Dockerfile

# Use official Python 3.11 slim image
FROM python:3.11-slim

# Set environment variables
ENV PYTHONUNBUFFERED=1 \
    PYTHONDONTWRITEBYTECODE=1 \
    PORT=7860 \
    HOME=/home/user \
    USE_OLLAMA=false \
    LLAMA_THREADS=2

# Install system dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
    build-essential \
    python3-dev \
    tesseract-ocr \
    curl \
    git \
    && apt-get clean && rm -rf /var/lib/apt/lists/*

# Set up user 1000 for Hugging Face Spaces compatibility
RUN useradd -m -u 1000 user
WORKDIR $HOME/app

# Pre-create required directories and assign ownership
RUN mkdir -p models vector_store documents && chown -R user:user models vector_store documents

# Copy requirements and install
COPY requirements.txt .
RUN pip install --no-cache-dir --upgrade pip && \
    pip install --no-cache-dir -r requirements.txt

# Download Gemma 3 1B Instruct GGUF model during build
RUN curl -L -o models/google_gemma-3-1b-it-Q4_K_M.gguf \
    "https://huggingface.co/bartowski/google_gemma-3-1b-it-GGUF/resolve/main/google_gemma-3-1b-it-Q4_K_M.gguf"

# Copy the rest of the application files
COPY --chown=user:user . .

# Switch to the non-root user
USER user

# Pre-download/cache the embedding model under the user's home cache
RUN python -c "from langchain_huggingface import HuggingFaceEmbeddings; HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2')"

# Expose Hugging Face Space port
EXPOSE 7860

# Start application using uvicorn on port 7860
CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]