Dockerfile

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

# Set environment variables
ENV PYTHONDONTWRITEBYTECODE=1 \
    PYTHONUNBUFFERED=1 \
    PYTHONIOENCODING=UTF-8 \
    HF_HOME=/app/cache

# Install system dependencies
RUN apt-get update && apt-get install -y --no-install-recommends curl \
    && rm -rf /var/lib/apt/lists/* \
    && useradd -m -u 1000 user

WORKDIR /app

# --- LAYER 1: Dependencies ---
COPY --chown=user:user requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# --- LAYER 2: Download Models (Cached) ---
# We download models for 384, 768, and 1024 dimensions.
# 384 dim: BAAI/bge-small-en-v1.5
# 768 dim: BAAI/bge-base-en-v1.5
# 1024 dim: BAAI/bge-large-en-v1.5
RUN python3 -c "from huggingface_hub import snapshot_download; \
    snapshot_download(repo_id='BAAI/bge-small-en-v1.5', local_dir='./models/bge-384'); \
    snapshot_download(repo_id='BAAI/bge-base-en-v1.5', local_dir='./models/bge-768'); \
    snapshot_download(repo_id='BAAI/bge-large-en-v1.5', local_dir='./models/bge-1024')"

# --- LAYER 3: Application Code ---
COPY --chown=user:user . .

# Ensure permissions
RUN mkdir -p $HF_HOME && chown -R user:user /app/cache && chown -R user:user /app/models

# Switch user
USER user

# Expose port
EXPOSE 7860

# Start script
CMD bash -c "while true; do curl -s https://xce009-embeddings-api.hf.space/ping >/dev/null && sleep 300; done & uvicorn main:app --host 0.0.0.0 --port 7860 --workers 4 --loop asyncio"