Dockerfile

# Use Ubuntu 22.04 as a public base image to avoid GHCR permissions
FROM ubuntu:22.04

# Disable interactive prompts and set timezone for tzdata
ENV DEBIAN_FRONTEND=noninteractive
ENV TZ=Etc/UTC

# Use bash with strict modes for debugging
SHELL ["/bin/bash", "-euxo", "pipefail", "-c"]

# 1. Install OS-level dependencies
RUN echo "### STEP 1: Installing OS-level dependencies" && \
    apt-get update && \
    apt-get install -y --no-install-recommends \
        tzdata \
        build-essential \
        cmake \
        libopenblas-dev \
        python3 \
        python3-pip \
        python3-opencv && \
    rm -rf /var/lib/apt/lists/*

# 2. Prepare the application directory
RUN echo "### STEP 2: Preparing application directory"
WORKDIR /app
COPY requirements.txt ./
COPY app.py ./
# Copy any other source files or directories needed by your app

# 3. Install Python dependencies (excluding llama-cpp-python)
RUN echo "### STEP 3: Installing Python dependencies" && \
    pip3 install --upgrade pip && \
    pip3 install --no-cache-dir -r requirements.txt

# 4. Build and install llama-cpp-python from source with OpenBLAS
RUN echo "### STEP 4: Building and installing llama-cpp-python with OpenBLAS" && \
    export CMAKE_ARGS="-DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS" && \
    pip3 install --no-cache-dir --force-reinstall --no-binary llama-cpp-python llama-cpp-python

# 5. Expose port and launch the application
RUN echo "### STEP 5: Finalizing Docker image"
EXPOSE 7860
CMD ["python3", "app.py"]