Commit Graph

75 Commits

Author SHA1 Message Date
fawney19
09e0f594ff refactor: 重构限流系统和健康监控,支持按 API 格式区分
- 将 adaptive_concurrency 重命名为 adaptive_rpm,从并发控制改为 RPM 控制
- 健康监控器支持按 API 格式独立管理健康度和熔断器状态
- 新增 model_permissions 模块,支持按格式配置允许的模型
- 重构前端提供商相关表单组件,新增 Collapsible UI 组件
- 新增数据库迁移脚本支持新的数据结构
2026-01-10 18:48:35 +08:00
fawney19
ea35efe440 perf: 优化 HTTP 客户端连接池复用
- 新增 get_proxy_client() 方法,相同代理配置复用同一客户端
- 添加 LRU 淘汰策略,代理客户端上限 50 个防止内存泄漏
- 新增 get_default_client_async() 异步线程安全版本
- 使用模块级锁避免类属性初始化竞态条件
- 优化 ConcurrencyManager 使用 Redis MGET 批量获取减少往返
- 添加 get_pool_stats() 连接池统计信息接口
2026-01-08 13:34:59 +08:00
fawney19
1521ce5a96 feat: 添加负载均衡调度模式
- 新增 load_balance 调度模式,同优先级内随机轮换
- 前端支持三种调度模式切换:缓存亲和、负载均衡、固定顺序
2026-01-08 03:20:04 +08:00
fawney19
d378630b38 perf: 添加多层缓存优化减少数据库查询
- 新增 ProviderCacheService 缓存 Provider 和 ProviderAPIKey 数据
- SystemConfigService 添加进程内缓存(TTL 60秒)
- API Key last_used_at 更新添加节流策略(60秒间隔)
- HTTP 连接池配置改为可配置,支持根据 Worker 数量自动计算
- 前端优先级管理改用 health_score 显示健康度
2026-01-08 02:34:59 +08:00
fawney19
e33d5b952c feat: 所有计费模板支持按次计费,调整端点默认重试次数为 2
- 为 Claude、OpenAI、豆包、Gemini 计费模板添加 request 维度
- 支持通过 price_per_request 配置按次计费(如图片生成模型)
- 将端点 max_retries 默认值从 3 改为 2(请求一次 + 重试一次)
2026-01-07 20:15:30 +08:00
fawney19
4345ac2ba2 fix: 添加系统配置项默认值,避免前端获取配置时报错
添加以下配置项到 DEFAULT_CONFIGS:
- email_suffix_mode/email_suffix_list (邮箱后缀限制)
- audit_log_retention_days (审计日志保留天数)
- smtp_* (SMTP 邮件服务器配置)
2026-01-07 20:01:17 +08:00
fawney19
a12b43ce5c refactor: 清理数据库字段命名歧义
- users 表:重命名 allowed_endpoints 为 allowed_api_formats(修正历史命名错误)
- api_keys 表:删除 allowed_endpoints 字段(未使用的功能)
- providers 表:删除 rate_limit 字段(与 rpm_limit 重复)
- usage 表:重命名 provider 为 provider_name(避免与 provider_id 外键混淆)

同步更新前后端所有相关代码
2026-01-07 19:53:32 +08:00
fawney19
fbe303a3cd fix: 补充 cost.py 中 Provider 未实现模型时的 else 分支处理 2026-01-07 16:13:53 +08:00
fawney19
0061fc04b7 feat: 添加访问令牌管理功能并升级至 0.2.4
- 新增 Management Token(访问令牌)功能,支持创建、更新、删除和管理
- 前端添加访问令牌管理页面,支持普通用户和管理员
- 后端实现完整的令牌生命周期管理 API
- 添加数据库迁移脚本创建 management_tokens 表
- Nginx 配置添加 gzip 压缩,优化响应传输
- Dialog 组件添加 persistent 属性,防止意外关闭
- 为管理后台 API 添加详细的中文文档注释
- 简化多处类型注解,统一代码风格
2026-01-07 14:55:07 +08:00
fawney19
00562dd1d4 feat: 添加 LDAP 认证支持
- 新增 LDAP 服务和 API 接口
- 添加 LDAP 配置管理页面
- 登录页面支持 LDAP/本地认证切换
- 数据库迁移支持 LDAP 相关字段
2026-01-06 14:38:42 +08:00
fawney19
35e29d46bd refactor: 抽取统一计费模块,支持配置驱动的多厂商计费
- 新增 src/services/billing/ 模块,包含计费计算器、模板和使用量映射
- 将 ChatAdapterBase 和 CliAdapterBase 中的计费逻辑重构为调用 billing 模块
- 为每个 adapter 添加 BILLING_TEMPLATE 类属性,指定计费模板
- 支持 Claude/OpenAI/Gemini 三种计费模板,支持阶梯计费和缓存 TTL 定价
- 新增 tests/services/billing/ 单元测试
2026-01-05 16:48:59 +08:00
fawney19
e5f12fddd9 feat: 流式预读增强与自适应并发算法优化
流式预读增强:
- 新增预读字节上限(64KB),防止无换行响应导致内存增长
- 预读结束后检测非 SSE 格式的错误响应(HTML 页面、纯 JSON 错误)
- 抽取 check_html_response 和 check_prefetched_response_error 到 utils.py

自适应并发算法优化(边界记忆 + 渐进探测):
- 缩容策略:从乘性减少改为边界 -1,一次 429 即可收敛到真实限制附近
- 扩容策略:普通扩容不超过已知边界,探测性扩容可谨慎突破(每次 +1)
- 仅在并发限制 429 时记录边界,避免 RPM/UNKNOWN 类型覆盖
2026-01-05 12:17:45 +08:00
fawney19
4fa9a1303a feat: 优化首字时间和 streaming 状态的记录时序
改进 streaming 状态更新机制:
- 统一在首次输出时记录 TTFB 并更新 streaming 状态
- 重构 CliMessageHandlerBase 中的状态更新逻辑,消除重复
- 确保 provider/key 信息在 streaming 状态更新时已可用

前端改进:
- 添加 first_byte_time_ms 字段支持
- 管理员接口支持返回 provider/api_key_name 字段
- 优化活跃请求轮询逻辑,更准确地判断是否需要刷新完整数据

数据库与 API:
- UsageService.get_active_requests_status 添加 include_admin_fields 参数
- 管理员接口调用时启用该参数以获取额外信息
2026-01-05 10:31:34 +08:00
fawney19
b34cfb676d fix: streaming 状态更新时传递 provider 相关 ID 信息
在 update_usage_status 方法中增加 provider_id、provider_endpoint_id
和 provider_api_key_id 参数,确保流式请求进入 streaming 状态时
能正确记录这些字段。
2026-01-05 09:12:03 +08:00
fawney19
3064497636 refactor: 改进上游错误消息的提取和传递
- 新增 extract_error_message 工具函数,统一错误消息提取逻辑
- 在 HTTPStatusError 异常上附加 upstream_response 属性,保留原始错误
- 优先使用上游响应内容作为错误消息,而非异常字符串表示
- 移除错误消息的长度限制(500/1000 字符)
- 修复边界条件检查,使用 startswith 匹配 "Unable to read" 前缀
- 简化 result.py 中的条件判断逻辑
2026-01-05 03:18:55 +08:00
fawney19
e7db76e581 refactor: API Key 过期时间改用日期选择器,rate_limit 支持无限制
- 前端:将过期时间设置从"天数输入"改为"日期选择器",更直观
- 后端:新增 expires_at 字段(ISO 日期格式),兼容旧版 expire_days
- rate_limit 字段现在支持 null 表示无限制,移除默认值 100
- 解析逻辑:过期时间设为当天 UTC 23:59:59.999999
2026-01-05 02:16:16 +08:00
fawney19
689339117a refactor: 提取 ModelMultiSelect 组件并支持失效模型检测
- 新增 ModelMultiSelect 组件,支持显示和移除已失效的模型
- 新增 useInvalidModels composable 检测 allowed_models 中的无效引用
- 重构 StandaloneKeyFormDialog 和 UserFormDialog 使用新组件
- 补充 GlobalModel 删除逻辑的设计说明注释
2026-01-05 01:20:58 +08:00
fawney19
b202765be4 perf: 优化流式响应 TTFB,将数据库状态更新移至 yield 后执行
- StreamUsageTracker: 先 yield 首个 chunk 再更新 streaming 状态
- EnhancedStreamUsageTracker: 同步添加 TTFB 记录和状态更新逻辑
- 确保客户端首字节响应不受数据库操作延迟影响
2026-01-05 00:13:23 +08:00
fawney19
3bbf3073df feat: 所有 Provider 失败时透传上游错误信息
- FallbackOrchestrator 在所有候选组合失败后保留最后的错误信息
- 从 httpx.HTTPStatusError 提取上游状态码和响应内容
- ProviderNotAvailableException 携带上游错误信息
- ErrorResponse 在返回错误时透传上游状态码和响应
2026-01-04 23:50:15 +08:00
fawney19
f46aaa2182 debug: 添加 streaming 状态更新时 provider 为空的调试日志
- base_handler: 更新 streaming 状态时检测并记录 provider 为空的情况
- cli_handler_base: 修复预读数据为空时未更新 streaming 状态的问题
- usage service: 检测状态变为 streaming 但 provider 仍为 pending 的异常
2026-01-04 23:16:01 +08:00
fawney19
a2f33a6c35 perf: 拆分热力图为独立接口并添加 Redis 缓存
- 新增独立热力图 API 端点 (/api/admin/usage/heatmap, /api/users/me/usage/heatmap)
- 添加 Redis 缓存层 (5分钟 TTL),减少数据库查询
- 用户角色变更时清除热力图缓存
- 前端并行加载统计数据和热力图,添加加载/错误状态显示
- 修复 cache_decorator 缺少 JSON 解析错误处理的问题
- 更新 docker-compose 启动命令提示
2026-01-04 22:42:58 +08:00
fawney19
c3a5878b1b feat: 优化用量查询分页和热力图性能
- 用量查询接口添加 limit/offset 分页参数支持
- 热力图统计从实时查询 Usage 表改为读取预计算的 StatsDaily/StatsUserDaily 表
- 修复 avg_response_time_ms 为 0 时被错误跳过的问题
2026-01-04 18:02:47 +08:00
RWDai
3e4309eba3 Enhance LDAP auth config handling 2026-01-04 16:27:02 +08:00
RWDai
414f45aa71 Fix LDAP authentication stability 2026-01-04 13:09:55 +08:00
RWDai
64bfa955f4 feat(ldap): 完善 LDAP 认证功能和安全性
- 添加 LDAP 配置类型定义,移除 any 类型
- 首次配置 LDAP 时强制要求设置绑定密码
- 根据认证类型区分登录标识验证(本地需邮箱,LDAP 允许用户名)
- 添加 LDAP 过滤器转义函数防止注入攻击
- 增加 LDAP 连接超时设置
- 添加账户来源冲突检查,防止 LDAP 覆盖本地账户
- 添加用户名冲突自动重命名机制
2026-01-04 11:18:28 +08:00
RWDai
9bfb295238 feat: add ldap login 2026-01-02 16:17:24 +08:00
fawney19
cddc22d2b3 refactor: 重构邮箱验证模块并修复代码审查问题
- 重构: 将 verification 模块重命名为 email,目录结构更清晰
- 新增: 独立的邮件配置管理页面 (EmailSettings.vue)
- 新增: 邮件模板管理功能(支持自定义 HTML 模板和预览)
- 新增: 查询验证状态 API,支持页面刷新后恢复验证流程
- 新增: 注册邮箱后缀白名单/黑名单限制功能
- 修复: 统一密码最小长度为 6 位(前后端一致)
- 修复: SMTP 连接添加 30 秒超时配置,防止 worker 挂起
- 修复: 邮件模板变量添加 HTML 转义,防止 XSS
- 修复: 验证状态清除改为 db.commit 后执行,避免竞态条件
- 优化: RegisterDialog 重写验证码输入组件,提升用户体验
- 优化: Input 组件支持 disableAutofill 属性
2026-01-01 02:10:19 +08:00
RWDai
6bd8cdb9cf feat: 实现邮箱验证注册功能
添加完整的邮箱验证注册系统,包括验证码发送、验证和限流控制:
  - 新增邮箱验证服务模块(email_sender, email_template, email_verification)
  - 更新认证API支持邮箱验证注册流程
  - 添加注册对话框和验证码输入组件
  - 完善IP限流器支持邮箱验证场景
  - 修复前端类型定义问题,升级esbuild依赖

  🤖 Generated with [Claude Code](https://claude.com/claude-code)

  Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
2025-12-30 17:15:48 +08:00
fawney19
d0ce798881 fix: TTL=0时启用Key随机轮换模式
- 当所有Key的cache_ttl_minutes都为0时,使用随机排序代替确定性哈希
- 将hashlib和random的import移到文件顶部
- 简化单Key场景的处理逻辑

Closes #57
2025-12-28 19:07:25 +08:00
fawney19
71bc2e6aab fix: 增加参数校验防止除零错误 2025-12-25 22:44:17 +08:00
fawney19
afb329934a fix: 修复端点健康统计时间分段计算的除零错误 2025-12-25 19:54:16 +08:00
fawney19
9dad194130 fix: 修复 API Key 访问限制字段无法清除的问题
- 统一前端创建和更新 API Key 时的空数组处理逻辑
- 后端创建和更新接口都支持空数组转 NULL(表示不限制)
- 开启自动刷新时立即刷新一次数据
2025-12-24 22:35:30 +08:00
fawney19
03ad16ea8a fix: 修复迁移脚本在全新安装时报错及改进统计回填逻辑
迁移脚本修复:
- 移除 AUTOCOMMIT 模式,改为在同一事务中创建索引
- 分别检查每个索引是否存在,只创建缺失的索引
- 修复全新安装时 AUTOCOMMIT 连接看不到未提交表的问题 (#46)

统计回填改进:
- 分别检查 StatsDaily 和 StatsDailyModel 的缺失日期
- 只回填实际缺失的数据而非连续区间
- 添加失败统计计数和 rollback 错误日志
2025-12-24 21:50:05 +08:00
fawney19
1d5c378343 feat: add TTFB timeout detection and improve stream handling
- Add stream first byte timeout (TTFB) detection to trigger failover
  when provider responds too slowly (configurable via STREAM_FIRST_BYTE_TIMEOUT)
- Add rate limit fail-open/fail-close strategy configuration
- Improve exception handling in stream prefetch with proper error classification
- Refactor UsageService with shared _prepare_usage_record method
- Add batch deletion for old usage records to avoid long transaction locks
- Update CLI adapters to use proper User-Agent headers for each CLI client
- Add composite indexes migration for usage table query optimization
- Fix streaming status display in frontend to show TTFB during streaming
- Remove sensitive JWT secret logging in auth service
2025-12-22 23:44:42 +08:00
fawney19
4e1aed9976 feat: add daily model statistics aggregation with stats_daily_model table 2025-12-20 02:39:10 +08:00
fawney19
af476ff21e feat: enhance error logging and upstream response tracking for provider failures 2025-12-19 15:29:48 +08:00
fawney19
3bbc1c6b66 feat: add provider compatibility error detection for intelligent failover
- Introduce ProviderCompatibilityException for unsupported parameter/feature errors
- Add COMPATIBILITY_ERROR_PATTERNS to detect provider-specific limitations
- Implement _is_compatibility_error() method in ErrorClassifier
- Prioritize compatibility error checking before client error validation
- Remove 'max_tokens' from CLIENT_ERROR_PATTERNS as it can indicate compatibility issues
- Enable automatic failover when provider doesn't support requested features
- Improve error classification accuracy with pattern matching for common compatibility issues
2025-12-19 13:28:26 +08:00
fawney19
c69a0a8506 refactor: remove stream smoothing config from system settings and improve base image caching
- Remove stream_smoothing configuration from SystemConfigService (moved to handler default)
- Remove stream smoothing UI controls from admin settings page
- Add AdminClearSingleAffinityAdapter for targeted cache invalidation
- Add clearSingleAffinity API endpoint to clear specific affinity cache entries
- Include global_model_id in affinity list response for UI deletion support
- Improve CI/CD workflow with hash-based base image change detection
- Add hash label to base image for reliable cache invalidation detection
- Use remote image inspection to determine if base image rebuild is needed
- Include Dockerfile.base in hash calculation for proper dependency tracking
2025-12-19 13:09:56 +08:00
fawney19
97425ac68f refactor: make stream smoothing parameters configurable and add models cache invalidation
- Move stream smoothing parameters (chunk_size, delay_ms) to database config
- Remove hardcoded stream smoothing constants from StreamProcessor
- Simplify dynamic delay calculation by using config values directly
- Add invalidate_models_list_cache() function to clear /v1/models endpoint cache
- Call cache invalidation on model create, update, delete, and bulk operations
- Update admin UI to allow runtime configuration of smoothing parameters
- Improve model listing freshness when models are modified
2025-12-19 11:03:46 +08:00
fawney19
070121717d refactor: consolidate stream smoothing into StreamProcessor with intelligent timing
- Move StreamSmoother functionality directly into StreamProcessor for better integration
- Create ContentExtractor strategy pattern for format-agnostic content extraction
- Implement intelligent dynamic delay calculation based on text length
- Support three text length tiers: short (char-by-char), medium (chunked), long (chunked)
- Remove manual chunk_size and delay_ms configuration - now auto-calculated
- Simplify admin UI to single toggle switch with auto timing adjustment
- Extract format detection logic to reusable content_extractors module
- Improve code maintainability with cleaner architecture
2025-12-19 09:46:22 +08:00
fawney19
85fafeacb8 feat: add stream smoothing feature for improved user experience
- Implement StreamSmoother class to split large content chunks into smaller pieces with delay
- Support OpenAI, Claude, and Gemini API response formats for smooth playback
- Add stream smoothing configuration to system settings (enable, chunk size, delay)
- Create streamlined API for stream smoothing with StreamSmoothingConfig dataclass
- Add admin UI controls for configuring stream smoothing parameters
- Use batch configuration loading to minimize database queries
- Enable typing effect simulation for better user experience in streaming responses
2025-12-19 03:15:19 +08:00
fawney19
5f0c1fb347 refactor: remove unused response normalizer module
- Delete unused ResponseNormalizer class and its initialization logic
- Remove response_normalizer and enable_response_normalization parameters from handlers
- Simplify chat adapter base initialization by removing normalizer setup
- Clean up unused imports in handler modules
2025-12-19 01:20:30 +08:00
fawney19
7b932d7afb refactor: optimize middleware with pure ASGI implementation and enhance security measures
- Replace BaseHTTPMiddleware with pure ASGI implementation in plugin middleware for better streaming response handling
- Add trusted proxy count configuration for client IP extraction in reverse proxy environments
- Implement audit log cleanup scheduler with configurable retention period
- Replace plaintext token logging with SHA256 hash fingerprints for security
- Fix database session lifecycle management in middleware
- Improve request tracing and error logging throughout the system
- Add comprehensive tests for pipeline architecture
2025-12-18 19:07:20 +08:00
fawney19
21587449c8 fix: improve error classification and logging system
- Enhance error classifier to properly handle API key failures with fallback support
- Add error reason/code parsing for better AWS and multi-provider compatibility
- Improve error message structure detection for non-standard formats
- Refactor file logging with size-based rotation (100MB) instead of daily
- Optimize production logging by disabling backtrace and diagnose
- Clean up model validation and remove redundant configurations
2025-12-18 10:57:31 +08:00
fawney19
b2a857c164 refactor: consolidate transaction management and remove legacy modules
- Remove unused context.py module (replaced by request.state)
- Remove provider_cache.py (no longer needed)
- Unify environment loading in config/settings.py instead of __init__.py
- Add deprecation warning for get_async_db() (consolidating on sync Session)
- Enhance database.py documentation with comprehensive transaction strategy
- Simplify audit logging to reuse request-level Session (no separate connections)
- Extract UsageService._build_usage_params() helper to reduce code duplication
- Update model and user cache implementations with refined transaction handling
- Remove unnecessary sessionmaker from pipeline
- Clean up audit service exception handling
2025-12-18 01:59:40 +08:00
fawney19
4d1d863916 refactor: improve authentication and user data handling
- Replace user cache queries with direct database queries to ensure data consistency
- Fix token_type parameter in verify_token calls (access token verification)
- Fix role-based permission check using dictionary ranking instead of string comparison
- Fix logout operation to use correct JWT claim name (user_id instead of sub)
- Simplify user authentication flow by removing unnecessary cache layer
- Optimize session initialization in main.py using create_session helper
- Remove unused imports and exception variables
2025-12-18 01:09:22 +08:00
fawney19
b579420690 refactor: optimize database session lifecycle and middleware architecture
- Improve database pool capacity logging with detailed configuration parameters
- Optimize database session dependency injection with middleware-managed lifecycle
- Simplify plugin middleware by delegating session creation to FastAPI dependencies
- Fix import path in auth routes (relative to absolute)
- Add safety checks for database session management across middleware exception handlers
- Ensure session cleanup only when not managed by middleware (avoid premature cleanup)
2025-12-18 00:35:46 +08:00
fawney19
9d5c84f9d3 refactor: add scheduling mode support and optimize system settings UI
- Add fixed_order and cache_affinity scheduling modes to CacheAwareScheduler
- Only apply cache affinity in cache_affinity mode; use fixed order otherwise
- Simplify Dialog components with title/description props
- Remove unnecessary button shadows in SystemSettings
- Optimize import dialog UI structure
- Update ModelAliasesTab shadow styling
- Fix fallback orchestrator type hints
- Add scheduling_mode configuration in system config
2025-12-17 19:15:08 +08:00
fawney19
bd11ebdbd5 fix: 修复个人设置页面深色模式切换后刷新失效的问题
- 前端使用 useDarkMode composable 统一主题切换逻辑
- 后端支持 system 主题值(之前只支持 auto)
- 主题以本地 localStorage 为准,避免刷新时被服务端旧值覆盖

Fixes #22
2025-12-17 18:02:19 +08:00
fawney19
1dac4cb156 refactor: optimize provider query and stats aggregation logic 2025-12-17 16:41:10 +08:00