Update utilities/extract_tables.py

utilities/extract_tables.py

@@ -1,25 +1,54 @@
-import os
-import requests
-
-EXTRACT_TABLES_API = "https://point9-extract-text-and-table.hf.space/api/tables"  # Replace with your space URL
-
-def extract_tables_remote(state):
-    filename = state["filename"]
-    path = state["temp_files"][filename]
-
-    with open(path, "rb") as f:
-        files = {"file": (filename, f, "application/pdf")}
-        data = {
-            "filename": filename,
-            "start_page": state.get("start_page", 1),
-            "end_page": state.get("end_page", 1),
-        }
-        headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_TOKEN')}"}
-        resp = requests.post(EXTRACT_TABLES_API, files=files, data=data, headers=headers)
-
-    if resp.status_code != 200:
-        raise RuntimeError(f"Extract tables API failed: {resp.text}")
-
-    js = resp.json()
-    state["tables"] = js.get("tables", js)
+# import os
+# import requests
+
+# EXTRACT_TABLES_API = "https://point9-extract-text-and-table.hf.space/api/tables"  # Replace with your space URL
+
+# def extract_tables_remote(state):
+#     filename = state["filename"]
+#     path = state["temp_files"][filename]
+
+#     with open(path, "rb") as f:
+#         files = {"file": (filename, f, "application/pdf")}
+#         data = {
+#             "filename": filename,
+#             "start_page": state.get("start_page", 1),
+#             "end_page": state.get("end_page", 1),
+#         }
+#         headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_TOKEN')}"}
+#         resp = requests.post(EXTRACT_TABLES_API, files=files, data=data, headers=headers)
+
+#     if resp.status_code != 200:
+#         raise RuntimeError(f"Extract tables API failed: {resp.text}")
+
+#     js = resp.json()
+#     state["tables"] = js.get("tables", js)
+#     return state
+
+import os
+import requests
+
+EXTRACT_TABLES_API = "https://point9-extract-text-and-table.hf.space/api/tables"  # Replace with your space URL
+
+def extract_tables_remote(state):
+    filename = state["filename"]
+    path = state["temp_files"][filename]
+    
+    # Extract filename with extension from the path
+    filename_with_extension = os.path.basename(path)
+    
+    with open(path, "rb") as f:
+        files = {"file": (filename_with_extension, f, "application/pdf")}
+        data = {
+            "filename": filename_with_extension,  # Use filename with extension
+            "start_page": state.get("start_page", 1),
+            "end_page": state.get("end_page", 1),
+        }
+        headers = {"Authorization": f"Bearer {os.getenv('HUGGINGFACE_API_TOKEN')}"}
+        resp = requests.post(EXTRACT_TABLES_API, files=files, data=data, headers=headers)
+
+    if resp.status_code != 200:
+        raise RuntimeError(f"Extract tables API failed: {resp.text}")
+
+    js = resp.json()
+    state["tables"] = js.get("tables", js)
     return state