Update backend/app/openrouter_client.py

backend/app/openrouter_client.py

@@ -55,7 +55,9 @@ def _pdf_to_images(pdf_bytes: bytes) -> List[bytes]:
 def _image_bytes_to_base64(image_bytes: bytes) -> str:
     """Convert image bytes to base64 data URL (JPEG format)."""
     b64 = base64.b64encode(image_bytes).decode("utf-8")
-    return f"data:image/jpeg;base64,{b64}"
+    data_url = f"data:image/jpeg;base64,{b64}"
+    print(f"[DEBUG] Base64 encoded image: {len(image_bytes)} bytes -> {len(data_url)} chars")
+    return data_url
 
 
 def _file_to_image_blocks(file_bytes: bytes, content_type: str) -> List[Dict[str, Any]]:
@@ -251,10 +253,38 @@ async def extract_fields_from_document(
         "X-Title": "Document Capture Demo",
     }
 
-    async with httpx.AsyncClient(timeout=180) as client:  # Increased timeout for long responses
-        resp = await client.post(OPENROUTER_BASE_URL, headers=headers, json=payload)
-        resp.raise_for_status()
-        data = resp.json()
+    # Calculate payload size
+    import sys
+    payload_str = json.dumps(payload)
+    payload_size_mb = len(payload_str.encode('utf-8')) / 1024 / 1024
+    
+    print(f"[INFO] Sending request to OpenRouter API...")
+    print(f"[INFO] Payload size: {payload_size_mb:.2f} MB, Images: {len(image_blocks)} blocks")
+    print(f"[INFO] Model: {MODEL_NAME}")
+    
+    if payload_size_mb > 10:
+        print(f"[WARNING] Payload is very large ({payload_size_mb:.2f} MB). This may cause slow responses or timeouts.")
+    
+    try:
+        # Use a longer timeout for large documents - 10 minutes
+        timeout = httpx.Timeout(600.0, connect=30.0)  # 10 min total, 30s connect
+        async with httpx.AsyncClient(timeout=timeout) as client:
+            print(f"[INFO] Making POST request to {OPENROUTER_BASE_URL}...")
+            print(f"[INFO] Timeout set to 10 minutes for large document processing...")
+            resp = await client.post(OPENROUTER_BASE_URL, headers=headers, json=payload)
+            print(f"[INFO] Received response: Status {resp.status_code}")
+            resp.raise_for_status()
+            data = resp.json()
+            print(f"[INFO] Response parsed successfully")
+    except httpx.TimeoutException:
+        print(f"[ERROR] Request to OpenRouter timed out after 5 minutes")
+        raise RuntimeError("Request to OpenRouter API timed out. The document may be too large or the API is slow. Please try again or use a smaller document.")
+    except httpx.HTTPStatusError as e:
+        print(f"[ERROR] HTTP error from OpenRouter: {e.response.status_code} - {e.response.text[:500]}")
+        raise RuntimeError(f"OpenRouter API error: {e.response.status_code} - {str(e)}")
+    except Exception as e:
+        print(f"[ERROR] Unexpected error calling OpenRouter: {type(e).__name__}: {str(e)}")
+        raise RuntimeError(f"Failed to call OpenRouter API: {str(e)}")
 
     # OpenRouter returns choices[0].message.content
     if "choices" not in data or len(data["choices"]) == 0: