masterllm / utilities /extract_tables.py
redhairedshanks1's picture
Update utilities/extract_tables.py
1482837
# 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
# Hardcoded API URL - DO NOT CHANGE
EXTRACT_TABLES_API = "https://point9-extract-text-and-table.hf.space/api/tables"
def extract_tables_remote(state):
"""
Extract tables from documents via API.
All file types are sent to the API (no local processing).
Args:
state: Dict with 'filename', 'temp_files', 'start_page', 'end_page'
Returns:
state: Dict with 'tables' key containing list of extracted tables
"""
filename = state["filename"]
path = state["temp_files"][filename]
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)
file_size = os.path.getsize(path)
print(f"\n{'='*60}")
print(f"πŸ“Š 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)
files = {"file": (file_basename, f, "application/pdf")}
# IMPORTANT: Convert page numbers to strings (matches curl -F format)
data = {
"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')}"}
print(f"πŸš€ 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"\nπŸ“₯ API Response:")
print(f"Status Code: {resp.status_code}")
if resp.status_code != 200:
print(f"❌ Error: {resp.text[:500]}")
raise RuntimeError(f"Extract tables API failed with status {resp.status_code}: {resp.text}")
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"\nπŸ“Š 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)}")