Update app caching in Dockerfile

Dockerfile

@@ -22,18 +22,15 @@ ENV SENTENCE_TRANSFORMERS_HOME="/home/user/.cache/huggingface/sentence-transform
 RUN mkdir -p $SENTENCE_TRANSFORMERS_HOME && \
     chown -R user:user /home/user/.cache/huggingface
 
-# Download and persist SentenceTransformer model during build stage
-RUN python -c "from sentence_transformers import SentenceTransformer; \
-    SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2', cache_folder='/home/user/.cache/huggingface/sentence-transformers')"
-
-# Copy model to ensure availability at runtime
-RUN cp -r /home/user/.cache/huggingface/sentence-transformers /app/model_cache
+# Download and persist the model properly (USE snapshot_download)
+RUN python -c "from huggingface_hub import snapshot_download; \
+    snapshot_download(repo_id='sentence-transformers/all-MiniLM-L6-v2', cache_dir='/app/model_cache')"
 
 # Ensure ownership and permissions remain intact
-RUN chown -R user:user /home/user/.cache/huggingface
+RUN chown -R user:user /app/model_cache
 
-# EXPOSE 7860 (HF automatically maps it)
+# Expose application port
 EXPOSE 7860
 
-# Run server with uvicorn
+# Run server
 CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -68,15 +68,17 @@ model_cache_dir = "/app/model_cache"
 from huggingface_hub import snapshot_download
 print("⏳ Checking or downloading the all-MiniLM-L6-v2 model from huggingface_hub...")
 # st.write("⏳ Checking or downloading the all-MiniLM-L6-v2 model from huggingface_hub...")
-if os.path.exists(model_cache_dir):
+# First, try loading from our copied cache
+if os.path.exists(model_cache_dir) and os.path.exists(os.path.join(model_cache_dir, "config.json")):
     print(f"✅ Found cached model at {model_cache_dir}")
     model_loc = model_cache_dir
-else: 
+# Else, try loading backup from snapshot_download
+else:
     try:
         model_loc = snapshot_download(
             repo_id="sentence-transformers/all-MiniLM-L6-v2",
-            cache_dir=hf_cache_dir, # Set directly also fine `os.environ["HF_HOME"]`
-            local_files_only=True   # 🚨 Avoids re-downloading / fetch from internet (set False for local dev)
+            cache_dir=hf_cache_dir,
+            local_files_only=True  # ✅ Ensure it's loaded from cache
         )
         print(f"✅ Model loaded from local cache: {model_loc}")
     except Exception as e:
@@ -85,10 +87,9 @@ else:
         model_loc = snapshot_download(
             repo_id="sentence-transformers/all-MiniLM-L6-v2",
             cache_dir=hf_cache_dir,
-            local_files_only=False   # ⬇️ Fallback
+            local_files_only=False  # ✅ Fallback to online download
         )
-        print(f"✅ Model directory reloaded: {model_loc}")
-# st.write(f"✅ Model directory: {model_loc}")
+        print(f"✅ Model directory after retry: {model_loc}")
 
 from sentence_transformers import SentenceTransformer
 print("📥 **Loading Embedding Model...**")