nltk fix

Dockerfile

@@ -12,14 +12,8 @@ COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
 # download NLTK data **after** nltk is installed
-RUN python - <<EOF
-import nltk
-nltk.download('punkt', download_dir='/usr/local/nltk_data')
-nltk.download('punkt_tab', download_dir='/usr/local/nltk_data')
-nltk.download('averaged_perceptron_tagger', download_dir='/usr/local/nltk_data')
-nltk.download('wordnet', download_dir='/usr/local/nltk_data')
-EOF
-ENV NLTK_DATA=/usr/local/nltk_data
+COPY nltk_data /app/nltk_data
+ENV NLTK_DATA=/app/nltk_data
 
 # copy application code
 COPY app.py rag_engine.py ./

rag_engine.py

@@ -10,10 +10,11 @@ from nltk.translate.bleu_score import sentence_bleu
 from nltk.tokenize import word_tokenize
 from rouge_score import rouge_scorer
 import nltk
-try:
-    nltk.data.find("tokenizers/punkt")
-except LookupError:
-    raise RuntimeError("NLTK punkt tokenizer not found inside container. Fix Dockerfile.")
+import os
+
+# Force NLTK to use the bundled data
+NLTK_LOCAL = os.path.join(os.path.dirname(__file__), "nltk_data")
+nltk.data.path.insert(0, NLTK_LOCAL)
 
 DEFAULT_LLMS = {
     "gpt2": "gpt2",