Improve Dockerfile HF_TOKEN handling with better error messages

Dockerfile

@@ -26,12 +26,22 @@ RUN uv sync
 # Create data directory if it doesn't exist
 RUN mkdir -p $HOME/app/data
 
-# Install huggingface_hub and datasets and login in one step
-ARG HF_TOKEN
-RUN uv pip install huggingface_hub datasets && \
-    uv run python -c "from huggingface_hub import login; login(token='${HF_TOKEN}')"
+# Install huggingface_hub and datasets
+RUN uv pip install huggingface_hub datasets
 
 # Download PDFs from Hugging Face dataset
+# Use a build arg for the token
+ARG HF_TOKEN
+# Echo first character of token (for debugging, redacted for security)
+RUN echo "HF_TOKEN starts with: ${HF_TOKEN:0:1}..."
+# Try to use the token more explicitly
+RUN if [ -n "${HF_TOKEN}" ]; then \
+        echo "Token is available, attempting login"; \
+        python -c "from huggingface_hub import login; login(token='${HF_TOKEN}')"; \
+    else \
+        echo "No token available, skipping login"; \
+    fi
+
 RUN python download_pdfs.py
 
 # Run preprocessing to generate the embeddings