Update utilities/extract_tables.py

utilities/extract_tables.py

@@ -49,6 +49,15 @@ def extract_tables_remote(state):
     
     # Extract just the filename (not full path) to match curl format
     file_basename = os.path.basename(path)
+    file_size = os.path.getsize(path)
+
+    print(f"\n{'='*60}")
+    print(f"📊 EXTRACT TABLES API CALL")
+    print(f"{'='*60}")
+    print(f"File: {filename}")
+    print(f"Basename: {file_basename}")
+    print(f"Size: {file_size} bytes")
+    print(f"API URL: {EXTRACT_TABLES_API}")
 
     with open(path, "rb") as f:
         # IMPORTANT: Use basename for the file tuple (matches curl format)
@@ -62,12 +71,44 @@ def extract_tables_remote(state):
         }
         headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_TOKEN')}"}
         
-        # Call API and wait for response
-        resp = requests.post(EXTRACT_TABLES_API, files=files, data=data, headers=headers, timeout=120)
+        print(f"🚀 Sending request...")
+        print(f"Data params: {data}")
+        
+        # Call API and wait for response - 5 minute timeout for large files
+        try:
+            resp = requests.post(EXTRACT_TABLES_API, files=files, data=data, headers=headers, timeout=300)
+        except requests.exceptions.Timeout:
+            print(f"❌ Request timed out after 300 seconds")
+            raise RuntimeError("Table extraction API timed out after 5 minutes")
+        except requests.exceptions.RequestException as e:
+            print(f"❌ Request exception: {str(e)}")
+            raise RuntimeError(f"API request failed: {str(e)}")
+
+    print(f"\n📥 API Response:")
+    print(f"Status Code: {resp.status_code}")
 
     if resp.status_code != 200:
-        raise RuntimeError(f"Extract tables API failed: {resp.text}")
+        print(f"❌ Error: {resp.text[:500]}")
+        raise RuntimeError(f"Extract tables API failed with status {resp.status_code}: {resp.text}")
 
-    js = resp.json()
-    state["tables"] = js.get("tables", js)
-    return state
+    try:
+        response_json = resp.json()
+        print(f"Response keys: {list(response_json.keys())}")
+        
+        # IMPORTANT: API returns {"status": "completed", "result": [...tables...], "process_id": "..."}
+        # NOT {"tables": [...]}
+        tables = response_json.get("result", []) or response_json.get("tables", [])
+        
+        print(f"\n📊 Extraction Result:")
+        print(f"API Status: {response_json.get('status', 'unknown')}")
+        print(f"Process ID: {response_json.get('process_id', 'none')}")
+        print(f"Tables found: {len(tables) if isinstance(tables, list) else 0}")
+        print(f"{'='*60}\n")
+        
+        state["tables"] = tables
+        return state
+        
+    except Exception as e:
+        print(f"❌ Error parsing response: {str(e)}")
+        print(f"Raw response: {resp.text[:500]}")
+        raise RuntimeError(f"Failed to parse API response: {str(e)}")