# Prisme — RAG multimodal (Hugging Face Space, SDK Docker, CPU gratuit)
FROM python:3.11-slim

ENV PYTHONDONTWRITEBYTECODE=1 \
    PYTHONUNBUFFERED=1 \
    PIP_NO_CACHE_DIR=1 \
    PIP_DISABLE_PIP_VERSION_CHECK=1

RUN apt-get update \
 && apt-get install -y --no-install-recommends ffmpeg libglib2.0-0 \
 && rm -rf /var/lib/apt/lists/*

# Hugging Face exécute le conteneur avec l'utilisateur 1000
RUN useradd -m -u 1000 user
USER user
ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH \
    HF_HOME=/home/user/.cache/huggingface \
    TIKTOKEN_CACHE_DIR=/home/user/.cache/tiktoken
WORKDIR /home/user/app

# PyTorch CPU (image beaucoup plus légère que la version CUDA)
COPY --chown=user requirements.txt .
RUN pip install --user torch --index-url https://download.pytorch.org/whl/cpu \
 && pip install --user -r requirements.txt

# Téléchargement des modèles au build : démarrage rapide du Space
RUN python -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('intfloat/multilingual-e5-base')" \
 && python -c "import tiktoken; tiktoken.get_encoding('o200k_base')"

COPY --chown=user . .

EXPOSE 7860
CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860", "--proxy-headers", "--forwarded-allow-ips", "*"]