refactor: remove adaptive rate limiting feature

Remove adaptive rate limiting since GitHub Copilot API does not provide
rate limit headers. The API only returns x-quota-snapshot-* headers which
track quota usage, not rate limits, and overage is permitted freely.

Removed:
- src/lib/rate-limit-parser.ts
- tests/rate-limit-parser.test.ts
- onHeaders callback from createChatCompletions
- Rate limit header parsing logic from handlers

The opt-in request queue remains functional for users who want to set
a fixed rate limit via --rate-limit flag.

Signed-off-by: leocavalcante <[email protected]>

Author: leocavalcante

src/lib/rate-limit-parser.ts

@@ -5,11 +5,6 @@ import { streamSSE, type SSEMessage } from "hono/streaming"
 
 import { awaitApproval } from "~/lib/approval"
 import { executeWithRateLimit } from "~/lib/rate-limit"
-import {
-  calculateOptimalDelay,
-  logRateLimitInfo,
-  parseRateLimitHeaders,
-} from "~/lib/rate-limit-parser"
 import { state } from "~/lib/state"
 import { getTokenCount } from "~/lib/tokenizer"
 import { isNullish } from "~/lib/utils"
@@ -51,16 +46,7 @@ export async function handleCompletion(c: Context) {
       consola.debug("Set max_tokens to:", JSON.stringify(payload.max_tokens))
     }
 
-    const response = await createChatCompletions(payload, (headers) => {
-      // Parse rate limit headers and update queue if applicable
-      const rateLimitInfo = parseRateLimitHeaders(headers)
-      logRateLimitInfo(rateLimitInfo)
-
-      const optimalDelay = calculateOptimalDelay(rateLimitInfo)
-      if (optimalDelay !== undefined) {
-        state.requestQueue.updateRateLimit(optimalDelay)
-      }
-    })
+    const response = await createChatCompletions(payload)
 
     if (isNonStreaming(response)) {
       consola.debug("Non-streaming response:", JSON.stringify(response))

src/routes/chat-completions/handler.ts

@@ -5,11 +5,6 @@ import { streamSSE } from "hono/streaming"
 
 import { awaitApproval } from "~/lib/approval"
 import { executeWithRateLimit } from "~/lib/rate-limit"
-import {
-  calculateOptimalDelay,
-  logRateLimitInfo,
-  parseRateLimitHeaders,
-} from "~/lib/rate-limit-parser"
 import { state } from "~/lib/state"
 import {
   createChatCompletions,
@@ -45,16 +40,7 @@ export async function handleCompletion(c: Context) {
       await awaitApproval()
     }
 
-    const response = await createChatCompletions(openAIPayload, (headers) => {
-      // Parse rate limit headers and update queue if applicable
-      const rateLimitInfo = parseRateLimitHeaders(headers)
-      logRateLimitInfo(rateLimitInfo)
-
-      const optimalDelay = calculateOptimalDelay(rateLimitInfo)
-      if (optimalDelay !== undefined) {
-        state.requestQueue.updateRateLimit(optimalDelay)
-      }
-    })
+    const response = await createChatCompletions(openAIPayload)
 
     if (isNonStreaming(response)) {
       consola.debug(

src/routes/messages/handler.ts

@@ -7,7 +7,6 @@ import { state } from "~/lib/state"
 
 export const createChatCompletions = async (
   payload: ChatCompletionsPayload,
-  onHeaders?: (headers: Headers) => void,
 ) => {
   if (!state.copilotToken) throw new Error("Copilot token not found")
 
@@ -40,11 +39,6 @@ export const createChatCompletions = async (
     throw new HTTPError("Failed to create chat completions", response)
   }
 
-  // Call the headers callback if provided
-  if (onHeaders) {
-    onHeaders(response.headers)
-  }
-
   if (payload.stream) {
     return events(response)
   }