feat: 在日志中显示 premium requests 剩余量

- 在请求完成时显示 [N left] 格式的 premium 配额剩余
- 实时调用 usage API 获取最新数据
- 支持所有 handler 路径 (streaming/non-streaming, ChatCompletions/Responses)

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

Co-Authored-By: Claude <noreply@anthropic.com>

Author: cuipengfei

.beads/issues.jsonl

@@ -1,2 +1,3 @@
 {"id":"copilot-api-0jp","title":"改进 /v1/messages 日志：IN/OUT 格式 + 实时 chunk 计数","description":"目标日志格式：\n[/v1/messages] IN  gpt-5-mini → gpt-5 (Responses)\n[/v1/messages] OUT gpt-5 (Responses) 42 chunks ✓\n\n要求：\n- IN：请求进来时打印，原始模型 → 实际模型 (API路径)\n- OUT：streaming 时用 \\r 实时更新 chunk 数，结束后加 ✓ 并换行\n- Non-streaming：打印 OUT model (API) ✓\n\n关键约束：\n- 不能消费 stream（只计数，不读取内容）\n- 不能破坏现有 streaming 行为\n- chunk 计数在循环中递增，不影响数据传输\n\n修改文件：src/routes/messages/handler.ts","status":"closed","priority":1,"issue_type":"task","created_at":"2025-11-28T23:33:01.113239475+08:00","updated_at":"2025-11-28T23:39:50.896628851+08:00","closed_at":"2025-11-28T23:39:50.896628851+08:00"}
 {"id":"copilot-api-b5q","title":"改进模型请求日志：覆盖两条路径并记录原始/实际模型","description":"当前日志只在 ChatCompletions 路径记录 openAIPayload.model。需要：\n1. 在两条路径（ChatCompletions 和 Responses API）都记录日志\n2. 记录原始请求的模型（anthropicPayload.model）\n3. 记录实际使用的模型（翻译后的模型名）\n\n相关文件：\n- src/routes/messages/handler.ts:76-78\n- src/routes/messages/non-stream-translation.ts:49 (translateModelName)\n- src/routes/messages/responses-translation.ts:64 (model 直接透传)","notes":"实施计划：\n\n1. 在 handleCompletion 函数中（路由分叉前）添加统一日志：\n   - 记录原始请求模型：anthropicPayload.model\n   - 记录使用的 API 路径：ChatCompletions 或 Responses\n\n2. ChatCompletions 路径：\n   - 在 handleWithChatCompletions 中记录 openAIPayload.model（翻译后的模型名）\n\n3. Responses API 路径：\n   - 在 handleWithResponsesApi 中添加类似日志\n   - 记录 responsesPayload.model（Responses API 不翻译模型名）\n\n4. 移除原有的单点日志：handler.ts:76\n\n5. 删除代码中的三个 TODO 注释","status":"closed","priority":2,"issue_type":"task","created_at":"2025-11-28T22:52:37.566511883+08:00","updated_at":"2025-11-28T22:56:59.002176965+08:00","closed_at":"2025-11-28T22:56:59.002176965+08:00"}
+{"id":"copilot-api-ojy","title":"显示 Premium Requests 剩余数量在日志中","description":"在请求完成日志 ↪ model chunks ✓ 后添加 [remaining/total] 显示 premium requests 剩余量","design":"## 实现方案\n\n1. 修改 OutLogOptions 接口添加 premium 字段\n2. 修改 formatOutLog 函数在 done 时显示 [remaining/total]\n3. 在所有 done:true 调用点获取 getCopilotUsage() 数据\n\n## 修改文件\n- src/routes/messages/handler.ts","status":"closed","priority":2,"issue_type":"feature","created_at":"2025-11-29T22:05:17.4622557+08:00","updated_at":"2025-11-29T22:07:47.3559476+08:00","closed_at":"2025-11-29T22:07:47.3559476+08:00"}

src/routes/messages/handler.ts

@@ -29,6 +29,7 @@ import {
   type ResponsesResult,
   type ResponseStreamEvent,
 } from "~/services/copilot/create-responses"
+import { getCopilotUsage } from "~/services/github/get-copilot-usage"
 
 import {
   type AnthropicMessagesPayload,
@@ -46,10 +47,37 @@ interface OutLogOptions {
   model: string
   chunks: number
   done: boolean
+  premium?: { remaining: number; total: number } | null
 }
 
-const formatOutLog = ({ model, chunks, done }: OutLogOptions): string =>
-  `\x1b[2K\r↪ ${model} ${chunks}${done ? " ✓" : ""}`
+const formatOutLog = ({
+  model,
+  chunks,
+  done,
+  premium,
+}: OutLogOptions): string => {
+  const base = `\x1b[2K\r↪ ${model} ${chunks}${done ? " ✓" : ""}`
+  if (done && premium) {
+    return `${base} [${premium.remaining} left]`
+  }
+  return base
+}
+
+const getPremiumInfo = async (): Promise<{
+  remaining: number
+  total: number
+} | null> => {
+  try {
+    const usage = await getCopilotUsage()
+    const pi = usage.quota_snapshots.premium_interactions
+    if (!pi.unlimited) {
+      return { remaining: pi.remaining, total: pi.entitlement }
+    }
+  } catch {
+    // Ignore errors, don't affect main flow
+  }
+  return null
+}
 
 export async function handleCompletion(c: Context) {
   await checkRateLimit(state)
@@ -102,8 +130,9 @@ const handleWithChatCompletions = async (
       "Translated Anthropic response:",
       JSON.stringify(anthropicResponse),
     )
+    const premium = await getPremiumInfo()
     process.stdout.write(
-      `${formatOutLog({ model: openAIPayload.model, chunks: 0, done: true })}\n`,
+      `${formatOutLog({ model: openAIPayload.model, chunks: 0, done: true, premium })}\n`,
     )
     return c.json(anthropicResponse)
   }
@@ -154,8 +183,9 @@ const handleWithChatCompletions = async (
       }
     } finally {
       clearInterval(pingInterval)
+      const premium = await getPremiumInfo()
       process.stdout.write(
-        `${formatOutLog({ model: openAIPayload.model, chunks: chunkCount, done: true })}\n`,
+        `${formatOutLog({ model: openAIPayload.model, chunks: chunkCount, done: true, premium })}\n`,
       )
     }
   })
@@ -245,8 +275,9 @@ const handleWithResponsesApi = async (
         }
       } finally {
         clearInterval(pingInterval)
+        const premium = await getPremiumInfo()
         process.stdout.write(
-          `${formatOutLog({ model: responsesPayload.model, chunks: chunkCount, done: true })}\n`,
+          `${formatOutLog({ model: responsesPayload.model, chunks: chunkCount, done: true, premium })}\n`,
         )
       }
     })
@@ -263,8 +294,9 @@ const handleWithResponsesApi = async (
     "Translated Anthropic response:",
     JSON.stringify(anthropicResponse),
   )
+  const premium = await getPremiumInfo()
   process.stdout.write(
-    `${formatOutLog({ model: responsesPayload.model, chunks: 0, done: true })}\n`,
+    `${formatOutLog({ model: responsesPayload.model, chunks: 0, done: true, premium })}\n`,
   )
   return c.json(anthropicResponse)
 }