Dockerfile

# Start from a basic Python 3.10 image
FROM python:3.10-slim

# Set frontend to noninteractive to avoid prompts
ENV DEBIAN_FRONTEND=noninteractive

# Install system dependencies needed for the pre-built wheel to work
RUN apt-get update && apt-get install -y --no-install-recommends \
    build-essential \
    cmake \
    && rm -rf /var/lib/apt/lists/*

# Set the working directory inside the container
WORKDIR /app

# Copy the requirements file first
COPY ./requirements.txt /app/requirements.txt

# Install Python packages
RUN pip install --no-cache-dir --upgrade pip

# This line is correct and fixed your build timeout
RUN pip install --no-cache-dir -r /app/requirements.txt \
    --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu

# Copy the main application file
COPY ./app.py /app/app.py

# Expose the port that FastAPI (uvicorn) will run on
EXPOSE 7860

# --- THIS IS THE FIX ---
# Instead of just running the script, we tell uvicorn to run the "app"
# object inside the "app.py" file. This starts the server.
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]