fix nltk access

Dockerfile

@@ -1,5 +1,7 @@
 FROM python:3.9-slim
 
+USER root
+
 WORKDIR /app
 
 # Install system dependencies
@@ -47,5 +49,9 @@ RUN mkdir -m 777 /tmp/NUMBA_CACHE_DIR /tmp/MPLCONFIGDIR
 ENV NUMBA_CACHE_DIR=/tmp/NUMBA_CACHE_DIR/
 ENV MPLCONFIGDIR=/tmp/MPLCONFIGDIR/
 
+# Pre-download NLTK data
+RUN python -m nltk.downloader -d /app/nltk_data punkt stopwords wordnet
+
+
 # Run the application
 CMD ["python", "app.py"]

app.py

@@ -22,6 +22,21 @@ UPLOAD_DIR = os.path.join(DATA_DIR, "uploads")
 RESULTS_DIR = os.path.join(DATA_DIR, "results")
 CONFIG_DIR = os.path.join(DATA_DIR, "configs")
 
+# Set NLTK_DATA directory to a writable location within the container
+os.environ["NLTK_DATA"] = "/app/nltk_data"
+os.makedirs("/app/nltk_data", exist_ok=True)
+
+# Pre-download NLTK data during startup
+try:
+    import nltk
+    for package in ['punkt', 'stopwords', 'wordnet']:
+        try:
+            nltk.download(package, quiet=True, download_dir='/app/nltk_data')
+        except:
+            pass
+except Exception as e:
+    print(f"Warning: NLTK setup failed: {e}")
+
 # Check if VERSA is installed
 def check_versa_installation():
     """Check if VERSA is properly installed"""