Dockerfile

FROM python:3.11-slim

# Install Node.js (for promptfoo) and system tools
RUN apt-get update && apt-get install -y \
    curl \
    git \
    nodejs \
    npm \
    && rm -rf /var/lib/apt/lists/*

# Install promptfoo globally
RUN npm install -g promptfoo

WORKDIR /app

# --- ULTIMATE FIX: Force Specific Platform ---
RUN pip download --only-binary :all: \
    --platform manylinux_2_17_x86_64 \
    --python-version 311 \
    --implementation cp \
    --abi cp311 \
    "llama-cpp-python" \
    --extra-index-url "https://abetlen.github.io/llama-cpp-python/whl/cpu" \
    -d /tmp/wheels

RUN pip install --no-cache-dir /tmp/wheels/llama*.whl && rm -rf /tmp/wheels

# --- Install Standard Requirements ---
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy application code
COPY . .

EXPOSE 7860

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