Update app.py

app.py

@@ -4,6 +4,7 @@ from ocr_utils import extract_text_from_pdf_with_tesseract_or_layoutlm
 import os
 import tempfile
 from tqdm import tqdm
+import subprocess
 
 # Initialize global state for failed records and uploaded files
 failed_records = []
@@ -14,6 +15,14 @@ def is_pdf_file(file_bytes):
     valid_pdf_header = b'%PDF-'
     return file_bytes.startswith(valid_pdf_header) if file_bytes else False
 
+def check_poppler():
+    """Check if poppler-utils is installed and in PATH."""
+    try:
+        subprocess.run(['pdftoppm', '-v'], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        return True
+    except FileNotFoundError:
+        return False
+
 def save_failed_record(pdf_name, object_name, error, mappings):
     """Log failed records for reconciliation."""
     global failed_records
@@ -39,11 +48,16 @@ def process_contract(uploaded_files, object_name, manual_mappings):
     for i, file_bytes in enumerate(uploaded_files):
         print(f"File {i} header: {file_bytes[:5]}")
 
+    # Check for poppler-utils
+    if not check_poppler():
+        return "❌ Error: poppler-utils is not installed or not in PATH. Please install it (e.g., 'sudo apt-get install poppler-utils' on Linux).", None, failed_records, "0/0"
+
     # Mock Salesforce object fields (replace with dynamic logic later)
     mock_object_fields = ["Name", "Description", "Amount", "Date"] if object_name else []
     total_files = len(uploaded_files)
     processed_files = 0
     results = []
+    ai_result = None  # Initialize to avoid UnboundLocalError
     with tqdm(total=total_files, desc="Processing PDFs") as pbar:
         for i, file_bytes in enumerate(uploaded_files):
             # Generate a filename based on index since name is not available with type="binary"