fix(api): echo X-OmniRoute-Compression response header (#6422) (#6441)

echo X-OmniRoute-Compression header on completions routes (#6422, 6/6). Reconciled with #6429 body.model echo. Integrated into release/v3.8.46.

src/app/api/v1/chat/completions/route.ts

@@ -8,6 +8,10 @@ import { acceptHeaderForcesStream } from "@omniroute/open-sse/utils/aiSdkCompat.
 import { withEarlyStreamKeepalive } from "@omniroute/open-sse/utils/earlyStreamKeepalive";
 import { resolveKeepaliveThreshold } from "@omniroute/open-sse/utils/keepaliveThreshold";
 import { checkChatAdmission } from "@/shared/middleware/chatBodyAdmission";
+import {
+  readCompressionRequestHeader,
+  withCompressionHeaderEcho,
+} from "@/shared/utils/compressionHeaderEcho";
 
 let initPromise = null;
 
@@ -99,14 +103,26 @@ export async function POST(request) {
     parsedBodyIsRecord && acceptHeaderForcesStream(acceptHeader, parsedBody.stream);
   const wantsStreaming = (parsedBodyIsRecord && parsedBody.stream === true) || acceptForcesStream;
 
+  // #6422 — capture the compression request header once so we can echo it back
+  // on the response when internal early-returns (idempotency cache, some combo
+  // paths) drop the meta the docs promise.
+  const compressionRequestHeader = readCompressionRequestHeader(request);
+
   if (wantsStreaming) {
     const reqId = generateRequestId();
-    return await withEarlyStreamKeepalive(handleChat(request, null, parsedBody, reqId), {
-      signal: request.signal,
-      thresholdMs: resolveKeepaliveThreshold(parsedBody?.model),
-      extraHeaders: { "X-Correlation-Id": reqId },
-    });
+    const streamedResponse = await withEarlyStreamKeepalive(
+      handleChat(request, null, parsedBody, reqId),
+      {
+        signal: request.signal,
+        thresholdMs: resolveKeepaliveThreshold(parsedBody?.model),
+        extraHeaders: { "X-Correlation-Id": reqId },
+      }
+    );
+    return withCompressionHeaderEcho(streamedResponse, compressionRequestHeader);
   }
 
-  return await handleChat(request, null, parsedBody);
+  return withCompressionHeaderEcho(
+    await handleChat(request, null, parsedBody),
+    compressionRequestHeader
+  );
 }

src/app/api/v1/completions/route.ts

@@ -3,6 +3,10 @@ import { buildClientRawRequest, handleChat } from "@/sse/handlers/chat";
 import { initTranslators } from "@omniroute/open-sse/translator/index.ts";
 import { createInjectionGuard } from "@/middleware/promptInjectionGuard";
 import { asTextCompletionResponse } from "./textCompletionTransform.ts";
+import {
+  readCompressionRequestHeader,
+  withCompressionHeaderEcho,
+} from "@/shared/utils/compressionHeaderEcho";
 
 let initPromise = null;
 const injectionGuard = createInjectionGuard();
@@ -40,6 +44,10 @@ export async function OPTIONS() {
 export async function POST(request: Request) {
   await ensureInitialized();
 
+  // #6422 — capture the compression request header once so we can echo it back
+  // on the response when internal early-returns drop the meta the docs promise.
+  const compressionRequestHeader = readCompressionRequestHeader(request);
+
   // Prompt injection guard
   try {
     const cloned = request.clone();
@@ -78,10 +86,14 @@ export async function POST(request: Request) {
         // #3571 — translate the chat-pipeline response back to the legacy
         // text-completion shape so OpenAI Completion clients (e.g. TabbyML) work.
         // Thread `body.model` so response `body.model` echoes the caller's
-        // requested identifier, matching the `x-omniroute-model` header.
-        return await asTextCompletionResponse(
-          await handleChat(newRequest, buildClientRawRequest(request, body)),
-          typeof body.model === "string" ? body.model : undefined
+        // requested identifier, matching the `x-omniroute-model` header, and
+        // echo the compression header on the way out.
+        return withCompressionHeaderEcho(
+          await asTextCompletionResponse(
+            await handleChat(newRequest, buildClientRawRequest(request, body)),
+            typeof body.model === "string" ? body.model : undefined
+          ),
+          compressionRequestHeader
         );
       }
     }
@@ -101,5 +113,8 @@ export async function POST(request: Request) {
   } catch {
     // ignore — asTextCompletionResponse falls back to upstream body.model
   }
-  return await asTextCompletionResponse(await handleChat(request), requestedModel);
+  return withCompressionHeaderEcho(
+    await asTextCompletionResponse(await handleChat(request), requestedModel),
+    compressionRequestHeader
+  );
 }