ci-workflow.yml

name: CI Workflow

on:
  pull_request:
    branches:
      - master
  push:
    branches:
      - master
  workflow_dispatch:
    inputs:
      auto_checkpoint:
        description: 'Database config - auto checkpoint'
        required: false
        default: true
        type: boolean

      buffer_pool_size:
        description: 'Database config - buffer pool size'
        required: false
        default: 67108864
        type: number

      max_num_threads:
        description: 'Database config - max number of threads'
        required: false
        default: 2
        type: number

      enable_compression:
        description: 'Database config - enable compression'
        required: false
        default: true
        type: boolean

      checkpoint_threshold:
        description: 'Database config - checkpoint threshold'
        required: false
        default: 16777216
        type: number
env:
  RUNTIME_CHECKS: 1
  USE_EXISTING_BINARY_DATASET: 1
  AUTO_CHECKPOINT: ${{ github.event.inputs.auto_checkpoint }}
  BUFFER_POOL_SIZE: ${{ github.event.inputs.buffer_pool_size }}
  MAX_NUM_THREADS: ${{ github.event.inputs.max_num_threads }}
  ENABLE_COMPRESSION: ${{ github.event.inputs.enable_compression }}
  CHECKPOINT_THRESHOLD: ${{ github.event.inputs.checkpoint_threshold }}
  WERROR: 1
  RUSTFLAGS: --deny warnings
  PIP_BREAK_SYSTEM_PACKAGES: 1

# Only allow one run in this group to run at a time, and cancel any runs in progress in this group.
# We use the workflow name and then add the pull request number, or (if it's a push to master), we use the name of the branch.
# See github's docs[1] and a relevant stack overflow answer[2]
#   [1]: https://docs.github.com/en/actions/using-jobs/using-concurrency
#   [2]: https://stackoverflow.com/questions/66335225/how-to-cancel-previous-runs-in-the-pr-when-you-push-new-commitsupdate-the-curre
concurrency:
  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
  cancel-in-progress: true

jobs:
  generate-binary-datasets:
    name: generate binary dataset
    needs: [sanity-checks, python-lint-check]
    runs-on: kuzu-self-hosted-testing
    env:
      NUM_THREADS: 32
      GEN: Ninja
      CC: gcc
      CXX: g++
    steps:
      - uses: actions/checkout@v4

      - name: Build
        run: make release

      - name: Generate datasets
        run: bash scripts/generate_binary_demo.sh

      - name: Generate and upload tinysnb
        run: |
          bash scripts/generate_binary_tinysnb.sh
          s3cmd get s3://kuzu-test/tinysnb/version.txt || echo "0" > version.txt
          if [ "$(cat tinysnb/version.txt)" == "$(cat version.txt)" ]; then
            echo "TinySNB dataset is up to date, skipping upload"
            rm -rf tinysnb version.txt
            exit 0
          fi
          echo "TinySNB dataset is outdated, uploading..."
          s3cmd del -r s3://kuzu-test/tinysnb/
          s3cmd sync ./tinysnb s3://kuzu-test/
          rm -rf tinysnb version.txt

      - name: Generate and upload ldbc-sf01
        run: |
          bash scripts/generate_binary_ldbc-sf01.sh
          s3cmd get s3://kuzu-test/ldbc01/version.txt || echo "0" > version.txt
          if [ "$(cat ldbc01/version.txt)" == "$(cat version.txt)" ]; then
            echo "LDBC-SF01 dataset is up to date, skipping upload"
            rm -rf ldbc01 version.txt
            exit 0
          fi
          echo "LDBC-SF01 dataset is outdated, uploading..."
          s3cmd del -r s3://kuzu-test/ldbc01/
          s3cmd sync ./ldbc01 s3://kuzu-test/
          rm -rf ldbc01 version.txt

      - name: Upload binary-demo
        uses: actions/upload-artifact@v4
        with:
          name: binary-demo
          path: dataset/binary-demo

  gcc-build-test:
    name: gcc build & test
    needs: [sanity-checks, python-lint-check, generate-binary-datasets]
    runs-on: kuzu-self-hosted-testing
    env:
      NUM_THREADS: 32
      CMAKE_BUILD_PARALLEL_LEVEL: 32
      TEST_JOBS: 16
      CLANGD_DIAGNOSTIC_JOBS: 32
      CLANGD_DIAGNOSTIC_INSTANCES: 6
      GEN: Ninja
      CC: gcc
      CXX: g++
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      RUN_ID: "$(hostname)-$(date +%s)"
      HTTP_CACHE_FILE: TRUE

    steps:
      - uses: actions/checkout@v4

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: ${{ github.workspace }}/dataset/binary-demo

      - name: Ensure Python dependencies
        run: |
          pip install torch~=2.2.0 --extra-index-url https://download.pytorch.org/whl/cpu
          pip install --user -r tools/python_api/requirements_dev.txt -f https://data.pyg.org/whl/torch-2.2.0+cpu.html

      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Test
        run: |
          ln -s /home/runner/ldbc-1-csv dataset/ldbc-1/csv
          make test

      - name: Test in mem
        env:
          IN_MEM_MODE: true
        run: |
          make test

      - name: Test checkpoint at threshold 0
        env:
          AUTO_CHECKPOINT: true
          CHECKPOINT_THRESHOLD: 0
        run: |
          make test

      - name: Python test
        run: make pytest

      - name: Node.js test
        run: make nodejstest

      - name: Java test
        run: make javatest

      - name: Rust test with pre-built library
        env:
          KUZU_SHARED: 1
          KUZU_INCLUDE_DIR: ${{ github.workspace }}/build/release/src
          KUZU_LIBRARY_DIR: ${{ github.workspace }}/build/release/src
        run: make rusttest

  gcc-build-test-in-mem-vector-size:
    name: gcc build & test in-mem only with various vector size
    needs: [gcc-build-test]
    runs-on: kuzu-self-hosted-testing
    env:
      NUM_THREADS: 32
      CMAKE_BUILD_PARALLEL_LEVEL: 32
      TEST_JOBS: 16
      CLANGD_DIAGNOSTIC_JOBS: 32
      CLANGD_DIAGNOSTIC_INSTANCES: 6
      GEN: Ninja
      CC: gcc
      CXX: g++
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      RUN_ID: "$(hostname)-$(date +%s)"
      HTTP_CACHE_FILE: TRUE

    steps:
      - uses: actions/checkout@v4

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: ${{ github.workspace }}/dataset/binary-demo

      - name: Ensure Python dependencies
        run: |
          pip install torch~=2.2.0 --extra-index-url https://download.pytorch.org/whl/cpu
          pip install --user -r tools/python_api/requirements_dev.txt -f https://data.pyg.org/whl/torch-2.2.0+cpu.html

      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Test vector 1024
        env:
          IN_MEM_MODE: true
        run: |
          ln -s /home/runner/ldbc-1-csv dataset/ldbc-1/csv
          make test VECTOR_CAPACITY_LOG2=10

  code-coverage:
    name: code coverage
    runs-on: ubuntu-22.04
    needs: [sanity-checks, python-lint-check, generate-binary-datasets]
    env:
      TEST_JOBS: 10
      CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
    steps:
      - uses: actions/checkout@v4

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: dataset/binary-demo

      - name: Install lcov
        run: sudo apt-get install -y lcov

      - name: Test with coverage
        run: make lcov NUM_THREADS=$(nproc)

      - name: Test with coverage in mem
        env:
          IN_MEM_MODE: true
        run: make lcov NUM_THREADS=$(nproc)

      - name: Generate coverage report
        run: |
          lcov --config-file .lcovrc -c -d ./ --no-external -o cover.info &&\
          lcov --remove cover.info $(< .github/workflows/lcov_exclude) -o cover.info

      - name: Upload coverage report
        uses: codecov/codecov-action@v4.5.0
        with:
          file: cover.info

  webassembly-build-test:
    name: webassembly build & test
    needs: [sanity-checks, generate-binary-datasets]
    runs-on: kuzu-self-hosted-testing
    env:
      WERROR: 0
      TEST_JOBS: 8
      NUM_THREADS: 32

    steps:
      - uses: actions/checkout@v4

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: dataset/binary-demo

      - name: Build & test
        run: |
          ln -s /home/runner/ldbc-1-csv dataset/ldbc-1/csv
          source /home/runner/emsdk/emsdk_env.sh
          make wasmtest

      - name: Build & test (single-thread mode)
        env:
          SINGLE_THREADED: true
        run: |
          make clean
          source /home/runner/emsdk/emsdk_env.sh
          make wasmtest

  rust-build-test:
    name: rust build & test
    needs: [rustfmt-check]
    runs-on: kuzu-self-hosted-testing
    env:
      CARGO_BUILD_JOBS: 32
      CMAKE_BUILD_PARALLEL_LEVEL: 32
      CC: gcc
      CXX: g++
    steps:
      - uses: actions/checkout@v4

      - name: Rust test
        run: make rusttest

      - name: Rust example
        working-directory: examples/rust
        run: cargo build --locked --features arrow

  gcc-build-test-with-asan:
    name: gcc build & test with asan
    needs: [gcc-build-test, generate-binary-datasets]
    runs-on: kuzu-self-hosted-testing
    env:
      NUM_THREADS: 32
      TEST_JOBS: 16
      GEN: Ninja
      CC: gcc
      CXX: g++
      WERROR: 0
    steps:
      - uses: actions/checkout@v4

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: dataset/binary-demo

      - name: Ensure Python dependencies
        run: |
          pip install torch~=2.2.0 --extra-index-url https://download.pytorch.org/whl/cpu
          pip install --user -r tools/python_api/requirements_dev.txt -f https://data.pyg.org/whl/torch-2.2.0+cpu.html

      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Test with ASAN
        run: make test ASAN=1
        env:
          ASAN_OPTIONS: "detect_leaks=1"

      - name: Test compression disabled with ASAN
        env:
          ENABLE_COMPRESSION: false
        run: |
          make test

  clang-build-test:
    name: clang build & test
    needs: [sanity-checks, python-lint-check, generate-binary-datasets]
    runs-on: kuzu-self-hosted-testing
    env:
      NUM_THREADS: 32
      CMAKE_BUILD_PARALLEL_LEVEL: 32
      TEST_JOBS: 16
      CC: clang
      CXX: clang++
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      RUN_ID: "$(hostname)-$(date +%s)"
      HTTP_CACHE_FILE: TRUE
    steps:
      - uses: actions/checkout@v4

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: dataset/binary-demo

      - name: Ensure Python dependencies
        run: |
          pip install torch~=2.2.0 --extra-index-url https://download.pytorch.org/whl/cpu
          pip install --user -r tools/python_api/requirements_dev.txt -f https://data.pyg.org/whl/torch-2.2.0+cpu.html

      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Test
        run: |
          ln -s /home/runner/ldbc-1-csv dataset/ldbc-1/csv
          make test

      - name: Test compression disabled
        env:
          ENABLE_COMPRESSION: false
        run: |
          make test

      - name: Python test
        run: make pytest

      - name: Node.js test
        run: make nodejstest

      - name: Java test
        run: make javatest

      - name: Rust test with pre-built library
        env:
          KUZU_SHARED: 1
          KUZU_INCLUDE_DIR: ${{ github.workspace }}/build/release/src
          KUZU_LIBRARY_DIR: ${{ github.workspace }}/build/release/src
        run: make rusttest

  clang-build-test-various-page-sizes:
    strategy:
      matrix:
        page_size_log2: [16, 18]
    name: clang build & test with page_size_log2=${{ matrix.page_size_log2 }}
    needs: [clang-build-test]
    runs-on: kuzu-self-hosted-testing
    env:
      NUM_THREADS: 32
      CMAKE_BUILD_PARALLEL_LEVEL: 32
      TEST_JOBS: 16
      CC: clang
      CXX: clang++
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      RUN_ID: "$(hostname)-$(date +%s)"
      HTTP_CACHE_FILE: TRUE
    steps:
      - uses: actions/checkout@v4

      - name: Ensure Python dependencies
        run: |
          pip install torch~=2.2.0 --extra-index-url https://download.pytorch.org/whl/cpu
          pip install --user -r tools/python_api/requirements_dev.txt -f https://data.pyg.org/whl/torch-2.2.0+cpu.html

      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Build
        run: |
          ln -s /home/runner/ldbc-1-csv dataset/ldbc-1/csv
          make release PAGE_SIZE_LOG2=${{ matrix.page_size_log2 }}

      - name: Generate binary demo
        run: |
          bash scripts/generate_binary_demo.sh

      - name: Test
        run: |
          make test PAGE_SIZE_LOG2=${{ matrix.page_size_log2 }}

  msvc-build-test:
    name: msvc build & test
    needs: [sanity-checks, python-lint-check, generate-binary-datasets]
    runs-on: self-hosted-windows
    env:
      # Shorten build path as much as possible
      CARGO_TARGET_DIR: ${{ github.workspace }}/rs
      CARGO_BUILD_JOBS: 18
      NUM_THREADS: 18
      CMAKE_BUILD_PARALLEL_LEVEL: 18
      TEST_JOBS: 9
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      PG_HOST: ${{ secrets.PG_HOST }}
      RUN_ID: "$(hostname)-$([Math]::Floor((Get-Date).TimeOfDay.TotalSeconds))"
      HTTP_CACHE_FILE: TRUE
      WERROR: 0
    steps:
      - uses: actions/checkout@v4

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: dataset/binary-demo

      - name: Ensure Python dependencies
        run: |
          pip install torch~=2.0.0 --extra-index-url https://download.pytorch.org/whl/cpu
          pip install --user -r tools/python_api/requirements_dev.txt -f https://data.pyg.org/whl/torch-2.0.0+cpu.html

      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Visual Studio Generator Build
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make all GEN="Visual Studio 17 2022"
          make clean

      - name: Test
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make test

      - name: Test compression disabled
        shell: cmd
        env:
          ENABLE_COMPRESSION: false
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make test

      - name: Test in mem
        shell: cmd
        env:
          IN_MEM_MODE: true
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make test

      - name: Python test
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make pytest

      - name: Node.js test
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make nodejstest

      - name: Rust test
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make rusttest

      - name: Java test
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make javatest

      - name: Rust test with pre-built library
        env:
          KUZU_SHARED: 1
          KUZU_INCLUDE_DIR: ${{ github.workspace }}/build/release/src
          KUZU_LIBRARY_DIR: ${{ github.workspace }}/build/release/src
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          set PATH=%PATH%;${{ github.workspace }}/build/release/src
          make rusttest

  sanity-checks:
    name: sanity checks
    runs-on: ubuntu-22.04
    steps:
      - uses: actions/checkout@v4

      - name: Check headers for include guards
        run: ./scripts/check-include-guards.sh src/include

      - name: Checks files for std::assert
        run: ./scripts/check-no-std-assert.sh src

      - name: Ensure generated grammar files are up to date
        run: |
          python3 scripts/antlr4/hash.py src/antlr4/keywords.txt src/antlr4/Cypher.g4 > tmphashfile
          cmp tmphashfile scripts/antlr4/hash.md5
          rm tmphashfile

  clang-format:
    name: clang format
    runs-on: ubuntu-24.04
    steps:
      - name: Install clang-format
        run: |
          sudo apt-get update
          sudo apt-get install -y clang-format-18

      - uses: actions/checkout@v4
        with:
          repository: ${{ github.event.pull_request.head.repo.full_name }}
          ref: ${{ github.event.pull_request.head.ref }}

      - name: Check and fix source format
        run: python3 scripts/run-clang-format.py --in-place --clang-format-executable /usr/bin/clang-format-18 -r src/

      - name: Check and fix test format
        run: python3 scripts/run-clang-format.py --in-place --clang-format-executable /usr/bin/clang-format-18 -r test/

      - name: Check and fix tools format
        run: python3 scripts/run-clang-format.py --in-place --clang-format-executable /usr/bin/clang-format-18 -r tools/

      - name: Check and fix extension format
        run: python3 scripts/run-clang-format.py --in-place --clang-format-executable /usr/bin/clang-format-18 -r extension/

      - name: Fail if any chang is made (master branch)
        if: github.ref == 'refs/heads/master'
        run: git diff --exit-code

      - name: Commit changes (non-master branch)
        uses: EndBug/add-and-commit@v9
        if: github.ref != 'refs/heads/master'
        with:
          author_name: "CI Bot"
          message: "Run clang-format"

  python-lint-check:
    name: python lint check
    runs-on: macos-14
    steps:
      - uses: actions/checkout@v4

      - name: Setup Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.11"

      - name: Run Python lint
        working-directory: tools/python_api
        run: |
          make requirements
          ./.venv/bin/ruff check src_py test --verbose

  rustfmt-check:
    name: rustfmt check
    runs-on: ubuntu-22.04
    steps:
      - name: Update Rust
        run: rustup update

      - uses: actions/checkout@v4

      - name: Check Rust API format
        working-directory: tools/rust_api
        run: cargo fmt --check

  benchmark:
    name: benchmark
    needs: [gcc-build-test, clang-build-test]
    uses: ./.github/workflows/benchmark-workflow.yml

  report-benchmark-result:
    name: report benchmark result
    needs: [benchmark]
    runs-on: ubuntu-22.04
    if: github.event_name == 'pull_request'
    steps:
      - name: Download comparison results
        uses: actions/download-artifact@v4
        with:
          name: comparison-results

      - name: Report benchmark result
        uses: thollander/actions-comment-pull-request@v2
        with:
          filePath: compare_result.md

  clang-tidy:
    name: clang tidy & clangd diagnostics check
    needs: [sanity-checks]
    runs-on: kuzu-self-hosted-testing
    env:
      GEN: Ninja
      NUM_THREADS: 32
      CC: clang
      CXX: clang++

    steps:
      - uses: actions/checkout@v4

      # For `napi.h` header.
      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Check for clangd diagnostics
        run: make clangd-diagnostics

      - name: Run clang-tidy
        run: make tidy

      - name: Run clang-tidy analyzer
        run: make tidy-analyzer

  macos-build-test:
    name: apple clang build & test
    needs: [sanity-checks, rustfmt-check, python-lint-check, generate-binary-datasets]
    runs-on: [self-hosted, macOS]
    env:
      TEST_JOBS: 16
      GEN: Ninja
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      PG_HOST: ${{ secrets.PG_HOST }}
      RUN_ID: "$(hostname)-$(date +%s)"
      HTTP_CACHE_FILE: TRUE
    steps:
      - uses: actions/checkout@v4

      - name: Determine NUM_THREADS
        run: |
          export NUM_THREADS=$(($(sysctl -n hw.logicalcpu) * 2 / 3))
          if [ $NUM_THREADS -lt 12 ]; then
            export NUM_THREADS=12
          fi
          echo "NUM_THREADS=$NUM_THREADS" >> $GITHUB_ENV
          echo "CMAKE_BUILD_PARALLEL_LEVEL=$NUM_THREADS" >> $GITHUB_ENV
          echo "NUM_THREADS=$NUM_THREADS"

      - name: Download binary-demo
        uses: actions/download-artifact@v4
        with:
          name: binary-demo
          path: dataset/binary-demo

      - name: Ensure Python dependencies
        run: |
          pip3 install torch~=2.2.0 --extra-index-url https://download.pytorch.org/whl/cpu
          pip3 install --user -r tools/python_api/requirements_dev.txt -f https://data.pyg.org/whl/torch-2.2.0+cpu.html

      - name: Ensure Node.js dependencies
        run: npm install --include=dev
        working-directory: tools/nodejs_api

      - name: Test
        run: |
          ulimit -n 10240
          ln -s /Users/runner/ldbc-1-csv dataset/ldbc-1/csv
          make test

      - name: Test in mem
        env:
          IN_MEM_MODE: true
        run: |
          make test

      - name: Test checkpoint at threshold 0
        env:
          AUTO_CHECKPOINT: true
          CHECKPOINT_THRESHOLD: 0
        run: |
          make test

      - name: Python test
        env:
          PYBIND11_PYTHON_VERSION: 3.12
        run: |
          ulimit -n 10240
          make pytest

      - name: C and C++ Examples
        run: |
          ulimit -n 10240
          make example

      - name: Node.js test
        run: |
          ulimit -n 10240
          make nodejstest

      - name: Java test
        run: |
          ulimit -n 10240
          export JAVA_HOME=`/usr/libexec/java_home`
          make javatest

      - name: Rust test
        run: |
          ulimit -n 10240
          source /Users/runner/.cargo/env
          make rusttest

      - name: Rust example
        working-directory: examples/rust
        run: |
          ulimit -n 10240
          source /Users/runner/.cargo/env
          cargo build --locked --features arrow

      # TODO(bmwinger): rust pre-built library test for macos
      # When last tried it had issues finding libkuzu.dylib when running the tests
      # Using LD_LIBRARY_PATH (and similar variables) did not work for some reason

  shell-test:
    name: shell test
    runs-on: ubuntu-latest
    needs: [sanity-checks]
    steps:
      - uses: actions/checkout@v4

      - name: Build
        run: make release NUM_THREADS=$(nproc)

      - name: Test
        working-directory: tools/shell/test
        run: |
          pip3 install pytest pexpect
          python3 -m pytest -v .

  linux-extension-test:
    name: linux extension test
    needs: [gcc-build-test, clang-build-test]
    runs-on: kuzu-self-hosted-testing
    env:
      NUM_THREADS: 32
      TEST_JOBS: 16
      CLANGD_DIAGNOSTIC_JOBS: 32
      CLANGD_DIAGNOSTIC_INSTANCES: 6
      GEN: Ninja
      CC: gcc
      CXX: g++
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      RUN_ID: "$(hostname)-$(date +%s)"
      HTTP_CACHE_FILE: TRUE
      ASAN_OPTIONS: detect_leaks=1

    steps:
      - uses: actions/checkout@v4

      - name: Update PostgreSQL host
        working-directory: extension/postgres/test/test_files
        env:
          FNAME: postgres.test
          FIND: "localhost"
        run: |
          node -e 'fs=require("fs");fs.readFile(process.env.FNAME,"utf8",(err,data)=>{if(err!=null)throw err;fs.writeFile(process.env.FNAME,data.replaceAll(process.env.FIND,process.env.PG_HOST),"utf8",e=>{if(e!=null)throw e;});});'
          cat postgres.test

      - name: Install dependencies
        run: pip install rangehttpserver

      # shell needs to be built first to generate the dataset provided by the server
      - name: Extension test build
        run: make extension-test-build

      - name: Extension test
        run: |
          python3 scripts/generate-tinysnb.py
          cd scripts/ && python3 http-server.py &
          make extension-test

      - name: Extension test in mem
        env:
          IN_MEM_MODE: true
          HTTP_CACHE_FILE: false
        run: |
          make extension-test && make clean

      # TODO(Royi) there is currently a known memory leak issue in the DuckDB postgres/delta/iceberg extension so we avoid running the postgres/delta extension tests with Leak Sanitizer
      - name: Extension test with asan
        run: |
          make clean && make extension-test ASAN=1 WERROR=0 EXTENSION_TEST_EXCLUDE_FILTER="(delta|postgres|iceberg)" && make clean

  macos-extension-test:
    name: macos extension test
    needs: [macos-build-test]
    runs-on: [self-hosted, macOS]
    env:
      TEST_JOBS: 16
      GEN: Ninja
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      PG_HOST: ${{ secrets.PG_HOST }}
      RUN_ID: "$(hostname)-$(date +%s)"
      HTTP_CACHE_FILE: TRUE
    steps:
      - uses: actions/checkout@v4

      - name: Determine NUM_THREADS
        run: |
          export NUM_THREADS=$(($(sysctl -n hw.logicalcpu) * 2 / 3))
          if [ $NUM_THREADS -lt 12 ]; then
            export NUM_THREADS=12
          fi
          echo "NUM_THREADS=$NUM_THREADS" >> $GITHUB_ENV
          echo "NUM_THREADS=$NUM_THREADS"

      - name: Update PostgreSQL host
        working-directory: extension/postgres/test/test_files
        env:
          FNAME: postgres.test
          FIND: "localhost"
        run: |
          node -e 'fs=require("fs");fs.readFile(process.env.FNAME,"utf8",(err,data)=>{if(err!=null)throw err;fs.writeFile(process.env.FNAME,data.replaceAll(process.env.FIND,process.env.PG_HOST),"utf8",e=>{if(e!=null)throw e;});});'
          cat postgres.test

      - name: Install dependencies
        run: pip3 install rangehttpserver

      # shell needs to be built first to generate the dataset provided by the server
      - name: Extension test build
        run: make extension-test-build

      - name: Extension test
        run: |
          python3 scripts/generate-tinysnb.py
          cd scripts/ && python3 http-server.py &
          make extension-test

      - name: Extension test in mem
        env:
          IN_MEM_MODE: true
          HTTP_CACHE_FILE: false
        run: |
          make extension-test && make clean

  windows-extension-test:
    name: windows extension test
    needs: [msvc-build-test]
    runs-on: self-hosted-windows
    env:
      # Shorten build path as much as possible
      CARGO_TARGET_DIR: ${{ github.workspace }}/rs
      CARGO_BUILD_JOBS: 18
      NUM_THREADS: 18
      TEST_JOBS: 9
      UW_S3_ACCESS_KEY_ID: ${{ secrets.UW_S3_ACCESS_KEY_ID }}
      UW_S3_SECRET_ACCESS_KEY: ${{ secrets.UW_S3_SECRET_ACCESS_KEY }}
      AWS_S3_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_S3_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_S3_ACCESS_KEY_ID }}
      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_S3_SECRET_ACCESS_KEY }}
      PG_HOST: ${{ secrets.PG_HOST }}
      RUN_ID: "$(hostname)-$([Math]::Floor((Get-Date).TimeOfDay.TotalSeconds))"
      HTTP_CACHE_FILE: TRUE
      WERROR: 0
    steps:
      - uses: actions/checkout@v4

      - name: Install dependencies
        run: pip install rangehttpserver

      - name: Update PostgreSQL host
        working-directory: extension/postgres/test/test_files
        env:
          FNAME: postgres.test
          FIND: "localhost"
        run: |
          $fname = $env:FNAME
          $find = $env:FIND
          $replace = $env:PG_HOST
          $content = Get-Content -Path $fname
          $content = $content -replace [regex]::Escape($find), $replace
          Set-Content -Path $fname -Value $content
          cat $fname

      - name: Extension test build
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make extension-test-build

      - name: Extension test
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          python3 scripts/generate-tinysnb.py
          if %errorlevel% neq 0 exit /b %errorlevel%
          cd scripts/ && start /b python http-server.py && cd ..
          make extension-test

      - name: Extension test in mem
        shell: cmd
        env:
          IN_MEM_MODE: true
          HTTP_CACHE_FILE: false
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          python3 scripts/generate-tinysnb.py
          if %errorlevel% neq 0 exit /b %errorlevel%
          cd scripts/ && start /b python http-server.py && cd ..
          make extension-test

      - name: Clean
        shell: cmd
        run: |
          call "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvars64.bat"
          make clean