Dockerfile

# Kerdos AI — Custom LLM Chat (Demo)
# Multi-stage Docker build — supports both Gradio UI and FastAPI REST server
#
# Build for Gradio (default):
#   docker build -t kerdos-rag .
#   docker run -p 7860:7860 -e HF_TOKEN=hf_... kerdos-rag
#
# Build for REST API:
#   docker build --build-arg MODE=api -t kerdos-rag-api .
#   docker run -p 8000:8000 -e HF_TOKEN=hf_... kerdos-rag-api

ARG MODE=serve

FROM python:3.11-slim AS base

# System dependencies for PyMuPDF and FAISS
RUN apt-get update && apt-get install -y --no-install-recommends \
    build-essential \
    libgomp1 \
    && rm -rf /var/lib/apt/lists/*

WORKDIR /app

# Install Python dependencies first (layer-cached)
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy source
COPY . .

# Install the package in editable mode so kerdos-rag CLI is available
RUN pip install --no-cache-dir -e .

# ── Gradio mode ───────────────────────────────────────────────
FROM base AS serve
EXPOSE 7860
ENV GRADIO_SERVER_NAME=0.0.0.0
ENV GRADIO_SERVER_PORT=7860
CMD ["kerdos-rag", "serve", "--host", "0.0.0.0", "--port", "7860"]

# ── FastAPI REST mode ─────────────────────────────────────────
FROM base AS api
EXPOSE 8000
CMD ["kerdos-rag", "api", "--host", "0.0.0.0", "--port", "8000"]

# Select the right stage based on BUILD ARG
FROM ${MODE}