Medical-Chatbot / Dockerfile
LiamKhoaLe's picture
Lazy load models
4f5341e
FROM python:3.11
# Create and use a non-root user (optional)
RUN useradd -m -u 1000 user
USER user
ENV PATH="/home/user/.local/bin:$PATH"
# Set working directory
WORKDIR /app
# Copy all project files to the container
COPY . .
# Install dependencies
RUN pip install --no-cache-dir -r requirements.txt
# Set Hugging Face cache directory to persist model downloads
ENV HF_HOME="/home/user/.cache/huggingface"
ENV SENTENCE_TRANSFORMERS_HOME="/home/user/.cache/huggingface/sentence-transformers"
ENV MEDGEMMA_HOME="/home/user/.cache/huggingface/sentence-transformers"
# Create cache directories and ensure permissions
RUN mkdir -p /app/model_cache /home/user/.cache/huggingface/sentence-transformers && \
chown -R user:user /app/model_cache /home/user/.cache/huggingface
# Control preloading to avoid exhausting build disk on HF Spaces
ENV PRELOAD_TRANSLATORS="0"
ENV EMBEDDING_HALF="0"
# Pre-load model in a separate script (translation preload disabled by default)
RUN python /app/download_model.py && python /app/warmup.py
# Ensure ownership and permissions remain intact
RUN chown -R user:user /app/model_cache
# Expose port
EXPOSE 7860
# Run the application
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "1"]