name: 'Build and Test on macOS'
description: 'Build and test MLX on macOS'

runs:
  using: "composite"
  steps:
    - name: Install dependencies
      shell: sh
      env:
        DEBUG: 1
        CMAKE_ARGS: "-DCMAKE_COMPILE_WARNING_AS_ERROR=ON"
      run: |
        uv pip install --upgrade pip
        uv pip install cmake setuptools nanobind==2.4.0
        uv pip install -e . -v

    - name: Generate package stubs
      shell: bash
      run: |
        uv pip install typing_extensions
        uv run --no-project setup.py generate_stubs

    - name: Install tests dependencies
      shell: sh
      run: |
        uv pip install numpy torch tensorflow unittest-xml-reporting

    - name: Run Python tests
      shell: bash
      env:
        LOW_MEMORY: 1
      run: |
        DEVICE=cpu uv run -m xmlrunner discover -v python/tests -o test-results/cpu
        DEVICE=gpu METAL_DEVICE_WRAPPER_TYPE=1 METAL_DEBUG_ERROR_MODE=0 uv run -m xmlrunner discover -v python/tests -o test-results/gpu
        mpirun --bind-to none -host localhost:8 -np 8 -x DYLD_LIBRARY_PATH=/opt/homebrew/lib/ python python/tests/mpi_test_distributed.py
        mlx.launch --verbose -n 8 python/tests/ring_test_distributed.py -v 2> >(tee -a stderr.log >&2)
        if $(grep "\[WARN\]" stderr.log); then echo "Distributed ring test failed"; exit 1; fi
    
    - name: Build example extension
      shell: bash
      run: |
        cd examples/extensions
        uv pip install -r requirements.txt
        uv run --no-project setup.py build_ext --inplace
        uv run --no-project test.py
    
    - name: Build CPP only
      shell: bash
      run: |
        mkdir -p build
        cd build
        cmake ..
        make -j $(sysctl -n hw.ncpu)
    
    - name: Run CPP tests
      shell: bash
      env:
        DEVICE: gpu
        METAL_DEVICE_WRAPPER_TYPE: 1
        METAL_DEBUG_ERROR_MODE: 0
      run: ./build/tests/tests
    
    - name: Build small binary with JIT
      shell: bash
      run: |
        mkdir -p build
        cd build
        cmake .. -DCMAKE_BUILD_TYPE=MinSizeRel \
          -DBUILD_SHARED_LIBS=ON \
          -DMLX_BUILD_CPU=OFF \
          -DMLX_BUILD_SAFETENSORS=OFF \
          -DMLX_BUILD_GGUF=OFF \
          -DMLX_METAL_JIT=ON
        make -j $(sysctl -n hw.ncpu)
    
    - name: Run Python tests with JIT
      shell: bash
      env:
        LOW_MEMORY: 1
        DEVICE: gpu
        METAL_DEVICE_WRAPPER_TYPE: 1
        METAL_DEBUG_ERROR_MODE: 0
      run: |
        CMAKE_ARGS="-DMLX_METAL_JIT=ON" \
          uv pip install -e . -v
        uv run -m xmlrunner discover \
            -v python/tests \
            -o test-results/gpu_jit