Update utilities/extract_tables.py

utilities/extract_tables.py

@@ -46,18 +46,24 @@ def extract_tables_remote(state):
 
     if not os.path.exists(path):
         raise RuntimeError(f"File not found: {path}")
+    
+    # Extract just the filename (not full path) to match curl format
+    file_basename = os.path.basename(path)
 
     with open(path, "rb") as f:
-        files = {"file": (filename, f, "application/pdf")}
+        # IMPORTANT: Use basename for the file tuple (matches curl format)
+        files = {"file": (file_basename, f, "application/pdf")}
+        
+        # IMPORTANT: Convert page numbers to strings (matches curl -F format)
         data = {
-            "filename": filename,
-            "start_page": state.get("start_page", 1),
-            "end_page": state.get("end_page", 1),
+            "filename": file_basename,  # Just filename, not full path
+            "start_page": str(state.get("start_page", 1)),  # String, not int
+            "end_page": str(state.get("end_page", 1)),  # String, not int
         }
         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)
+        resp = requests.post(EXTRACT_TABLES_API, files=files, data=data, headers=headers, timeout=120)
 
     if resp.status_code != 200:
         raise RuntimeError(f"Extract tables API failed: {resp.text}")