i/Aether
1
0
Fork 0
mirror of https://github.com/fawney19/Aether.git synced 2026-09-01 17:00:21 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat: 扩展 cache creation token 细分统计与 effective_input_tokens 计费逻辑

Browse Source 
- 新增 cache_creation_ephemeral_5m/1h_input_tokens 字段，区分不同 TTL 的缓存写入 token
- 引入 effective_input_tokens（扣除 cache read 后的有效输入 token），暴露给 usage 接口
- billing 规则生成器支持 5m/1h ephemeral cache 独立定价与分级计费
- usage_mapper 增加 Claude/Anthropic 格式映射，修复 OpenAI responses 格式字段兼容性
- 迁移逻辑增强：支持 checksum 容错、applied/pending 数量日志、逐步执行信息输出
- executor 抽离 LocalExecutionRequestOutcome 类型，统一 sync/stream 路径返回语义
- provider-transport auth 层新增 complete passthrough headers 构建逻辑
- 前端 usage 类型全面补充 effective_input_tokens、cache_creation_tokens、total_input_context 字段
This commit is contained in:
fawney19
2026-04-10 17:44:55 +08:00
parent 5014e2f5fd
commit 010ab127e2
64 changed files with 4217 additions and 477 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
frontend/src/api/dashboard.ts
View File

@@ -128,6 +128,7 @@ export interface RequestDetail {
}
provider: string
api_format?: string
endpoint_api_format?: string
model: string
target_model?: string | null // 映射后的目标模型名
tokens: {
@@ -142,6 +143,7 @@ export interface RequestDetail {
}
// Additional token fields
input_tokens?: number
effective_input_tokens?: number
output_tokens?: number
total_tokens?: number
cache_creation_input_tokens?: number

12
frontend/src/api/me.ts
View File

@@ -51,6 +51,7 @@ export interface UsageRecordDetail {
provider: string
model: string
input_tokens: number
effective_input_tokens?: number
output_tokens: number
total_tokens: number
cost: number // 官方费率
@@ -80,9 +81,12 @@ export interface ModelSummary {
model: string
requests: number
input_tokens: number
effective_input_tokens?: number
output_tokens: number
total_tokens: number
cache_read_tokens?: number
cache_creation_tokens?: number
total_input_context?: number
cache_hit_rate?: number
total_cost_usd: number
actual_total_cost_usd?: number // 倍率消耗（仅管理员可见）
@@ -92,8 +96,12 @@ export interface ModelSummary {
export interface ProviderSummary {
provider: string
requests: number
effective_input_tokens?: number
total_tokens: number
output_tokens?: number
cache_read_tokens?: number
cache_creation_tokens?: number
total_input_context?: number
cache_hit_rate?: number
total_cost_usd: number
success_rate: number | null
@@ -104,8 +112,12 @@ export interface ProviderSummary {
export interface ApiFormatSummary {
api_format: string
request_count: number
effective_input_tokens?: number
total_tokens: number
output_tokens?: number
cache_read_tokens: number
cache_creation_tokens?: number
total_input_context?: number
cache_hit_rate: number
total_cost_usd: number
avg_response_time_ms: number

14
frontend/src/api/usage.ts
View File

@@ -10,6 +10,7 @@ export interface UsageRecord {
provider_name?: string
model: string
input_tokens: number
effective_input_tokens?: number
output_tokens: number
cache_creation_input_tokens?: number
cache_read_input_tokens?: number
@@ -38,6 +39,10 @@ export interface UsageByModel {
model: string
request_count: number
total_tokens: number
effective_input_tokens?: number
total_input_context?: number
output_tokens?: number
cache_creation_tokens?: number
total_cost: number
avg_response_time?: number
cache_read_tokens?: number
@@ -58,6 +63,10 @@ export interface UsageByProvider {
provider: string
request_count: number
total_tokens: number
effective_input_tokens?: number
total_input_context?: number
output_tokens?: number
cache_creation_tokens?: number
total_cost: number
actual_cost: number
avg_response_time_ms: number
@@ -71,6 +80,10 @@ export interface UsageByApiFormat {
api_format: string
request_count: number
total_tokens: number
effective_input_tokens?: number
total_input_context?: number
output_tokens?: number
cache_creation_tokens?: number
total_cost: number
actual_cost: number
avg_response_time_ms: number
@@ -276,6 +289,7 @@ export const usageApi = {
id: string
status: 'pending' | 'streaming' | 'completed' | 'failed' | 'cancelled'
input_tokens: number
effective_input_tokens?: number | null
output_tokens: number
cache_creation_input_tokens?: number | null
cache_read_input_tokens?: number | null

15
frontend/src/features/usage/components/RequestDetailDrawer.vue
View File

@@ -194,7 +194,7 @@
<!-- 阶梯标题 -->
<div class="text-xs text-muted-foreground flex items-center gap-2 flex-wrap">
<span class="font-medium text-foreground">Token 计费</span>
<span class="text-muted-foreground/60">(输入 {{ formatNumber(detail.tokens?.input || detail.input_tokens || 0) }} + 缓存创建 {{ cacheCreationSummaryText }} + 缓存读取 {{ formatNumber(detail.cache_read_input_tokens || 0) }})</span>
<span class="text-muted-foreground/60">(输入 {{ formatNumber(displayInputTokens) }} + 缓存创建 {{ cacheCreationSummaryText }} + 缓存读取 {{ formatNumber(detail.cache_read_input_tokens || 0) }})</span>
<Badge
v-if="displayTiers.length > 1"
variant="outline"
@@ -260,7 +260,7 @@
<div class="flex items-center">
<div class="flex items-center flex-1">
<span class="text-xs text-muted-foreground w-[56px]">输入</span>
<span class="text-sm font-semibold font-mono flex-1 text-center">{{ detail.tokens?.input || detail.input_tokens || 0 }}</span>
<span class="text-sm font-semibold font-mono flex-1 text-center">{{ displayInputTokens }}</span>
<span class="text-xs font-mono">${{ (detail.cost?.input || detail.input_cost || 0).toFixed(6) }}</span>
</div>
<Separator
@@ -701,6 +701,7 @@ import { dashboardApi, type RequestDetail } from '@/api/dashboard'
import { formatApiFormat } from '@/api/endpoints/types/api-format'
import { formatShortRequestId } from '@/utils/format'
import { log } from '@/utils/logger'
import { getEffectiveInputTokens } from '../token-normalization'
// 子组件
import RequestHeadersContent from './RequestDetailDrawer/RequestHeadersContent.vue'
@@ -774,6 +775,16 @@ let timelineMountTimer: ReturnType<typeof setTimeout> | null = null
const fullRequestId = computed(() => detail.value?.request_id || detail.value?.id || '-')
const displayRequestId = computed(() => formatShortRequestId(fullRequestId.value))
const displayInputTokens = computed(() => {
if (!detail.value) return 0
return getEffectiveInputTokens({
effective_input_tokens: detail.value.effective_input_tokens,
input_tokens: detail.value.input_tokens ?? detail.value.tokens?.input,
cache_read_input_tokens: detail.value.cache_read_input_tokens,
api_format: detail.value.api_format,
endpoint_api_format: detail.value.endpoint_api_format,
})
})
// 监听标签页切换
watch(activeTab, (newTab) => {

2
frontend/src/features/usage/components/UsageApiFormatTable.vue
View File

@@ -53,7 +53,7 @@
</TableCell>
<TableCell class="text-right py-2 px-2">
<div class="flex flex-col items-end text-xs gap-0.5 whitespace-nowrap">
<span>{{ formatTokens(item.total_input_context || 0) }} / {{ formatTokens(item.output_tokens || 0) }}</span>
<span>{{ formatTokens(item.effective_input_tokens ?? item.total_input_context ?? 0) }} / {{ formatTokens(item.output_tokens || 0) }}</span>
<span class="text-muted-foreground">{{ formatTokens((item.cache_read_tokens || 0) + (item.cache_creation_tokens || 0)) }}</span>
</div>
</TableCell>

2
frontend/src/features/usage/components/UsageModelTable.vue
View File

@@ -53,7 +53,7 @@
</TableCell>
<TableCell class="text-right py-2 px-2">
<div class="flex flex-col items-end text-xs gap-0.5 whitespace-nowrap">
<span>{{ formatTokens(model.total_input_context || 0) }} / {{ formatTokens(model.output_tokens || 0) }}</span>
<span>{{ formatTokens(model.effective_input_tokens ?? model.total_input_context ?? 0) }} / {{ formatTokens(model.output_tokens || 0) }}</span>
<span class="text-muted-foreground">{{ formatTokens(model.cache_read_tokens || 0) }}</span>
</div>
</TableCell>

2
frontend/src/features/usage/components/UsageProviderTable.vue
View File

@@ -56,7 +56,7 @@
</TableCell>
<TableCell class="text-right py-2 px-2">
<div class="flex flex-col items-end text-xs gap-0.5 whitespace-nowrap">
<span>{{ formatTokens(provider.totalInputContext || 0) }} / {{ formatTokens(provider.outputTokens || 0) }}</span>
<span>{{ formatTokens(provider.effectiveInputTokens ?? provider.totalInputContext ?? 0) }} / {{ formatTokens(provider.outputTokens || 0) }}</span>
<span class="text-muted-foreground">{{ formatTokens((provider.cacheReadTokens || 0) + (provider.cacheCreationTokens || 0)) }}</span>
</div>
</TableCell>

9
frontend/src/features/usage/components/UsageRecordsTable.vue
View File

@@ -273,7 +273,7 @@
>-</span>
<span class="text-muted-foreground/50">|</span>
<!-- Tokens -->
<span>{{ formatTokens(record.input_tokens || 0) }}/{{ formatTokens(record.output_tokens || 0) }}</span>
<span>{{ formatTokens(getRecordEffectiveInputTokens(record)) }}/{{ formatTokens(record.output_tokens || 0) }}</span>
</div>
</div>
</div>
@@ -552,7 +552,7 @@
<TableCell class="text-right py-4 w-[140px]">
<div class="flex flex-col items-end text-xs gap-0.5">
<div class="flex items-center gap-1">
<span>{{ formatTokens(record.input_tokens || 0) }}</span>
<span>{{ formatTokens(getRecordEffectiveInputTokens(record)) }}</span>
<span class="text-muted-foreground">/</span>
<span>{{ formatTokens(record.output_tokens || 0) }}</span>
</div>
@@ -677,6 +677,7 @@ import {
import { RefreshCcw, Search } from 'lucide-vue-next'
import { formatTokens, formatCurrency } from '@/utils/format'
import { formatDateTime } from '../composables'
import { getEffectiveInputTokens } from '../token-normalization'
import { useRowClick } from '@/composables/useRowClick'
import { formatApiFormat } from '@/api/endpoints/types/api-format'
import type { DateRangeParams, UsageRecord } from '../types'
@@ -777,6 +778,10 @@ function handleRowClick(event: MouseEvent, id: string) {
emit('showDetail', id)
}
function getRecordEffectiveInputTokens(record: UsageRecord): number {
return getEffectiveInputTokens(record)
}
// useDebounceFn 自动处理清理，无需 onUnmounted
// 判断是否应该显示格式转换信息

7
frontend/src/features/usage/composables/useUsageData.ts
View File

@@ -110,6 +110,7 @@ export function useUsageData(options: UseUsageDataOptions) {
model: item.model,
request_count: item.request_count || 0,
total_tokens: item.total_tokens || 0,
effective_input_tokens: typeof raw.effective_input_tokens === 'number' ? raw.effective_input_tokens : 0,
total_input_context: typeof raw.total_input_context === 'number' ? raw.total_input_context : 0,
output_tokens: typeof raw.output_tokens === 'number' ? raw.output_tokens : 0,
cache_read_tokens: typeof raw.cache_read_tokens === 'number' ? raw.cache_read_tokens : 0,
@@ -129,6 +130,7 @@ export function useUsageData(options: UseUsageDataOptions) {
provider: item.provider,
requests: item.request_count,
totalTokens: item.total_tokens || 0,
effectiveInputTokens: item.effective_input_tokens || 0,
totalInputContext: item.total_input_context || 0,
outputTokens: item.output_tokens || 0,
cacheReadTokens: item.cache_read_tokens || 0,
@@ -151,6 +153,7 @@ export function useUsageData(options: UseUsageDataOptions) {
api_format: item.api_format,
request_count: item.request_count || 0,
total_tokens: item.total_tokens || 0,
effective_input_tokens: item.effective_input_tokens || 0,
total_input_context: item.total_input_context || 0,
output_tokens: item.output_tokens || 0,
cache_read_tokens: item.cache_read_tokens || 0,
@@ -188,6 +191,7 @@ export function useUsageData(options: UseUsageDataOptions) {
model: item.model,
request_count: item.requests || 0,
total_tokens: item.total_tokens || 0,
effective_input_tokens: item.effective_input_tokens || 0,
total_input_context: item.total_input_context || 0,
output_tokens: item.output_tokens || 0,
cache_read_tokens: item.cache_read_tokens || 0,
@@ -201,6 +205,7 @@ export function useUsageData(options: UseUsageDataOptions) {
provider: item.provider,
requests: item.requests || 0,
totalTokens: item.total_tokens || 0,
effectiveInputTokens: item.effective_input_tokens || 0,
totalInputContext: item.total_input_context || 0,
outputTokens: item.output_tokens || 0,
cacheReadTokens: item.cache_read_tokens || 0,
@@ -234,6 +239,7 @@ export function useUsageData(options: UseUsageDataOptions) {
api_format: item.api_format,
request_count: item.request_count || 0,
total_tokens: item.total_tokens || 0,
effective_input_tokens: item.effective_input_tokens || 0,
total_input_context: item.total_input_context || 0,
output_tokens: item.output_tokens || 0,
cache_read_tokens: item.cache_read_tokens || 0,
@@ -384,6 +390,7 @@ export function useUsageData(options: UseUsageDataOptions) {
status: mergedStatus,
provider: protectProvider ? existing.provider : (record.provider || existing.provider),
input_tokens: existing.input_tokens || record.input_tokens,
effective_input_tokens: existing.effective_input_tokens ?? record.effective_input_tokens,
output_tokens: existing.output_tokens || record.output_tokens,
cache_creation_input_tokens: existing.cache_creation_input_tokens ?? record.cache_creation_input_tokens,
cache_read_input_tokens: existing.cache_read_input_tokens ?? record.cache_read_input_tokens,

40
frontend/src/features/usage/token-normalization.ts Normal file
View File

@@ -0,0 +1,40 @@
type UsageTokenLike = {
effective_input_tokens?: number | null
input_tokens?: number | null
cache_read_input_tokens?: number | null
api_format?: string | null
endpoint_api_format?: string | null
}
function toNonNegativeNumber(value: number | null | undefined): number {
return typeof value === 'number' && Number.isFinite(value) ? Math.max(value, 0) : 0
}
function apiFamily(apiFormat: string | null | undefined): string {
return String(apiFormat || '')
.split(':', 1)[0]
.trim()
.toLowerCase()
}
export function getEffectiveInputTokens(usage: UsageTokenLike): number {
const explicit = toNonNegativeNumber(usage.effective_input_tokens)
if (explicit > 0) {
return explicit
}
const inputTokens = toNonNegativeNumber(usage.input_tokens)
const cacheReadTokens = toNonNegativeNumber(usage.cache_read_input_tokens)
if (inputTokens === 0 || cacheReadTokens === 0) {
return inputTokens
}
switch (apiFamily(usage.endpoint_api_format || usage.api_format)) {
case 'openai':
case 'gemini':
case 'google':
return Math.max(inputTokens - cacheReadTokens, 0)
default:
return inputTokens
}
}

4
frontend/src/features/usage/types.ts
View File

@@ -22,6 +22,7 @@ export interface ModelStatsItem {
model: string
request_count: number
total_tokens: number
effective_input_tokens?: number
total_input_context?: number
output_tokens?: number
cache_read_tokens?: number
@@ -41,6 +42,7 @@ export interface ProviderStatsItem {
provider: string
requests: number
totalTokens: number
effectiveInputTokens?: number
totalInputContext?: number
outputTokens?: number
cacheReadTokens?: number
@@ -57,6 +59,7 @@ export interface ApiFormatStatsItem {
api_format: string
request_count: number
total_tokens: number
effective_input_tokens?: number
total_input_context?: number
output_tokens?: number
cache_read_tokens?: number
@@ -92,6 +95,7 @@ export interface UsageRecord {
endpoint_api_format?: string // 端点原生格式
has_format_conversion?: boolean // 是否发生了格式转换
input_tokens: number
effective_input_tokens?: number
output_tokens: number
cache_creation_input_tokens?: number
cache_read_input_tokens?: number

1
frontend/src/views/shared/Usage.vue
View File

@@ -391,6 +391,7 @@ async function pollActiveRequests() {
if (shouldApply) {
// 进行中状态也需要持续更新（provider/key/TTFB 可能在 streaming 后才落库）
record.input_tokens = update.input_tokens
record.effective_input_tokens = update.effective_input_tokens ?? record.effective_input_tokens
record.output_tokens = update.output_tokens
record.cache_creation_input_tokens = update.cache_creation_input_tokens ?? undefined
record.cache_read_input_tokens = update.cache_read_input_tokens ?? undefined