# syntax=docker/dockerfile:1
# =============================================================================
# Placement Policy Advisor — single-container image.
# Corpus verification, model download, and vector indexing all run at BUILD time
# so the (ephemeral, free-tier) runtime filesystem already contains everything
# and the first request pays no download/indexing penalty.
# =============================================================================
FROM python:3.11-slim

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

# onnxruntime (via FastEmbed) needs libgomp at runtime. Install as root first.
RUN apt-get update \
 && apt-get install -y --no-install-recommends libgomp1 \
 && rm -rf /var/lib/apt/lists/*

# --- Non-root user (UID 1000) ------------------------------------------------
RUN useradd -m -u 1000 user
USER user
ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH
WORKDIR /home/user/app

# --- Python dependencies (installed into the user site) ----------------------
COPY --chown=user:user requirements.txt ./
RUN pip install --user --upgrade pip \
 && pip install --user -r requirements.txt

# --- Application source ------------------------------------------------------
COPY --chown=user:user . ./

# --- Build-time data preparation ---------------------------------------------
# 1) verify curated Contextmd/ against source PDFs (quality gate),
# 2) cache embedding model + init Qdrant,
# 3) chunk Contextmd/ + embed + upsert into the local Qdrant collection.
RUN python convert_docs.py \
 && python download_models.py \
 && python data_indexer.py

# --- Networking --------------------------------------------------------------
# APP_PORT is read from the environment at runtime by main.py; this ARG/ENV pair
# only provides the concrete value EXPOSE requires and a sensible default.
ARG APP_PORT=7860
ENV APP_PORT=${APP_PORT}
EXPOSE ${APP_PORT}

CMD ["python", "main.py"]