ci.yaml

name: GlareDB CI

on:
  push:
    branches:
      - main
  pull_request:
  workflow_dispatch:

concurrency:
  group: build-ci-${{ github.ref }}
  cancel-in-progress: true

jobs:
  build:
    name: Build
    runs-on: ubuntu-latest-8-cores
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: toolchain cache
        with:
          path: |
            ~/.rustup/toolchains/
          key: ${{ runner.os }}-toolchain-${{ hashFiles('**/rust-toolchain.toml') }}
      - uses: actions/cache@v4
        name: cargo cache
        with:
          path: |
            ~/.cargo/
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - uses: actions/cache@v4
        name: build cache
        with:
          path: |
            target/
            !target/**/glaredb
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - uses: actions/cache@v4
        name: glaredb cache
        with:
          path: target/debug/glaredb
          key: ${{ github.run_id }}
      - name: build
        run: just build

  static-analysis:
    name: Lint and Format
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: toolchain cache
        with:
          path: |
            ~/.rustup/toolchains/
          key: ${{ runner.os }}-toolchain-${{ hashFiles('**/rust-toolchain.toml') }}
      - uses: actions/cache@v4
        name: cargo cache
        with:
          path: |
            ~/.cargo/
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}

      - name: clippy
        run: just clippy

      - name: format
        run: just fmt-check

  unit-tests:
    name: Unit Tests
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: toolchain cache
        with:
          path: |
            ~/.rustup/toolchains/
          key: ${{ runner.os }}-toolchain-${{ hashFiles('**/rust-toolchain.toml') }}
      - uses: actions/cache@v4
        name: cargo cache
        with:
          path: |
            ~/.cargo/
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}

      - name: run tests
        run: just unit-tests

  python-binding-tests:
    name: Python Binding Tests
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: toolchain cache
        with:
          path: |
            ~/.rustup/toolchains/
          key: ${{ runner.os }}-toolchain-${{ hashFiles('**/rust-toolchain.toml') }}
      - uses: actions/cache@v4
        name: cargo cache
        with:
          path: |
            ~/.cargo/
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - uses: actions/cache@v4
        name: build cache
        with:
          path: |
            target/
            !target/**/glaredb
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}

      - name: build python bindings
        run: just python build

      - name: run python binding tests
        run: just python test

  nodejs-bindings-tests:
    name: Node.js Binding Tests
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - uses: actions/checkout@v4

      - uses: actions/setup-node@v4
        with:
          node-version: 20
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: toolchain cache
        with:
          path: |
            ~/.rustup/toolchains/
          key: ${{ runner.os }}-toolchain-${{ hashFiles('**/rust-toolchain.toml') }}
      - uses: actions/cache@v4
        name: cargo cache
        with:
          path: |
            ~/.cargo/
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - uses: actions/cache@v4
        name: build cache
        with:
          path: |
            target/
            !target/**/glaredb
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - name: build node.js bindings
        run: just js build-debug
      - name: run node.js binding tests
        run: just js test

  pg-protocol:
    name: PG Protocol Tests
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: toolchain cache
        with:
          path: |
            ~/.rustup/toolchains/
          key: ${{ runner.os }}-toolchain-${{ hashFiles('**/rust-toolchain.toml') }}
      - uses: actions/cache@v4
        name: cargo cache
        with:
          path: |
            ~/.cargo/
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - uses: actions/cache@v4
        name: build cache
        with:
          path: |
            target/
            !target/**/glaredb
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - name: pg protocol test (script)
        run: |
          PROTOC=`just protoc && just --evaluate PROTOC` ./scripts/protocol-test.sh

      - name: pg protocol tests (slt runner)
        run: just slt-bin-debug 'pgproto/*'

  sql-logic-tests:
    name: SQL Logic Tests
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    strategy:
      matrix:
        protocol: ["postgres", "flightsql", "rpc"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: glaredb cache
        with:
          path: target/debug/glaredb
          key: ${{ github.run_id }}
      - name: public sql logic tests DEBUG
        if: ${{ env.ACTIONS_STEP_DEBUG == 'true' }}
        run: |
          if [[ "${{ matrix.protocol }}" == "rpc" ]]; then
            just rpc-tests
          else
            just slt-bin-debug --protocol=${{ matrix.protocol }} 'sqllogictests/*'
          fi
      - name: public sql logic tests
        if: ${{ env.ACTIONS_STEP_DEBUG != 'true' }}
        run: |
          if [[ "${{ matrix.protocol }}" == "rpc" ]]; then
            just rpc-tests
          else
            just slt-bin-debug --protocol=${{ matrix.protocol }} 'sqllogictests/*'
          fi

  process-integration-tests:
    name: Process Integration Tests (pytest)
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - name: install poetry
        uses: snok/install-poetry@v1
      - name: install python
        uses: actions/setup-python@v5
        with:
          python-version: "3.11"
          cache: poetry
          cache-dependency-path: tests/poetry.lock
      - uses: actions/cache@v4
        name: toolchain cache
        with:
          path: |
            ~/.rustup/toolchains/
          key: ${{ runner.os }}-toolchain-${{ hashFiles('**/rust-toolchain.toml') }}
      - uses: actions/cache@v4
        name: cargo cache
        with:
          path: |
            ~/.cargo/
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - uses: actions/cache@v4
        name: build cache
        with:
          path: |
            target/
            !target/**/glaredb
          key: ${{ runner.os }}-cargo-${{ hashFiles('**/Cargo.lock') }}
      - uses: actions/cache@v4
        name: glaredb cache
        with:
          path: target/debug/glaredb
          key: ${{ github.run_id }}
      - uses: actions/cache@v4
        name: py cache
        with:
          path: |
            tests/.venv/
          key: ${{ runner.os }}-poetry-${{ hashFiles('**/Cargo.lock') }}
      - name: setup pytest
        run: just pytest-setup
      - name: run pytest
        run: just pytest

  service-integration-tests:
    name: Service Integration Tests (SLT)
    if: github.event_name != 'pull_request' || github.event.pull_request.head.repo.owner.login == 'GlareDB'
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: glaredb cache
        with:
          path: target/debug/glaredb
          key: ${{ github.run_id }}

      - name: GCP authenticate
        uses: google-github-actions/auth@v2
        with:
          credentials_json: ${{ secrets.GCP_SERVICE_ACCOUNT_JSON }}

      - name: setup GCP
        uses: google-github-actions/setup-gcloud@v2

      - name: prepare testdata
        run: ./scripts/prepare-testdata.sh

      - name: run tests (slt)
        env:
          GCP_SERVICE_ACCOUNT_KEY: ${{ secrets.GCP_SERVICE_ACCOUNT_JSON }}
          GCP_PROJECT_ID: glaredb-artifacts
          AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
          AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
          AZURE_ACCESS_KEY: ${{ secrets.AZURE_ACCESS_KEY }}
          AZURE_ACCOUNT: ${{ secrets.AZURE_ACCOUNT }}
          MINIO_ACCESS_KEY: glaredb
          MINIO_SECRET_KEY: glaredb_test
          TEST_BUCKET: glaredb-test-bucket
        run: |
          # Prepare SLT (BigQuery)
          if ./scripts/files-changed-in-branch.sh \
            "scripts/prepare-testdata.sh" \
            "scripts/create-test-bigquery-db.sh" \
            "testdata/sqllogictests_datasources_common/data" \
            "testdata/sqllogictests_bigquery/data"
          then
            export GCP_PROJECT_ID=glaredb-dev-playground
            export BIGQUERY_DATASET_ID=$(./scripts/create-test-bigquery-db.sh)
          else
            export BIGQUERY_DATASET_ID=glaredb_test
          fi

          # Prepare SLT (Object store)
          export GCS_BUCKET_NAME=glaredb-test
          export AWS_S3_REGION=us-east-1
          export AWS_S3_BUCKET_NAME=glaredb-test

          # Unset application default credentials. We don't want to unknowingly
          # depend on this.
          unset GOOGLE_APPLICATION_CREDENTIALS

          # Prepare SLT (Postgres)
          POSTGRES_TEST_DB=$(./scripts/create-test-postgres-db.sh)
          export POSTGRES_CONN_STRING=$(echo "$POSTGRES_TEST_DB" | sed -n 1p)
          export POSTGRES_TUNNEL_SSH_CONN_STRING=$(echo "$POSTGRES_TEST_DB" | sed -n 2p)

          # Prepare SLT (MinIO)
          ./scripts/create-test-minio-store.sh

          # Prepare SLT (fake GCS server)
          ./scripts/create-test-gcs-store.sh

          echo "-------------------------------- WITHOUT TUNNEL TEST --------------------------------"
          # Run all data source tests without running tunnel tests or the basic
          # SLT tests.
          just slt-bin --exclude 'sqllogictests/*' \
            --exclude '*/tunnels/ssh' \
            --exclude 'sqllogictests_snowflake/*' \
            --exclude 'sqllogictests_cassandra/*' \
            --exclude 'sqllogictests_clickhouse/*' \
            --exclude 'sqllogictests_sqlserver/*' \
            --exclude 'sqllogictests_mongodb/*' \
            --exclude 'sqllogictests_mysql/*' \

          echo "-------------------------------- WITH TUNNEL TEST --------------------------------"
          # SSH tests are prone to fail if we make a lot of connections at the
          # same time. Hence, it makes sense to run all the SSH tests one-by-one
          # in order to test the SSH tunnels (which is our aim).
          just slt-bin  --jobs=1 '*/tunnels/ssh' --exclude 'sqllogictests_mysql/*'

          # TODO: move these into the datasource-integration-tests job

          echo "-------------------------------- RPC TESTS --------------------------------"

          echo "--------------------------  BigQuery -------------------------- "
          just slt-bin --protocol=rpc 'sqllogictests_bigquery/*'
          echo "--------------------------  Iceberg -------------------------- "
          just slt-bin --protocol=rpc 'sqllogictests_iceberg/*'

          echo "--------------------------  Native -------------------------- "
          just slt-bin --protocol=rpc 'sqllogictests_native/*'

          echo "--------------------------  Object Store -------------------------- "
          just slt-bin --protocol=rpc 'sqllogictests_object_store/*'

          echo "--------------------------  Postgres -------------------------- "
          just slt-bin --protocol=rpc --exclude '*/tunnels/ssh' 'sqllogictests_postgres/*'

          echo "--------------------------  Fake GCS server with a subdirectory -------------------------- "
          just slt-bin -l gs://$TEST_BUCKET/path/to/folder/1 -o service_account_path=/tmp/fake-gcs-creds.json 'sqllogictests_native/*'
          just slt-bin -l gs://$TEST_BUCKET/path/to/folder/2 -o service_account_path=/tmp/fake-gcs-creds.json 'sqllogictests_native/*'

  minio-integration-tests:
    name: MinIO Integration Tests (SLT::MinIO)
    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: glaredb cache
        with:
          path: target/debug/glaredb
          key: ${{ github.run_id }}

      - name: run tests (slt)
        env:
          MINIO_ACCESS_KEY: glaredb
          MINIO_SECRET_KEY: glaredb_test
          TEST_BUCKET: glaredb-test-bucket
        run: |
          # prepare testdata
          ./scripts/prepare-testdata.sh

          # Prepare SLT (MinIO)
          ./scripts/create-test-minio-store.sh

          echo "--------------------------  MinIO (S3) -------------------------- "

          # PG protocol
          just slt-bin --location http://localhost:9100 \
            --option access_key_id=$MINIO_ACCESS_KEY \
            --option secret_access_key=$MINIO_SECRET_KEY \
            --option bucket=$TEST_BUCKET \
            'sqllogictests/*' \
            'sqllogictests_native/*'

          # Flight protocol
          just slt-bin --location http://localhost:9100 \
            --option access_key_id=$MINIO_ACCESS_KEY \
            --option secret_access_key=$MINIO_SECRET_KEY \
            --option bucket=$TEST_BUCKET \
            'sqllogictests/*' \
            'sqllogictests_native/*' \
            --protocol=flightsql

          echo "--------------------------  MinIO (S3) With a sub-directory -------------------------- "
          # PG protocol
          just slt-bin -l http://localhost:9100/$TEST_BUCKET/path/to/folder \
              -o access_key_id=$MINIO_ACCESS_KEY \
              -o secret_access_key=$MINIO_SECRET_KEY \
              'sqllogictests/*'

          # Flight protocol
          just slt-bin -l http://localhost:9100/$TEST_BUCKET/path/to/folder \
              -o access_key_id=$MINIO_ACCESS_KEY \
              -o secret_access_key=$MINIO_SECRET_KEY \
              'sqllogictests/*' \
              --protocol=flightsql

  datasource-integration-tests:
    name: Datasource Integration (${{matrix.settings.name}})
    strategy:
      matrix:
        settings:
          - name: Clickhouse
            path: "sqllogictests_clickhouse/*"
            prepare: |
              ./scripts/prepare-testdata.sh
              source ./scripts/ci-install-clickhouse.sh
              export CLICKHOUSE_CONN_STRING=$(./scripts/create-test-clickhouse-db.sh)
          - name: Cassandra
            path: "sqllogictests_cassandra/*"
            prepare: |
              export CASSANDRA_CONN_STRING=$(./scripts/create-test-cassandra-db.sh | tail -n 1)
          - name: Mysql
            path: "sqllogictests_mysql/*"
            prepare: |
              ./scripts/prepare-testdata.sh
              export MYSQL_CONN_STRING=$(./scripts/create-test-mysql-db.sh)
              export MYSQL_TUNNEL_SSH_CONN_STRING=$(echo "$MYSQL_CONN_STRING" | sed -n 2p)
              export MYSQL_CONN_STRING=$(echo "$MYSQL_CONN_STRING" | sed -n 1p)
          - name: MongoDB
            path: "sqllogictests_mongodb/*"
            prepare: |
              ./scripts/prepare-testdata.sh
              export MONGO_CONN_STRING=$(./scripts/create-test-mongo-db.sh)
          - name: Sqlserver
            path: "sqllogictests_sqlserver/*"
            prepare: |
              ./scripts/prepare-testdata.sh
              export SQL_SERVER_CONN_STRING=$(./scripts/create-test-sqlserver-db.sh)

    runs-on: ubuntu-latest-8-cores
    needs: ["build"]
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: glaredb cache
        with:
          path: target/debug/glaredb
          key: ${{ github.run_id }}
      - name: run tests (slt)
        run: |
          ${{matrix.settings.prepare}}

          just slt-bin ${{matrix.settings.path}}
          just slt-bin --protocol=rpc --exclude '*/tunnels/ssh' ${{matrix.settings.path}}

          # Run flight SLTs for everything except clickhouse.
          #
          # <https://github.com/GlareDB/glaredb/issues/2389>
          if [[ "${{ matrix.settings.name }}" != 'Clickhouse' ]]; then
             just slt-bin --protocol=flightsql --exclude '*/tunnels/ssh' ${{matrix.settings.path}}
          fi

  service-integration-tests-snowflake:
    if: github.event_name != 'pull_request' || github.event.pull_request.head.repo.owner.login == 'GlareDB'
    name: Snowflake Service Integration Tests (SLT::Snowflake)
    runs-on: ubuntu-latest-8-cores
    needs: ["service-integration-tests"]
    concurrency:
      group: snowflake-integration-tests
      cancel-in-progress: false
    steps:
      - name: checkout
        uses: actions/checkout@v4
      - uses: extractions/setup-just@v1
        with:
          just-version: "1.23.0"
      - uses: actions/cache@v4
        name: glaredb cache
        with:
          path: target/debug/glaredb
          key: ${{ github.run_id }}

      - name: snowflake setup (SnowSQL)
        run: |
          curl -o snowsql.bash \
            https://sfc-repo.snowflakecomputing.com/snowsql/bootstrap/1.2/linux_x86_64/snowsql-1.2.24-linux_x86_64.bash
          mkdir -p ~/bin
          SNOWSQL_DEST=~/bin SNOWSQL_LOGIN_SHELL=~/.profile bash snowsql.bash

      - name: prepare testdata
        run: ./scripts/prepare-testdata.sh

      - name: run tests (slt)
        env:
          SNOWFLAKE_USERNAME: ${{ secrets.SNOWFLAKE_USERNAME }}
          SNOWFLAKE_PASSWORD: ${{ secrets.SNOWFLAKE_PASSWORD }}
        run: |
          # Prepare SLT (Snowflake)
          export PATH="$HOME/bin:$PATH"
          if ./scripts/files-changed-in-branch.sh \
            "scripts/prepare-testdata.sh" \
            "scripts/create-test-snowflake-db.sh" \
            "testdata/sqllogictests_datasources_common/data" \
            "testdata/sqllogictests_snowflake/data"
          then
            export SNOWFLAKE_DATABASE=$(./scripts/create-test-snowflake-db.sh)
          else
            export SNOWFLAKE_DATABASE=glaredb_test
          fi

          just slt-bin 'sqllogictests_snowflake/*'
          just slt-bin --protocol=rpc 'sqllogictests_snowflake/*'