Update interim/app.py

interim/app.py

@@ -6,19 +6,23 @@ import time
 import streamlit as st
 import nltk
 import tempfile
+import subprocess
+
+# Pin NLTK to version 3.9.1
+REQUIRED_NLTK_VERSION = "3.9.1"
+subprocess.run([sys.executable, "-m", "pip", "install", f"nltk=={REQUIRED_NLTK_VERSION}"])
 
 # Set up temporary directory for NLTK resources
 nltk_data_path = os.path.join(tempfile.gettempdir(), "nltk_data")
 os.makedirs(nltk_data_path, exist_ok=True)
-nltk.data.path = [nltk_data_path]  # Force NLTK to use only the temp directory
+nltk.data.path.append(nltk_data_path)
 
-# Force clean download of 'punkt'
+# Download 'punkt_tab' for compatibility
 try:
-    print("Ensuring NLTK 'punkt' resource is downloaded...")
-    if not os.path.exists(os.path.join(nltk_data_path, "tokenizers/punkt")):
-        nltk.download("punkt", download_dir=nltk_data_path)
+    print("Ensuring NLTK 'punkt_tab' resource is downloaded...")
+    nltk.download("punkt_tab", download_dir=nltk_data_path)
 except Exception as e:
-    print(f"Error downloading NLTK 'punkt': {e}")
+    print(f"Error downloading NLTK 'punkt_tab': {e}")
     raise e
 
 sys.path.append(os.path.abspath("."))