Regression.yml

name: Regression
on:
  workflow_dispatch:
  repository_dispatch:
  push:
    branches:
      - '**'
      - '!main'
      - '!feature'
    paths-ignore:
      - '**.md'
      - 'tools/**'
      - '!tools/pythonpkg/**'
      - '.github/patches/duckdb-wasm/**'
      - '.github/workflows/**'
      - '!.github/workflows/Regression.yml'

  pull_request:
    types: [opened, reopened, ready_for_review]
    paths-ignore:
      - '**.md'
      - 'tools/**'
      - '!tools/pythonpkg/**'
      - '.github/patches/duckdb-wasm/**'
      - '.github/workflows/**'
      - '!.github/workflows/Regression.yml'


concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}-${{ github.head_ref || '' }}-${{ github.base_ref || '' }}-${{ github.ref != 'refs/heads/main' || github.sha }}
  cancel-in-progress: true

env:
  GH_TOKEN: ${{ secrets.GH_TOKEN }}
  BASE_BRANCH: ${{ github.base_ref || (endsWith(github.ref, '_feature') && 'feature' || 'main') }}

jobs:
 regression-test-benchmark-runner:
  name: Regression Tests
  runs-on: ubuntu-20.04
  env:
    CC: gcc-10
    CXX: g++-10
    GEN: ninja
    BUILD_BENCHMARK: 1
    BUILD_TPCH: 1
    BUILD_TPCDS: 1
    BUILD_HTTPFS: 1
    BUILD_JEMALLOC: 1
    CORE_EXTENSIONS: "inet"

  steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - uses: actions/setup-python@v5
      with:
        python-version: '3.12'

    - name: Install
      shell: bash
      run: sudo apt-get update -y -qq && sudo apt-get install -y -qq ninja-build && pip install requests

    - name: Setup Ccache
      uses: hendrikmuhs/ccache-action@main
      with:
        key: ${{ github.job }}
        save: ${{ github.ref == 'refs/heads/main' || github.repository != 'duckdb/duckdb' }}

    - name: Build
      shell: bash
      run: |
        make
        git clone --branch ${{ env.BASE_BRANCH }} https://github.com/duckdb/duckdb.git --depth=1
        cd duckdb
        make
        cd ..

    - name: Set up benchmarks
      shell: bash
      run: |
        cp -r benchmark duckdb/

    - name: Regression Test Micro
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/micro.csv --verbose --threads=2

    - name: Regression Test Ingestion Perf
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/ingestion.csv --verbose --threads=2

    - name: Regression Test TPCH
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/tpch.csv --verbose --threads=2

    - name: Regression Test TPCH-PARQUET
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/tpch_parquet.csv --verbose --threads=2

    - name: Regression Test TPCDS
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/tpcds.csv --verbose --threads=2

    - name: Regression Test H2OAI
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/h2oai.csv --verbose --threads=2

    - name: Regression Test IMDB
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/imdb.csv --verbose --threads=2

    - name: Regression Test CSV
      if: always()
      shell: bash
      run: |
        python scripts/regression_test_runner.py --old=duckdb/build/release/benchmark/benchmark_runner --new=build/release/benchmark/benchmark_runner --benchmarks=.github/regression/csv.csv --verbose --threads=2

 regression-test-storage:
  name: Storage Size Regression Test
  runs-on: ubuntu-20.04
  env:
    CC: gcc-10
    CXX: g++-10
    GEN: ninja
    BUILD_TPCH: 1
    BUILD_TPCDS: 1

  steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - uses: actions/setup-python@v5
      with:
        python-version: '3.12'

    - name: Install
      shell: bash
      run: sudo apt-get update -y -qq && sudo apt-get install -y -qq ninja-build && pip install requests

    - name: Setup Ccache
      uses: hendrikmuhs/ccache-action@main
      with:
        key: ${{ github.job }}
        save: ${{ github.ref == 'refs/heads/main' || github.repository != 'duckdb/duckdb' }}

    - name: Build
      shell: bash
      run: |
        make
        git clone --branch ${{ env.BASE_BRANCH }} https://github.com/duckdb/duckdb.git --depth=1
        cd duckdb
        make
        cd ..

    - name: Regression Test
      shell: bash
      run: |
        python scripts/regression_test_storage_size.py --old=duckdb/build/release/duckdb --new=build/release/duckdb

    - name: Test for incompatibility
      shell: bash
      run: |
        if (cmp test/sql/storage_version/storage_version.db duckdb/test/sql/storage_version/storage_version.db); then
          echo "storage_changed=false" >> $GITHUB_ENV
        else
          echo "storage_changed=true" >> $GITHUB_ENV
        fi

    - name: Regression Compatibility Test (testing bidirectional compatibility)
      shell: bash
      if: env.storage_changed == 'false'
      run: |
        # Regenerate test/sql/storage_version.db with newer version -> read with older version
        python3 scripts/generate_storage_version.py
        ./duckdb/build/release/duckdb test/sql/storage_version/storage_version.db
        # Regenerate test/sql/storage_version.db with older version -> read with newer version (already performed as part of test.slow)
        cd duckdb
        python3 ../scripts/generate_storage_version.py
        ../build/release/duckdb duckdb/test/sql/storage_version/storage_version.db
        cd ..

    - name: Regression Compatibility Test (testing storage version has been bumped)
      shell: bash
      if: env.storage_changed == 'true'
      run: |
        python3 scripts/generate_storage_version.py
        cd duckdb
        python3 scripts/generate_storage_version.py
        cd ..
        if (cmp -i 8 -n 12 test/sql/storage_version.db duckdb/test/sql/storage_version.db); then
           echo "Expected storage format to be bumped, but this is not the case"
           echo "This might fail spuriously if changes to content of test database / generation script happened"
           exit 1
        else
           echo "Storage bump detected, all good!"
        fi

 regression-test-binary-size:
  name: Regression test binary size
  runs-on: ubuntu-20.04
  env:
    CC: gcc-10
    CXX: g++-10
    GEN: ninja
    BUILD_TPCH: 1
    BUILD_TPCDS: 1
    BUILD_JSON: 1
    BUILD_PARQUET: 1
    EXTENSION_STATIC_BUILD: 1
  steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - uses: actions/setup-python@v5
      with:
        python-version: '3.12'

    - name: Install
      shell: bash
      run: sudo apt-get update -y -qq && sudo apt-get install -y -qq ninja-build && pip install requests

    - name: Setup Ccache
      uses: hendrikmuhs/ccache-action@main
      with:
        key: ${{ github.job }}
        save: ${{ github.ref == 'refs/heads/main' || github.repository != 'duckdb/duckdb' }}

    - name: Build
      shell: bash
      run: |
        make
        git clone --branch ${{ env.BASE_BRANCH }} https://github.com/duckdb/duckdb.git --depth=1
        cd duckdb
        make
        cd ..

    - name: Regression Test Extension binary size
      shell: bash
      run: |
        python scripts/regression_test_extension_size.py --old 'duckdb/build/release/extension' --new build/release/extension --expect json,parquet,tpch,tpcds

 regression-test-python:
  name: Regression Test (Python Client)
  runs-on: ubuntu-20.04
  env:
    CC: gcc-10
    CXX: g++-10
    GEN: ninja

  steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - uses: actions/setup-python@v5
      with:
        python-version: '3.12'

    - name: Install
      shell: bash
      run: |
        sudo apt-get update -y -qq && sudo apt-get install -y -qq ninja-build
        pip install numpy pytest pandas mypy psutil pyarrow

    - name: Setup Ccache
      uses: hendrikmuhs/ccache-action@main
      with:
        key: ${{ github.job }}
        save: ${{ github.ref == 'refs/heads/main' || github.repository != 'duckdb/duckdb' }}

    - name: Build Current Version
      shell: bash
      run: |
        cd tools/pythonpkg
        pip install --use-pep517 . --user
        cd ../..

    - name: Run New Version
      shell: bash
      run: |
        python scripts/regression_test_python.py --threads=2 --out-file=new.csv

    - name: Cleanup New Version
      shell: bash
      run: |
        cd tools/pythonpkg
        ./clean.sh
        cd ../..

    - name: Build Current
      shell: bash
      run: |
        git clone --branch ${{ env.BASE_BRANCH }} https://github.com/duckdb/duckdb.git --depth=1
        cd duckdb/tools/pythonpkg
        pip install --use-pep517 . --user
        cd ../../..

    - name: Run Current Version
      shell: bash
      run: |
        python scripts/regression_test_python.py --threads=2 --out-file=current.csv

    - name: Regression Test
      shell: bash
      run: |
        cp -r benchmark duckdb/
        python scripts/regression_check.py --old=current.csv --new=new.csv

 regression-test-plan-cost:
  name: Regression Test Join Order Plan Cost
  runs-on: ubuntu-20.04
  env:
    CC: gcc-10
    CXX: g++-10
    GEN: ninja
    BUILD_TPCH: 1
    BUILD_HTTPFS: 1

  steps:
    - uses: actions/checkout@v4
      with:
        fetch-depth: 0

    - uses: actions/setup-python@v5
      with:
        python-version: '3.12'

    - name: Install
      shell: bash
      run: sudo apt-get update -y -qq && sudo apt-get install -y -qq ninja-build && pip install tqdm

    - name: Setup Ccache
      uses: hendrikmuhs/ccache-action@main
      with:
        key: ${{ github.job }}
        save: ${{ github.ref == 'refs/heads/main' || github.repository != 'duckdb/duckdb' }}

    - name: Build
      shell: bash
      run: |
        make
        git clone --branch ${{ env.BASE_BRANCH }} https://github.com/duckdb/duckdb.git --depth=1
        cd duckdb
        make
        cd ..

    - name: Set up benchmarks
      shell: bash
      run: |
        cp -r benchmark duckdb/

    - name: Regression Test IMDB
      if: always()
      shell: bash
      run: |
        python scripts/plan_cost_runner.py --old=duckdb/build/release/duckdb --new=build/release/duckdb --dir=benchmark/imdb_plan_cost

    - name: Regression Test TPCH
      if: always()
      shell: bash
      run: |
        python scripts/plan_cost_runner.py --old=duckdb/build/release/duckdb --new=build/release/duckdb --dir=benchmark/tpch_plan_cost