prerelease.yml

name: "Pre-Release"

on:
  push:
    branches: ["main-dev"]
  pull_request:
    branches: ["main-dev"]

env:
  BUILD_TYPE: Release
  GH_TOKEN: ${{ secrets.SEMANTIC_RELEASE_TOKEN }}
  PYTHON_VERSION: 3.11
  SWIFT_VERSION: 5.9
  PYTHONUTF8: 1

# Sets permissions of the GITHUB_TOKEN to allow deployment to GitHub Pages
permissions:
  contents: read

jobs:
  versioning:
    name: Update Version
    runs-on: ubuntu-24.04
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
          persist-credentials: false
      - name: Run TinySemVer
        uses: ashvardanian/tinysemver@v2.0.7
        with:
          verbose: "true"
          version-file: "VERSION"
          update-version-in: |
            Cargo.toml:^version = "(\d+\.\d+\.\d+)"
            package.json:"version": "(\d+\.\d+\.\d+)"
            CMakeLists.txt:VERSION (\d+\.\d+\.\d+)
          update-major-version-in: |
            include/stringzilla/stringzilla.h:^#define STRINGZILLA_VERSION_MAJOR (\d+)
          update-minor-version-in: |
            include/stringzilla/stringzilla.h:^#define STRINGZILLA_VERSION_MINOR (\d+)
          update-patch-version-in: |
            include/stringzilla/stringzilla.h:^#define STRINGZILLA_VERSION_PATCH (\d+)
          dry-run: "true"

  test_ubuntu_gcc:
    name: Ubuntu (GCC 12)
    runs-on: ubuntu-22.04
    env:
      CC: gcc-12
      CXX: g++-12

    steps:
      - uses: actions/checkout@v4

        # C/C++
        # If the compilation fails, we want to log the compilation commands in addition to
        # the standard output.
      - name: Install dependencies
        run: |
          sudo apt update
          sudo apt install -y cmake build-essential libjemalloc-dev libomp-dev gcc-12 g++-12

      - name: Build C/C++
        run: |
          cmake -B build_artifacts \
            -DCMAKE_BUILD_TYPE=RelWithDebInfo \
            -DCMAKE_EXPORT_COMPILE_COMMANDS=1 \
            -DSTRINGZILLA_BUILD_BENCHMARK=1 \
            -DSTRINGZILLA_BUILD_TEST=1

          cmake --build build_artifacts --config RelWithDebInfo > build_artifacts/logs.txt 2>&1 || {
            echo "Compilation failed. Here are the logs:"
            cat build_artifacts/logs.txt
            echo "The original compilation commands:"
            cat build_artifacts/compile_commands.json
            echo "CPU Features:"
            lscpu
            echo "GCC Version:"
            gcc-12 --version
            echo "G++ Version:"
            g++-12 --version
            exit 1
          }
      - name: Test C++
        run: build_artifacts/stringzilla_test_cpp20
      - name: Check stringzillite is built correctly
        run: test -z "$(ldd build_artifacts/libstringzillite.so | grep '.so')"
      - name: Test on Real World Data
        run: |
          build_artifacts/stringzilla_bench_search ${DATASET_PATH}     # for substring search
          build_artifacts/stringzilla_bench_token ${DATASET_PATH}      # for hashing, equality comparisons, etc.
          build_artifacts/stringzilla_bench_similarity ${DATASET_PATH} # for edit distances and alignment scores
          build_artifacts/stringzilla_bench_sort ${DATASET_PATH}       # for sorting arrays of strings
          build_artifacts/stringzilla_bench_container ${DATASET_PATH}  # for STL containers with string keys
        env:
          DATASET_PATH: ./README.md
        # Don't overload GitHub with our benchmarks.
        # The results in such an unstable environment will be meaningless anyway.
        if: 0

        # Python
      - name: Set up Python ${{ env.PYTHON_VERSION }}
        uses: actions/setup-python@v5.2.0
        with:
          python-version: ${{ env.PYTHON_VERSION }}
      - name: Build Python
        run: |
          python -m pip install --upgrade pip
          pip install pytest pytest-repeat numpy pyarrow
          python -m pip install .
      - name: Test Python
        run: pytest scripts/test.py -s -x

        # JavaScript
      # - name: Set up Node.js
      #   uses: actions/setup-node
      #   with:
      #     node-version: 18
      # - name: Build and test JavaScript
      #   run: npm ci && npm test

      # Rust
      - name: Test Rust
        uses: actions-rs/toolchain@v1
        with:
          toolchain: stable
          override: true

  test_ubuntu_clang:
    name: Ubuntu (Clang 16)
    runs-on: ubuntu-22.04
    env:
      CC: clang-16
      CXX: clang++-16

    steps:
      - uses: actions/checkout@v4

        # C/C++
        # Clang 16 isn't available from default repos on Ubuntu 22.04, so we have to install it manually
      - name: Install dependencies
        run: |
          sudo apt update
          sudo apt install -y cmake build-essential libjemalloc-dev
          wget https://apt.llvm.org/llvm.sh
          chmod +x llvm.sh
          sudo ./llvm.sh 16

      - name: Build C/C++
        run: |
          cmake -B build_artifacts \
            -DCMAKE_BUILD_TYPE=RelWithDebInfo \
            -DCMAKE_EXPORT_COMPILE_COMMANDS=1 \
            -DSTRINGZILLA_BUILD_BENCHMARK=1 \
            -DSTRINGZILLA_BUILD_TEST=1

          cmake --build build_artifacts --config RelWithDebInfo > build_artifacts/logs.txt 2>&1 || {
            echo "Compilation failed. Here are the logs:"
            cat build_artifacts/logs.txt
            echo "The original compilation commands:"
            cat build_artifacts/compile_commands.json
            echo "CPU Features:"
            lscpu
            echo "Clang Version:"
            clang-16 --version
            echo "Clang++ Version:"
            clang++-16 --version
            exit 1
          }
      - name: Test C++
        run: build_artifacts/stringzilla_test_cpp20
      - name: Check stringzillite is built correctly
        run: test -z "$(ldd build_artifacts/libstringzillite.so | grep '.so')"
      - name: Test on Real World Data
        run: |
          build_artifacts/stringzilla_bench_search ${DATASET_PATH}     # for substring search
          build_artifacts/stringzilla_bench_token ${DATASET_PATH}      # for hashing, equality comparisons, etc.
          build_artifacts/stringzilla_bench_similarity ${DATASET_PATH} # for edit distances and alignment scores
          build_artifacts/stringzilla_bench_sort ${DATASET_PATH}       # for sorting arrays of strings
          build_artifacts/stringzilla_bench_container ${DATASET_PATH}  # for STL containers with string keys
        env:
          DATASET_PATH: ./README.md
        # Don't overload GitHub with our benchmarks.
        # The results in such an unstable environment will be meaningless anyway.
        if: 0

        # Python
      - name: Set up Python ${{ env.PYTHON_VERSION }}
        uses: actions/setup-python@v5.2.0
        with:
          python-version: ${{ env.PYTHON_VERSION }}
      - name: Build Python
        run: |
          python -m pip install --upgrade pip
          pip install pytest pytest-repeat numpy pyarrow
          python -m pip install .
      - name: Test Python
        run: pytest scripts/test.py -s -x

        # Rust
      - name: Test Rust
        uses: actions-rs/toolchain@v1
        with:
          toolchain: stable
          override: true

        # Swift
        # Fails due to: https://github.com/swift-actions/setup-swift/issues/591
      # - name: Set up Swift ${{ env.SWIFT_VERSION }}
      #   uses: swift-actions/setup-swift@v1
      #   with:
      #     swift-version: ${{ env.SWIFT_VERSION }}
      # - name: Build Swift
      #   run: swift build -c release --static-swift-stdlib
      # - name: Test Swift
      #   run: swift test -c release --enable-test-discovery

  # Temporary workaround to run Swift tests on Linux
  # Based on: https://github.com/swift-actions/setup-swift/issues/591#issuecomment-1685710678
  test_ubuntu_swift:
    name: Swift on Linux
    runs-on: ubuntu-22.04
    container: swift:5.9
    steps:
      - uses: actions/checkout@v4
      - name: Test Swift
        run: swift test

  test_ubuntu_cross_compilation:
    name: Clang Cross Compilation
    runs-on: ubuntu-22.04
    env:
      CC: clang-16
      CXX: clang++-16
      AR: llvm-ar
      NM: llvm-nm
      RANLIB: llvm-ranlib

    strategy:
      fail-fast: false
      matrix:
        include:
          - arch: amd64
            target: x86_64-linux-gnu
          - arch: arm64
            target: aarch64-linux-gnu

    steps:
      - uses: actions/checkout@v4

        # C/C++
        # We need to install the cross-compilation toolchain for ARM64 and ARMHF
        # Clang 16 isn't available from default repos on Ubuntu 22.04, so we have to install it manually
      - name: Install dependencies
        run: |
          sudo apt-get update
          sudo apt-get install -y make build-essential crossbuild-essential-arm64 crossbuild-essential-armhf  libjemalloc-dev
          wget https://apt.llvm.org/llvm.sh
          chmod +x llvm.sh
          sudo ./llvm.sh 16

      - name: Build C/C++
        run: |
          cmake -B build_artifacts \
            -DCMAKE_BUILD_TYPE=RelWithDebInfo \
            -DCMAKE_EXPORT_COMPILE_COMMANDS=1 \
            -DCMAKE_C_COMPILER_TARGET=${{ matrix.target }} \
            -DCMAKE_CXX_COMPILER_TARGET=${{ matrix.target }} \
            -DCMAKE_SYSTEM_NAME=Linux \
            -DCMAKE_SYSTEM_PROCESSOR=${{ matrix.arch }} \
            -DSTRINGZILLA_BUILD_SHARED=1 \
            -DSTRINGZILLA_BUILD_BENCHMARK=1 \
            -DSTRINGZILLA_BUILD_TEST=1

          cmake --build build_artifacts --config RelWithDebInfo

        # We can't run the produced builds, but we can make sure they exist
      - name: Test artifacts presense
        run: |
          test -e build_artifacts/libstringzillite.so
          test -e build_artifacts/libstringzilla_shared.so
          test -e build_artifacts/stringzilla_test_cpp20
      
      - name: Check stringzillite is built correctly
        run: test -z "$(ldd build_artifacts/libstringzillite.so | grep '.so')"

  test_macos:
    name: MacOS
    runs-on: macos-14

    steps:
      - uses: actions/checkout@v4

        # C/C++
      - name: Install dependencies
        run: |
          brew update
          brew install cmake
      - name: Build C/C++
        run: |
          cmake -B build_artifacts \
            -DCMAKE_BUILD_TYPE=RelWithDebInfo \
            -DCMAKE_EXPORT_COMPILE_COMMANDS=1 \
            -DSTRINGZILLA_BUILD_BENCHMARK=1 \
            -DSTRINGZILLA_BUILD_TEST=1
          cmake --build build_artifacts --config RelWithDebInfo
      - name: Test C++
        run: build_artifacts/stringzilla_test_cpp17
      - name: Check stringzillite is built correctly
        run: test -z "$(ldd build_artifacts/libstringzillite.so | grep '.so')"
      - name: Test on Real World Data
        run: |
          build_artifacts/stringzilla_bench_search ${DATASET_PATH}     # for substring search
          build_artifacts/stringzilla_bench_token ${DATASET_PATH}      # for hashing, equality comparisons, etc.
          build_artifacts/stringzilla_bench_similarity ${DATASET_PATH} # for edit distances and alignment scores
          build_artifacts/stringzilla_bench_sort ${DATASET_PATH}       # for sorting arrays of strings
          build_artifacts/stringzilla_bench_container ${DATASET_PATH}  # for STL containers with string keys
        env:
          DATASET_PATH: ./README.md
        # Don't overload GitHub with our benchmarks.
        # The results in such an unstable environment will be meaningless anyway.
        if: 0

        # Python
      - name: Set up Python ${{ env.PYTHON_VERSION }}
        uses: actions/setup-python@v5.2.0
        with:
          python-version: ${{ env.PYTHON_VERSION }}
      - name: Build Python
        run: |
          python -m pip install --upgrade pip
          pip install pytest pytest-repeat numpy pyarrow
          python -m pip install .
        env:
          MACOSX_DEPLOYMENT_TARGET: "11.0"
      - name: Test Python
        run: pytest scripts/test.py -s -x

        # Swift
      - name: Set up Swift ${{ env.SWIFT_VERSION }}
        uses: swift-actions/setup-swift@v1
        with:
          swift-version: ${{ env.SWIFT_VERSION }}
      - name: Build Swift
        run: swift build
      - name: Test Swift
        run: swift test

        # Rust
      - name: Test Rust
        uses: actions-rs/toolchain@v1
        with:
          toolchain: stable
          override: true

  test_windows:
    name: Windows
    runs-on: windows-2022
    steps:
      - uses: actions/checkout@v4
      - uses: ilammy/msvc-dev-cmd@v1

      - name: Build C/C++
        shell: cmd
        run: |
          cmake -GNinja -B build_artifacts ^
            -DCMAKE_BUILD_TYPE=RelWithDebInfo ^
            -DCMAKE_EXPORT_COMPILE_COMMANDS=1 ^
            -DSTRINGZILLA_BUILD_BENCHMARK=1 ^
            -DSTRINGZILLA_BUILD_TEST=1

          cmake --build build_artifacts --config RelWithDebInfo > build_artifacts/logs.txt 2>&1 || (
            echo "Compilation failed. Here are the logs:"
            type build_artifacts\logs.txt
            echo "The original compilation commands:"
            type build_artifacts\compile_commands.json
            echo:
            echo "CPU Features:"
            wmic cpu list /format:list
            exit 1
          )
      - name: Test C++
        run: .\build_artifacts\stringzilla_test_cpp20.exe
      - name: Check stringzillite is built correctly
        shell: cmd
        run: dumpbin /nologo /dependents stringzillite.dll | findstr /r "^[^a-z].*\.dll" && exit /b 1 || exit /b 0
      - name: Test on Real World Data
        run: |
          .\build_artifacts\stringzilla_bench_search.exe ${DATASET_PATH}     # for substring search
          .\build_artifacts\stringzilla_bench_token.exe ${DATASET_PATH}      # for hashing, equality comparisons, etc.
          .\build_artifacts\stringzilla_bench_similarity.exe ${DATASET_PATH} # for edit distances and alignment scores
          .\build_artifacts\stringzilla_bench_sort.exe ${DATASET_PATH}       # for sorting arrays of strings
          .\build_artifacts\stringzilla_bench_container.exe ${DATASET_PATH}  # for STL containers with string keys
        env:
          DATASET_PATH: ./README.md
        # Don't overload GitHub with our benchmarks.
        # The results in such an unstable environment will be meaningless anyway.
        if: 0

        # Python
      - name: Set up Python ${{ env.PYTHON_VERSION }}
        uses: actions/setup-python@v5.2.0
        with:
          python-version: ${{ env.PYTHON_VERSION }}
      - name: Build Python
        run: |
          python -m pip install --upgrade pip
          pip install pytest pytest-repeat numpy pyarrow
          python -m pip install .
      - name: Test Python
        run: pytest scripts/test.py -s -x

  test_windows_cross_compilation:
    name: Windows Cross Compilation ${{ matrix.target }}
    runs-on: windows-latest
    
    strategy:
      fail-fast: false
      matrix:
        include:
          - arch: x64
            target: x64
          - arch: x86
            target: x86
          - arch: amd64_arm64
            target: arm64
            
    steps:
      - uses: actions/checkout@v4
      - uses: ilammy/msvc-dev-cmd@v1
        with:
          arch: ${{ matrix.arch }}
      
      - name: Build C/C++
        shell: cmd
        run: |
          cmake -GNinja -B build_artifacts ^
            -DCMAKE_BUILD_TYPE=RelWithDebInfo ^
            -DCMAKE_EXPORT_COMPILE_COMMANDS=1 ^
            -DSTRINGZILLA_BUILD_SHARED=1 ^
            -DSTRINGZILLA_BUILD_BENCHMARK=1 ^
            -DSTRINGZILLA_BUILD_TEST=1

          cmake --build build_artifacts --config RelWithDebInfo
        
        # We can't run the produced builds, but we can make sure they exist
      - name: Test artifacts presense
        shell: cmd
        run: |
          type build_artifacts\stringzillite.dll >NUL
          type build_artifacts\stringzilla_shared.dll >NUL
          type build_artifacts\stringzilla_test_cpp20.exe >NUL
      
      - name: Check stringzillite is built correctly
        shell: cmd
        run: dumpbin /nologo /dependents stringzillite.dll | findstr /r "^[^a-z].*\.dll" && exit /b 1 || exit /b 0
  
  test_alpine:
    name: Alpine Linux
    runs-on: ubuntu-24.04
    container:
      image: alpine:latest
      options: --privileged # If needed for certain Docker operations
    steps:
      - uses: actions/checkout@v3

        # C/C++
      - name: Build C/C++
        run: |
          apk add --update make cmake g++ gcc
          cmake -B build_artifacts \
            -DCMAKE_BUILD_TYPE=RelWithDebInfo \
            -DCMAKE_EXPORT_COMPILE_COMMANDS=1 \
            -DSTRINGZILLA_BUILD_BENCHMARK=1 \
            -DSTRINGZILLA_BUILD_TEST=1
          cmake --build build_artifacts --config RelWithDebInfo
      - name: Test C++
        run: build_artifacts/stringzilla_test_cpp20
      - name: Check stringzillite is built correctly
        run: test -z "$(ldd build_artifacts/libstringzillite.so | grep '.so')"

        # Python
      - name: Build Python
        run: |
          apk add python3 py3-pip python3-dev
          pip install --break-system-packages pytest pytest-repeat
          pip install --break-system-packages .
      - name: Test Python
        run: pytest scripts/test.py -s -x

        # Rust
      - name: Test Rust
        uses: actions-rs/toolchain@v1
        with:
          toolchain: stable
          override: true

  build_wheels:
    name: Build Python ${{ matrix.python-version }} for ${{ matrix.os }}
    runs-on: ${{ matrix.os }}
    needs:
      [
        test_ubuntu_gcc,
        test_ubuntu_clang,
        test_macos,
        test_windows,
        test_alpine,
      ]
    strategy:
      matrix:
        os: [ubuntu-24.04, macos-14, windows-2022]
        python-version: ["36", "37", "38", "39", "310", "311", "312"]
    steps:
      - uses: actions/checkout@v4
      - name: Set up Python
        uses: actions/setup-python@v5.2.0
        with:
          python-version: 3.x

        # We only need QEMU for Linux builds
      - name: Setup QEMU
        if: matrix.os == 'ubuntu-24.04'
        uses: docker/setup-qemu-action@v3
      - name: Install cibuildwheel
        run: python -m pip install cibuildwheel
      - name: Build wheels
        run: cibuildwheel --output-dir wheelhouse
        env:
          CIBW_BUILD: cp${{ matrix.python-version }}-*