === 'number') merged.status_code = raw.status_code
if (typeof raw.error_type === 'string') merged.error_type = raw.error_type
const rawPoolGroupId = typeof raw.pool_group_id === 'string' ? raw.pool_group_id.trim() : ''
@@ -1045,6 +1056,10 @@ const currentAttempt = computed(() => {
return selectedGroup.value.allAttempts[selectedAttemptIndex.value] || selectedGroup.value.primary
})
watch(currentAttempt, (attempt) => {
emit('selectAttempt', attempt ?? null)
}, { immediate: true })
const currentGroupTitle = computed(() => {
if (!selectedGroup.value || !currentAttempt.value) return ''
if (selectedGroup.value.isPoolGroup) {
@@ -1224,7 +1239,7 @@ const navigateGroup = (direction: number) => {
// 加载请求追踪数据
const isSilentRefresh = ref(false)
const loadTrace = async (silent = false) => {
if (!props.requestId) return
if (!props.requestId || props.traceData) return
isSilentRefresh.value = silent
@@ -1234,7 +1249,7 @@ const loadTrace = async (silent = false) => {
error.value = null
try {
trace.value = await requestTraceApi.getRequestTrace(props.requestId)
internalTrace.value = await requestTraceApi.getRequestTrace(props.requestId)
} catch (err: unknown) {
if (!silent) {
error.value = parseApiError(err, '加载失败')
@@ -1304,12 +1319,31 @@ watch(groupedTimeline, (newGroups) => {
selectedAttemptIndex.value = 0
}, { immediate: true })
// 监听 requestId 变化
watch(() => props.requestId, () => {
selectedGroupIndex.value = 0
selectedAttemptIndex.value = 0
loadTrace()
}, { immediate: true })
// 监听 requestId / 外部 trace 变化
watch(
[() => props.requestId, () => props.traceData],
() => {
selectedGroupIndex.value = 0
selectedAttemptIndex.value = 0
if (props.traceData) {
internalTrace.value = null
loading.value = false
error.value = null
return
}
if (!props.requestId) {
internalTrace.value = null
loading.value = false
error.value = null
return
}
void loadTrace()
},
{ immediate: true },
)
defineExpose({ refresh: () => loadTrace(true) })
@@ -1,325 +0,0 @@
<template>
<Card class="p-4">
<div class="flex items-center justify-between mb-3">
<p class="text-sm font-semibold">
{{ title }}
</p>
<div
v-if="displayLegendItems.length > 0"
class="flex items-center gap-2 flex-wrap justify-end text-[11px]"
>
<div
v-for="item in displayLegendItems"
:key="item.id"
class="flex items-center gap-1"
>
<div
class="w-2.5 h-2.5 rounded-full"
:style="{ backgroundColor: item.color }"
/>
<span class="text-muted-foreground">{{ item.name }}</span>
</div>
<span
v-if="hiddenLegendCount > 0"
class="text-muted-foreground"
>
+{{ hiddenLegendCount }} 更多
</span>
</div>
</div>
<div
v-if="loading"
class="h-[160px] flex items-center justify-center"
>
<div class="text-sm text-muted-foreground">
Loading...
</div>
</div>
<div
v-else-if="hasData"
class="h-[160px]"
>
<ScatterChart
:data="chartData"
:options="chartOptions"
/>
</div>
<div
v-else
class="h-[160px] flex items-center justify-center text-sm text-muted-foreground"
>
暂无请求间隔数据
</div>
</Card>
</template>
<script setup lang="ts">
import { computed, ref, onMounted, onBeforeUnmount, watch } from 'vue'
import Card from '@/components/ui/card.vue'
import ScatterChart from '@/components/charts/ScatterChart.vue'
import { cacheAnalysisApi, type IntervalTimelineResponse } from '@/api/cache'
import { meApi } from '@/api/me'
import type { ChartData, ChartOptions } from 'chart.js'
import { log } from '@/utils/logger'
const props = withDefaults(defineProps<{
title: string
isAdmin: boolean
hours?: number
}>(), {
hours: 24 // 默认当天
})
const loading = ref(false)
const timelineData = ref<IntervalTimelineResponse | null>(null)
const primaryColor = ref('201, 100, 66') // 默认主题色
let loadRequestId = 0
const ADMIN_TIMELINE_LIMIT = 1500
const USER_TIMELINE_LIMIT = 1200
// 获取主题色
function getPrimaryColor(): string {
if (typeof window === 'undefined') return '201, 100, 66'
// CSS 变量定义在 body 上,不是 documentElement
const body = document.body
const style = getComputedStyle(body)
const rgb = style.getPropertyValue('--color-primary-rgb').trim()
return rgb || '201, 100, 66'
}
onMounted(() => {
primaryColor.value = getPrimaryColor()
void loadData()
})
// 预定义的颜色列表(用于区分不同用户/模型)
const COLORS = [
'rgba(59, 130, 246, 0.7)', // blue
'rgba(236, 72, 153, 0.7)', // pink
'rgba(34, 197, 94, 0.7)', // green
'rgba(249, 115, 22, 0.7)', // orange
'rgba(168, 85, 247, 0.7)', // purple
'rgba(234, 179, 8, 0.7)', // yellow
'rgba(14, 165, 233, 0.7)', // sky
'rgba(239, 68, 68, 0.7)', // red
'rgba(20, 184, 166, 0.7)', // teal
'rgba(99, 102, 241, 0.7)', // indigo
]
const hasData = computed(() =>
timelineData.value && timelineData.value.points && timelineData.value.points.length > 0
)
// 判断是否有多个分组(管理员按用户分组,普通用户按模型分组)
const hasMultipleGroups = computed(() => {
if (props.isAdmin) {
// 管理员视图:按用户分组
return timelineData.value?.users && Object.keys(timelineData.value.users).length > 1
} else {
// 用户视图:按模型分组
return timelineData.value?.models && timelineData.value.models.length > 1
}
})
// 图例最多显示数量
const MAX_LEGEND_ITEMS = 6
// 图例项(管理员显示用户,普通用户显示模型)
const legendItems = computed(() => {
if (props.isAdmin && timelineData.value?.users) {
// 管理员视图:显示用户图例
const users = Object.entries(timelineData.value.users)
if (users.length <= 1) return []
return users.map(([userId, username], index) => ({
id: userId,
name: username || userId.slice(0, 8),
color: COLORS[index % COLORS.length]
}))
} else if (timelineData.value?.models && timelineData.value.models.length > 1) {
// 用户视图:显示模型图例
return timelineData.value.models.map((model, index) => ({
id: model,
name: formatModelName(model),
color: COLORS[index % COLORS.length]
}))
}
return []
})
// 显示的图例项(限制数量)
const displayLegendItems = computed(() => {
return legendItems.value.slice(0, MAX_LEGEND_ITEMS)
})
// 隐藏的图例数量
const hiddenLegendCount = computed(() => {
return Math.max(0, legendItems.value.length - MAX_LEGEND_ITEMS)
})
// 格式化模型名称,使其更简洁
function formatModelName(model: string): string {
// 常见的简化规则
if (model.includes('claude')) {
// claude-3-5-sonnet-20241022 -> Claude 3.5 Sonnet
const match = model.match(/claude-(\d+)-(\d+)?-?(\w+)?/i)
if (match) {
const major = match[1]
const minor = match[2]
const variant = match[3]
let name = `Claude ${major}`
if (minor) name += `.${minor}`
if (variant) name += ` ${variant.charAt(0).toUpperCase() + variant.slice(1)}`
return name
}
}
// 其他模型保持原样但截断
return model.length > 20 ? `${model.slice(0, 17) }...` : model
}
// 构建图表数据
const chartData = computed<ChartData<'scatter'>>(() => {
if (!timelineData.value?.points) {
return { datasets: [] }
}
const points = timelineData.value.points
// 管理员视图且有多个用户:按用户分组
if (props.isAdmin && timelineData.value.users && Object.keys(timelineData.value.users).length > 1) {
const userIds = Object.keys(timelineData.value.users)
const userColorMap: Record<string, string> = {}
userIds.forEach((userId, index) => {
userColorMap[userId] = COLORS[index % COLORS.length]
})
// 按用户分组数据
const groupedData: Record<string, Array<{ x: string; y: number }>> = {}
for (const point of points) {
const userId = point.user_id || 'unknown'
if (!groupedData[userId]) {
groupedData[userId] = []
}
groupedData[userId].push({ x: point.x, y: point.y })
}
// 创建每个用户的 dataset
const datasets = Object.entries(groupedData).map(([userId, data]) => ({
label: timelineData.value?.users?.[userId] || userId.slice(0, 8),
data,
backgroundColor: userColorMap[userId] || 'rgba(59, 130, 246, 0.6)',
borderColor: userColorMap[userId] || 'rgba(59, 130, 246, 0.8)',
pointRadius: 3,
pointHoverRadius: 5,
}))
return { datasets }
}
// 用户视图且有多个模型:按模型分组
if (!props.isAdmin && timelineData.value.models && timelineData.value.models.length > 1) {
const models = timelineData.value.models
const modelColorMap: Record<string, string> = {}
models.forEach((model, index) => {
modelColorMap[model] = COLORS[index % COLORS.length]
})
// 按模型分组数据
const groupedData: Record<string, Array<{ x: string; y: number }>> = {}
for (const point of points) {
const model = point.model || 'unknown'
if (!groupedData[model]) {
groupedData[model] = []
}
groupedData[model].push({ x: point.x, y: point.y })
}
// 创建每个模型的 dataset
const datasets = Object.entries(groupedData).map(([model, data]) => ({
label: formatModelName(model),
data,
backgroundColor: modelColorMap[model] || 'rgba(59, 130, 246, 0.6)',
borderColor: modelColorMap[model] || 'rgba(59, 130, 246, 0.8)',
pointRadius: 3,
pointHoverRadius: 5,
}))
return { datasets }
}
// 单用户或单模型:使用主题色
return {
datasets: [{
label: '请求间隔',
data: points.map(p => ({ x: p.x, y: p.y })),
backgroundColor: `rgba(${primaryColor.value}, 0.6)`,
borderColor: `rgba(${primaryColor.value}, 0.8)`,
pointRadius: 3,
pointHoverRadius: 5,
}]
}
})
const chartOptions = computed<ChartOptions<'scatter'>>(() => ({
plugins: {
legend: {
display: false // 使用自定义图例
},
tooltip: {
callbacks: {
label: (context) => {
const point = context.raw as { x: string; y: number; _originalY?: number }
const realY = point._originalY ?? point.y
const datasetLabel = context.dataset.label || ''
if (hasMultipleGroups.value) {
return `${datasetLabel}: ${realY.toFixed(1)} 分钟`
}
return `间隔: ${realY.toFixed(1)} 分钟`
}
}
}
}
}))
async function loadData() {
const requestId = ++loadRequestId
loading.value = true
try {
const limit = props.isAdmin ? ADMIN_TIMELINE_LIMIT : USER_TIMELINE_LIMIT
if (props.isAdmin) {
// 管理员:获取所有用户数据(按比例采样)
const data = await cacheAnalysisApi.getIntervalTimeline({
hours: props.hours,
include_user_info: true,
limit,
})
if (requestId !== loadRequestId) return
timelineData.value = data
} else {
// 普通用户:获取自己的数据
const data = await meApi.getIntervalTimeline({
hours: props.hours,
limit,
})
if (requestId !== loadRequestId) return
timelineData.value = data
}
} catch (error) {
if (requestId !== loadRequestId) return
log.error('加载请求间隔时间线失败:', error)
timelineData.value = null
} finally {
if (requestId === loadRequestId) {
loading.value = false
}
}
}
watch([() => props.hours, () => props.isAdmin], () => {
void loadData()
})
onBeforeUnmount(() => {
loadRequestId++
})
</script>
@@ -239,6 +239,35 @@
<!-- 响应结果 -->
<template v-if="replayResult">
<div
v-if="replayResult.mapping"
class="border-b"
>
<div class="px-4 py-2 text-[11px] text-muted-foreground flex flex-col gap-1">
<div class="flex flex-wrap items-center gap-2">
<span class="font-mono truncate">{{ replayResult.mapping.source_model }}</span>
<span class="text-muted-foreground/50">→</span>
<span class="font-mono truncate">{{ replayResult.mapping.resolved_model }}</span>
<Badge
variant="outline"
class="text-[10px] px-1.5 py-0 h-4"
>
{{ formatReplayMode(replayResult.mapping.replay_mode) }}
</Badge>
</div>
<div class="flex flex-wrap gap-2 text-muted-foreground/60">
<span>Provider: {{ replayResult.mapping.target_provider }}</span>
<span>Endpoint: {{ replayResult.mapping.target_endpoint_id }}</span>
<span>Format: {{ replayResult.mapping.target_api_format || '-' }}</span>
</div>
<div
v-if="replayResult.mapping.mapping_source && replayResult.mapping.mapping_source !== 'none'"
class="text-muted-foreground/50"
>
{{ formatMappingSource(replayResult.mapping.mapping_source) }}
</div>
</div>
</div>
<!-- 响应头(可折叠) -->
<div class="border-b">
<button
@@ -293,7 +322,7 @@
<script setup lang="ts">
import { ref, computed, watch } from 'vue'
import { dashboardApi, type ReplayResponse, type RequestDetail } from '@/api/dashboard'
import { requestDetailsApi, type ReplayResponse, type RequestDetail } from '@/api/request-details'
import { getProvidersSummary } from '@/api/endpoints/providers'
import { getProviderKeys } from '@/api/endpoints/keys'
import type { EndpointAPIKey } from '@/api/endpoints/types'
@@ -376,6 +405,32 @@ const responseHeaderCount = computed(() => {
return Object.keys(replayResult.value.response_headers).length
})
function formatReplayMode(mode?: string) {
switch (mode) {
case 'same_endpoint_reuse':
return '同端点复用'
case 'same_provider_remap':
return '同 Provider 重映射'
case 'cross_provider_remap':
return '跨 Provider 重映射'
default:
return mode || '-'
}
}
function formatMappingSource(source?: string) {
switch (source) {
case 'original_target_model':
return '映射来源: 复用原目标模型'
case 'model_mapping':
return '映射来源: 模型映射'
case 'none':
return '映射来源: 未映射'
default:
return source ? `映射来源: ${source}` : ''
}
}
// ---- 生命周期 ----
watch(() => props.isOpen, async (isOpen) => {
@@ -427,7 +482,7 @@ async function doReplay() {
if (selectedProviderId.value) params.provider_id = selectedProviderId.value
if (selectedKeyId.value) params.api_key_id = selectedKeyId.value
replayResult.value = await dashboardApi.replayRequest(
replayResult.value = await requestDetailsApi.replayRequest(
props.requestId,
Object.keys(params).length > 0 ? params : undefined,
)

Some files were not shown because too many files have changed in this diff Show More