name: test

concurrency:
  # For PRs, later CI runs preempt previous ones. e.g. a force push on a PR
  # cancels running CI jobs and starts all new ones.
  #
  # For non-PR pushes, concurrency.group needs to be unique for every distinct
  # CI run we want to have happen. Use run_id, which in practice means all
  # non-PR CI runs will be allowed to run without preempting each other.
  group: ${{ github.workflow }}-$${{ github.pull_request.number || github.run_id }}
  cancel-in-progress: true

on:
  pull_request:
    paths:
      - '**/*'
      - '!docs/**'
      - '!README.md'

jobs:
  changes:
    runs-on: ubuntu-latest
    outputs:
      GENERATE: ${{ steps.changes.outputs.GENERATE }}
      GENERATE_CUDA: ${{ steps.changes.outputs.GENERATE_CUDA }}
      GENERATE_ROCM: ${{ steps.changes.outputs.GENERATE_ROCM }}
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0
      - id: changes
        run: |
          changed() {
            git diff-tree -r --no-commit-id --name-only \
              $(git merge-base ${{ github.event.pull_request.base.sha }} ${{ github.event.pull_request.head.sha }}) \
              ${{ github.event.pull_request.head.sha }} \
              | xargs python3 -c "import sys; from pathlib import Path; print(any(Path(x).match(glob) for x in sys.argv[1:] for glob in '$*'.split(' ')))"
          }

          {
            echo GENERATE=$(changed 'llm/llama.cpp' 'llm/patches/**' 'llm/ext_server/**' 'llm/generate/**')
            echo GENERATE_CUDA=$(changed 'llm/llama.cpp' 'llm/patches/**' 'llm/ext_server/**' 'llm/generate/**')
            echo GENERATE_ROCM=$(changed 'llm/llama.cpp' 'llm/patches/**' 'llm/ext_server/**' 'llm/generate/**')
          } >>$GITHUB_OUTPUT

  generate:
    needs: [changes]
    if: ${{ needs.changes.outputs.GENERATE == 'True' }}
    strategy:
      matrix:
        os: [ubuntu-latest, macos-latest, windows-2019]
        arch: [amd64, arm64]
        exclude:
          - os: ubuntu-latest
            arch: arm64
          - os: windows-2019
            arch: arm64
    runs-on: ${{ matrix.os }}
    env:
      GOARCH: ${{ matrix.arch }}
      CGO_ENABLED: '1'
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-go@v5
        with:
          go-version: "stable"
          cache: true
      - run: go get ./...
      - run: |
          $gopath=(get-command go).source | split-path -parent
          $gccpath=(get-command gcc).source | split-path -parent
          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
          cd $env:GITHUB_WORKSPACE
          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
          $env:PATH="$gopath;$gccpath;$env:PATH"
          echo $env:PATH
          go generate -x ./...
        if: ${{ startsWith(matrix.os, 'windows-') }}
        name: 'Windows Go Generate'
      - run: go generate -x ./...
        if: ${{ ! startsWith(matrix.os, 'windows-') }}
        name: 'Unix Go Generate'
      - run: go build .
      - uses: actions/upload-artifact@v4
        with:
          name: ${{ matrix.os }}-${{ matrix.arch }}-libraries
          path: |
            llm/build/**/bin/*
            llm/build/**/*.a
  generate-cuda:
    needs: [changes]
    if: ${{ needs.changes.outputs.GENERATE_CUDA == 'True' }}
    strategy:
      matrix:
        cuda-version:
          - '11.8.0'
    runs-on: linux
    container: nvidia/cuda:${{ matrix.cuda-version }}-devel-ubuntu20.04
    steps:
      - run: |
          apt-get update && apt-get install -y git build-essential curl
          curl -fsSL https://github.com/Kitware/CMake/releases/download/v3.28.1/cmake-3.28.1-linux-x86_64.tar.gz \
            | tar -zx -C /usr --strip-components 1
        env:
          DEBIAN_FRONTEND: noninteractive
      - uses: actions/checkout@v4
      - uses: actions/setup-go@v4
        with:
          go-version-file: go.mod
          cache: true
      - run: go get ./...
      - run: |
          git config --global --add safe.directory /__w/ollama/ollama
          go generate -x ./...
        env:
          OLLAMA_SKIP_CPU_GENERATE: '1'
      - uses: actions/upload-artifact@v4
        with:
          name: cuda-${{ matrix.cuda-version }}-libraries
          path: |
            llm/build/**/bin/*
            dist/windows-amd64/**
  generate-rocm:
    needs: [changes]
    if: ${{ needs.changes.outputs.GENERATE_ROCM == 'True' }}
    strategy:
      matrix:
        rocm-version:
          - '6.1.2'
    runs-on: linux
    container: rocm/dev-ubuntu-20.04:${{ matrix.rocm-version }}
    steps:
      - run: |
          apt-get update && apt-get install -y git build-essential curl rocm-libs
          curl -fsSL https://github.com/Kitware/CMake/releases/download/v3.28.1/cmake-3.28.1-linux-x86_64.tar.gz \
            | tar -zx -C /usr --strip-components 1
        env:
          DEBIAN_FRONTEND: noninteractive
      - uses: actions/checkout@v4
      - uses: actions/setup-go@v4
        with:
          go-version-file: go.mod
          cache: true
      - run: go get ./...
      - run: |
          git config --global --add safe.directory /__w/ollama/ollama
          go generate -x ./...
        env:
          OLLAMA_SKIP_CPU_GENERATE: '1'
      - uses: actions/upload-artifact@v4
        with:
          name: rocm-${{ matrix.rocm-version }}-libraries
          path: |
            llm/build/**/bin/*
            dist/windows-amd64/**

  # ROCm generation step
  generate-windows-rocm:
    needs: [changes]
    if: ${{ needs.changes.outputs.GENERATE_ROCM == 'True' }}
    runs-on: windows
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-go@v5
        with:
          go-version: "stable"
          cache: true
      - name: 'Install ROCm'
        run: |
          $ErrorActionPreference = "Stop"
          write-host "downloading AMD HIP Installer"
          Invoke-WebRequest -Uri "https://download.amd.com/developer/eula/rocm-hub/AMD-Software-PRO-Edition-24.Q3-WinSvr2022-For-HIP.exe" -OutFile "${env:RUNNER_TEMP}\rocm-install.exe"
          write-host "Installing AMD HIP"
          Start-Process "${env:RUNNER_TEMP}\rocm-install.exe" -ArgumentList '-install' -NoNewWindow -Wait
          write-host "Completed AMD HIP"
      - name: 'Verify ROCm'
        run: |
          & 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' --version
      - run: go get ./...
      - run: |
          $gopath=(get-command go).source | split-path -parent
          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
          cd $env:GITHUB_WORKSPACE
          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
          $env:PATH="$gopath;$env:PATH"
          $env:OLLAMA_SKIP_CPU_GENERATE="1"
          $env:HIP_PATH=$(Resolve-Path 'C:\Program Files\AMD\ROCm\*\bin\clang.exe' | split-path | split-path)
          go generate -x ./...
        name: go generate
        env:
          OLLAMA_SKIP_CPU_GENERATE: '1'
      # TODO - do we need any artifacts?

  # CUDA generation step
  generate-windows-cuda:
    needs: [changes]
    if: ${{ needs.changes.outputs.GENERATE_CUDA == 'True' }}
    runs-on: windows
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-go@v5
        with:
          go-version: "stable"
          cache: true
      - name: 'Install CUDA'
        run: |
          $ErrorActionPreference = "Stop"
          write-host "downloading CUDA Installer"
          Invoke-WebRequest -Uri "https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.89_win10.exe" -OutFile "${env:RUNNER_TEMP}\cuda-install.exe"
          write-host "Installing CUDA"
          Start-Process "${env:RUNNER_TEMP}\cuda-install.exe" -ArgumentList '-s' -NoNewWindow -Wait
          write-host "Completed CUDA"
          $cudaPath=((resolve-path "c:\Program Files\NVIDIA*\CUDA\v*\bin\nvcc.exe")[0].path | split-path | split-path)
          $cudaVer=($cudaPath | split-path -leaf ) -replace 'v(\d+).(\d+)', '$1_$2' 
          echo "$cudaPath\bin" >> $env:GITHUB_PATH
          echo "CUDA_PATH=$cudaPath" >> $env:GITHUB_ENV
          echo "CUDA_PATH_V${cudaVer}=$cudaPath" >> $env:GITHUB_ENV
          echo "CUDA_PATH_VX_Y=CUDA_PATH_V${cudaVer}" >> $env:GITHUB_ENV
      - name: 'Verify CUDA'
        run: nvcc -V
      - run: go get ./...
      - name: go generate
        run: |
          $gopath=(get-command go).source | split-path -parent
          $cudabin=(get-command nvcc).source | split-path
          & "C:\Program Files (x86)\Microsoft Visual Studio\2019\Enterprise\Common7\Tools\Launch-VsDevShell.ps1"
          cd $env:GITHUB_WORKSPACE
          $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
          $env:PATH="$gopath;$cudabin;$env:PATH"
          $env:OLLAMA_SKIP_CPU_GENERATE="1"
          go generate -x ./...
        env:
          OLLAMA_SKIP_CPU_GENERATE: '1'
      # TODO - do we need any artifacts?

  lint:
    strategy:
      matrix:
        os: [ubuntu-latest, macos-latest, windows-2019]
        arch: [amd64, arm64]
        exclude:
          - os: ubuntu-latest
            arch: arm64
          - os: windows-2019
            arch: arm64
          - os: macos-latest
            arch: amd64
    runs-on: ${{ matrix.os }}
    env:
      GOARCH: ${{ matrix.arch }}
      CGO_ENABLED: '1'
    steps:
      - uses: actions/checkout@v4
        with:
          submodules: recursive
      - uses: actions/setup-go@v5
        with:
          go-version: "stable"
          cache: false
      - run: |
          case ${{ matrix.arch }} in
            amd64) echo ARCH=x86_64 ;;
            arm64) echo ARCH=arm64 ;;
          esac >>$GITHUB_ENV
        shell: bash
      - run: |
          mkdir -p llm/build/linux/$ARCH/stub/bin
          touch llm/build/linux/$ARCH/stub/bin/ollama_llama_server
        if: ${{ startsWith(matrix.os, 'ubuntu-') }}
      - run: |
          mkdir -p llm/build/darwin/$ARCH/stub/bin
          touch llm/build/darwin/$ARCH/stub/bin/ollama_llama_server
        if: ${{ startsWith(matrix.os, 'macos-') }}
      - uses: golangci/golangci-lint-action@v6
        with:
          args: --timeout 8m0s -v ${{ startsWith(matrix.os, 'windows-') && '' || '--disable gofmt --disable goimports' }}
  test:
    strategy:
      matrix:
        os: [ubuntu-latest, macos-latest, windows-2019]
        arch: [amd64]
        exclude:
          - os: ubuntu-latest
            arch: arm64
          - os: windows-2019
            arch: arm64
    runs-on: ${{ matrix.os }}
    env:
      GOARCH: ${{ matrix.arch }}
      CGO_ENABLED: '1'
      OLLAMA_CPU_TARGET: 'static'
      OLLAMA_SKIP_CPU_GENERATE: '1'
      OLLAMA_SKIP_METAL_GENERATE: '1'
    steps:
      - uses: actions/checkout@v4
        with:
          submodules: recursive
      - uses: actions/setup-go@v5
        with:
          go-version: "stable"
          cache: true
      - run: |
          case ${{ matrix.arch }} in
            amd64) echo ARCH=x86_64 ;;
            arm64) echo ARCH=arm64 ;;
          esac >>$GITHUB_ENV
        shell: bash
      - run: |
          mkdir -p llm/build/linux/$ARCH/stub/bin
          touch llm/build/linux/$ARCH/stub/bin/ollama_llama_server
        if: ${{ startsWith(matrix.os, 'ubuntu-') }}
      - run: |
          mkdir -p llm/build/darwin/$ARCH/stub/bin
          touch llm/build/darwin/$ARCH/stub/bin/ollama_llama_server
        if: ${{ startsWith(matrix.os, 'macos-') }}
        shell: bash
      - run: go generate ./...
      - run: go build
      - run: go test -v ./...
      - uses: actions/upload-artifact@v4
        with:
          name: ${{ matrix.os }}-binaries
          path: ollama