# ════════════════════════════════════════════════════════════════
# Stage 1: Build the React frontend
# ════════════════════════════════════════════════════════════════
FROM node:20-alpine AS frontend-builder

WORKDIR /app/frontend

# Install dependencies first (better layer caching)
COPY frontend/package*.json ./
RUN npm ci --prefer-offline && npm cache clean --force

# Build
COPY frontend .
RUN npm run build

# ════════════════════════════════════════════════════════════════
# Stage 2: Build the FastAPI backend (HybridRAG)
# ════════════════════════════════════════════════════════════════
FROM python:3.10-slim AS backend

WORKDIR /app

# ── System dependencies ────────────────────────────────────────
RUN apt-get update && apt-get install -y \
    build-essential \
    curl \
    git \
    git-lfs \
    && rm -rf /var/lib/apt/lists/* \
    && git lfs install --skip-repo

# ── Non-root user (security best practice) ────────────────────
RUN addgroup --system appgroup && \
    adduser --system --ingroup appgroup --no-create-home appuser

# ── CPU-only PyTorch FIRST (saves ~1.8GB vs full CUDA package) ─
RUN pip install --no-cache-dir \
    torch==2.2.2+cpu \
    --index-url https://download.pytorch.org/whl/cpu

# ── Python production dependencies ────────────────────────────
COPY backend/requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# ── Application code ──────────────────────────────────────────
COPY backend ./backend

# ── Validate data files are real content (not LFS pointers) ───
RUN find /app/backend/data -type f \( -name "*.pdf" -o -name "*.docx" -o -name "*.xlsx" \) | \
    while read f; do \
      size=$(stat -c%s "$f"); \
      if [ "$size" -lt 500 ]; then \
        echo "WARNING: $f is an LFS pointer stub ($size bytes) - RAG will be empty!"; \
      fi; \
    done

# ── React frontend static files ───────────────────────────────
RUN mkdir -p ./backend/static
COPY --from=frontend-builder /app/frontend/dist ./backend/static

# ── Permissions ───────────────────────────────────────────────
# ChromaDB will be downloaded at startup from HF Dataset into data/chroma_db
RUN mkdir -p /app/backend/data/chroma_db /app/backend/storage/chroma_db /app/backend/storage/chat_db && \
    chown -R appuser:appgroup /app

USER appuser

# ── Tạo thư mục cache /tmp (writable bởi appuser) ────────────
RUN mkdir -p /tmp/huggingface /tmp/sentence-transformers /tmp/transformers /tmp/.cache && \
    chown -R appuser:appgroup /tmp/huggingface /tmp/sentence-transformers /tmp/transformers /tmp/.cache

# ── Environment ───────────────────────────────────────────────
ENV PYTHONPATH=/app/backend \
    PYTHONUNBUFFERED=1 \
    PYTHONDONTWRITEBYTECODE=1 \
    HF_HOME=/tmp/huggingface \
    HUGGINGFACE_HUB_CACHE=/tmp/huggingface/hub \
    SENTENCE_TRANSFORMERS_HOME=/tmp/sentence-transformers \
    TRANSFORMERS_CACHE=/tmp/transformers \
    XDG_CACHE_HOME=/tmp/.cache

# ── HuggingFace Spaces port ───────────────────────────────────
EXPOSE 7860

# ── Health check ──────────────────────────────────────────────
HEALTHCHECK --interval=30s --timeout=15s --start-period=600s --retries=3 \
    CMD curl -f http://localhost:7860/api/v1/health || exit 1

# ── Run app ───────────────────────────────────────────────────
CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860", \
     "--workers", "1", "--log-level", "info"]