FROM python:3.10

ENV PYTHONDONTWRITEBYTECODE=1 \
    PYTHONUNBUFFERED=1 \
    PIP_NO_CACHE_DIR=1 \
    NLTK_DATA=/usr/local/share/nltk_data

WORKDIR /app

COPY requirements.txt .

RUN pip install --upgrade pip && \
    pip install -r requirements.txt && \
    python -c "import nltk; nltk.download('wordnet', download_dir='/usr/local/share/nltk_data'); nltk.download('omw-1.4', download_dir='/usr/local/share/nltk_data')"

COPY . .

RUN mkdir -p /app/data /app/results
RUN python download_beir_datasets.py

EXPOSE 7860

CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]