perf: use shared HTTP client pool in pypi_download_file

backend/app/pypi_proxy.py

@@ -642,6 +642,8 @@ async def pypi_download_file(
     upstream: Optional[str] = None,
     db: Session = Depends(get_db),
     storage: S3Storage = Depends(get_storage),
+    http_client: HttpClientManager = Depends(get_http_client),
+    cache: CacheService = Depends(get_cache),
 ):
     """
     Download a package file, caching it in Orchard.
@@ -723,7 +725,9 @@ async def pypi_download_file(
             headers.update(_build_auth_headers(matched_source))
         auth = _get_basic_auth(matched_source) if matched_source else None
 
-        timeout = httpx.Timeout(300.0, connect=PROXY_CONNECT_TIMEOUT)  # 5 minutes for large files
+        # Use shared HTTP client from pool with longer timeout for file downloads
+        client = http_client.get_client()
+        download_timeout = httpx.Timeout(connect=30.0, read=300.0, write=300.0, pool=30.0)
 
         # Initialize extracted dependencies list
         extracted_deps = []
@@ -731,78 +735,79 @@ async def pypi_download_file(
         # Fetch the file
         logger.info(f"PyPI proxy: fetching {filename} from {upstream_url}")
 
-        async with httpx.AsyncClient(timeout=timeout, follow_redirects=False) as client:
+        response = await client.get(
+            upstream_url,
+            headers=headers,
+            auth=auth,
+            timeout=download_timeout,
+        )
+
+        # Handle redirects manually
+        redirect_count = 0
+        while response.status_code in (301, 302, 303, 307, 308) and redirect_count < 5:
+            redirect_url = response.headers.get('location')
+            if not redirect_url:
+                break
+
+            if not redirect_url.startswith('http'):
+                redirect_url = urljoin(upstream_url, redirect_url)
+
+            logger.info(f"PyPI proxy: following redirect to {redirect_url}")
+
+            # Don't send auth to different hosts
+            redirect_headers = {"User-Agent": "Orchard-PyPI-Proxy/1.0"}
+            redirect_auth = None
+            if urlparse(redirect_url).netloc == urlparse(upstream_url).netloc:
+                redirect_headers.update(headers)
+                redirect_auth = auth
+
             response = await client.get(
-                upstream_url,
-                headers=headers,
-                auth=auth,
+                redirect_url,
+                headers=redirect_headers,
+                auth=redirect_auth,
+                follow_redirects=False,
+                timeout=download_timeout,
+            )
+            redirect_count += 1
+
+        if response.status_code != 200:
+            # Parse upstream error for policy/blocking messages
+            error_detail = _parse_upstream_error(response)
+            logger.warning(f"PyPI proxy: upstream returned {response.status_code} for {filename}: {error_detail}")
+            raise HTTPException(
+                status_code=response.status_code,
+                detail=f"Upstream error: {error_detail}"
             )
 
-            # Handle redirects manually
-            redirect_count = 0
-            while response.status_code in (301, 302, 303, 307, 308) and redirect_count < 5:
-                redirect_url = response.headers.get('location')
-                if not redirect_url:
-                    break
+        content_type = response.headers.get('content-type', 'application/octet-stream')
 
-                if not redirect_url.startswith('http'):
-                    redirect_url = urljoin(upstream_url, redirect_url)
+        # Stream to temp file to avoid loading large packages into memory
+        # This keeps memory usage constant regardless of package size
+        # Using async iteration to avoid blocking the event loop
+        tmp_path = None
+        try:
+            with tempfile.NamedTemporaryFile(delete=False, suffix=f"_{filename}") as tmp_file:
+                tmp_path = tmp_file.name
+                async for chunk in response.aiter_bytes(chunk_size=65536):  # 64KB chunks
+                    tmp_file.write(chunk)
 
-                logger.info(f"PyPI proxy: following redirect to {redirect_url}")
+            # Store in S3 from temp file (computes hash and deduplicates automatically)
+            with open(tmp_path, 'rb') as f:
+                result = storage.store(f)
+            sha256 = result.sha256
+            size = result.size
+            s3_key = result.s3_key
 
-                # Don't send auth to different hosts
-                redirect_headers = {"User-Agent": "Orchard-PyPI-Proxy/1.0"}
-                redirect_auth = None
-                if urlparse(redirect_url).netloc == urlparse(upstream_url).netloc:
-                    redirect_headers.update(headers)
-                    redirect_auth = auth
+            # Extract dependencies from the temp file before cleaning up
+            extracted_deps = _extract_dependencies_from_file(tmp_path, filename)
+            if extracted_deps:
+                logger.info(f"PyPI proxy: extracted {len(extracted_deps)} dependencies from {filename}")
 
-                response = await client.get(
-                    redirect_url,
-                    headers=redirect_headers,
-                    auth=redirect_auth,
-                    follow_redirects=False,
-                )
-                redirect_count += 1
-
-            if response.status_code != 200:
-                # Parse upstream error for policy/blocking messages
-                error_detail = _parse_upstream_error(response)
-                logger.warning(f"PyPI proxy: upstream returned {response.status_code} for {filename}: {error_detail}")
-                raise HTTPException(
-                    status_code=response.status_code,
-                    detail=f"Upstream error: {error_detail}"
-                )
-
-            content_type = response.headers.get('content-type', 'application/octet-stream')
-
-            # Stream to temp file to avoid loading large packages into memory
-            # This keeps memory usage constant regardless of package size
-            # Using async iteration to avoid blocking the event loop
-            tmp_path = None
-            try:
-                with tempfile.NamedTemporaryFile(delete=False, suffix=f"_{filename}") as tmp_file:
-                    tmp_path = tmp_file.name
-                    async for chunk in response.aiter_bytes(chunk_size=65536):  # 64KB chunks
-                        tmp_file.write(chunk)
-
-                # Store in S3 from temp file (computes hash and deduplicates automatically)
-                with open(tmp_path, 'rb') as f:
-                    result = storage.store(f)
-                sha256 = result.sha256
-                size = result.size
-                s3_key = result.s3_key
-
-                # Extract dependencies from the temp file before cleaning up
-                extracted_deps = _extract_dependencies_from_file(tmp_path, filename)
-                if extracted_deps:
-                    logger.info(f"PyPI proxy: extracted {len(extracted_deps)} dependencies from {filename}")
-
-                logger.info(f"PyPI proxy: downloaded {filename}, {size} bytes, sha256={sha256[:12]}")
-            finally:
-                # Clean up temp file
-                if tmp_path and os.path.exists(tmp_path):
-                    os.unlink(tmp_path)
+            logger.info(f"PyPI proxy: downloaded {filename}, {size} bytes, sha256={sha256[:12]}")
+        finally:
+            # Clean up temp file
+            if tmp_path and os.path.exists(tmp_path):
+                os.unlink(tmp_path)
 
         # Check if artifact already exists
         existing = db.query(Artifact).filter(Artifact.id == sha256).first()