continuous-integration.yaml

# @author Ivan Senic
name: Continuous Integration

# runs on
# * pushes and pull requests on the "main" (pull request only for specific paths)
# * manual trigger
on:
  push:
    branches: [ "main", "feature/*" ]

  pull_request:
    branches: [ "main", "feature/*" ]

  workflow_dispatch:

# cancel same workflows in progress for pull request branches
concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: ${{ github.ref != 'refs/heads/main' }}

# needed when a workflow wants to use OIDC (OpenID Connect) to authenticate to cloud
permissions:
  id-token: write
  contents: write
  packages: write
  pull-requests: write

# global env vars, available in all jobs and steps
env:
  MAVEN_OPTS: '-Xmx4g'
  DS_ARTIFACTORY_USERNAME: ${{ secrets.DS_ARTIFACTORY_USERNAME }}
  DS_ARTIFACTORY_PASSWORD: ${{ secrets.DS_ARTIFACTORY_PASSWORD }}

# Jobs structure:
#
# 1. Runs unit tests
# 2. Then 2 jobs in parallel
#  a) Integration tests with DSE 6.9
#  b) Integration tests with HCD
jobs:

  # runs unit tests
  build:
    name: Unit tests
    runs-on: ubuntu-latest

    # max run time 12 minutes
    timeout-minutes: 12

    steps:
      - uses: actions/checkout@v7

      - name: Set up JDK 21
        uses: actions/setup-java@v5
        with:
          distribution: 'temurin'
          java-version: '21'
          cache: maven

      - name: Setup Maven
        run: |
          mkdir -p ~/.m2
          cat <<EOF > ~/.m2/settings.xml
          <settings>
            <servers>
              <server>
                <id>stargate-central</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
              </server>
              <server>
                <id>stargate-snapshots</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
             </server>
              <server>
                <id>artifactory</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
              </server>
              <server>
                <id>artifactory-snapshots</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
              </server>
              <server>
                <id>artifactory-releases</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
             </server>
           </servers>
          </settings>
          EOF

      - name: Build & Test
        run: |
          ./mvnw -B -ntp -Dstyle.color=always clean test

      # Publish baseline coverage for main branch (for future PR comparisons)
      - name: Publish baseline coverage
        if: github.ref == 'refs/heads/main' && github.event_name == 'push'
        uses: actions/upload-artifact@v7
        with:
          name: baseline-coverage
          path: target/site/jacoco/jacoco.xml
          retention-days: 30

      # Download baseline coverage from main branch for PR comparison
      - name: Download baseline coverage
        if: github.event_name == 'pull_request'
        continue-on-error: true
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          # Get the latest successful run on main that has the baseline-coverage artifact
          run_id=$(gh run list --workflow=continuous-integration.yaml --branch=main --status=success --limit=1 --json databaseId --jq='.[0].databaseId')
          if [ -n "$run_id" ]; then
            echo "Downloading baseline coverage from run $run_id"
            gh run download $run_id --name baseline-coverage --dir baseline-coverage
          else
            echo "No successful workflow run found on main branch with coverage artifact"
          fi

      - name: Calculate and report coverage delta
        if: github.event_name == 'pull_request'
        id: coverage-delta
        continue-on-error: true
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          PR_NUMBER: ${{ github.event.pull_request.number }}
          REPO: ${{ github.repository }}
        run: |
          python3 << 'EOF'
          import os
          import json
          import urllib.request

          def get_coverage(jacoco_dir):
              """Calculate coverage from JaCoCo XML file"""
              import xml.etree.ElementTree as ET

              xml_file = os.path.join(jacoco_dir, 'jacoco.xml')
              if os.path.exists(xml_file):
                  try:
                      tree = ET.parse(xml_file)
                      root = tree.getroot()
                      # Find report-level counter (direct child of root with largest total)
                      # Strip namespace if present
                      max_total = 0
                      best_covered = 0
                      best_missed = 0
                      for child in root:
                          tag = child.tag.split('}')[-1] if '}' in child.tag else child.tag
                          if tag == 'counter' and child.get('type') == 'INSTRUCTION':
                              covered = int(child.get('covered', 0))
                              missed = int(child.get('missed', 0))
                              total = covered + missed
                              if total > max_total:
                                  max_total = total
                                  best_covered = covered
                                  best_missed = missed
                      total = best_covered + best_missed
                      pct = (best_covered / total * 100) if total > 0 else 0
                      print(f'XML parse ({xml_file}): covered={best_covered}, missed={best_missed}, total={total}, pct={pct:.2f}%')
                      return pct
                  except Exception as e:
                      print(f'XML parse error: {e}')
                      pass

              return 0

          def github_api(method, endpoint, data=None):
              url = f"https://api.github.com/repos/{os.environ['REPO']}/{endpoint}"
              headers = {
                  'Authorization': f"token {os.environ['GH_TOKEN']}",
                  'Accept': 'application/vnd.github.v3+json'
              }
              req = urllib.request.Request(url, headers=headers, method=method)
              if data:
                  req.data = json.dumps(data).encode()
              return json.loads(urllib.request.urlopen(req).read())

          # Debug: Check if CSV exists
          import os as os_check
          csv_exists = os_check.path.exists('target/site/jacoco/jacoco.csv')
          xml_exists = os_check.path.exists('target/site/jacoco/jacoco.xml')
          print(f'Unit test files: CSV={csv_exists}, XML={xml_exists}')

          current = get_coverage('target/site/jacoco')
          baseline = get_coverage('baseline-coverage')
          print(f'Unit test coverage: current={current:.2f}%, baseline={baseline:.2f}%')
          has_baseline = os.path.exists('baseline-coverage/jacoco.xml')

          # Write outputs for job summary
          with open(os.environ['GITHUB_OUTPUT'], 'a') as f:
              f.write(f'current_coverage={current:.2f}\n')
              if has_baseline:
                  delta = current - baseline
                  f.write(f'baseline_coverage={baseline:.2f}\n')
                  f.write(f'delta={delta:.2f}\n')
                  f.write(f'has_baseline=true\n')
                  print(f'Coverage: {baseline:.2f}% (main) → {current:.2f}% (PR) = {delta:+.2f}%')

                  # Post PR comment with delta
                  emoji = '📈' if delta > 0 else '📉' if delta < 0 else '➡️'
                  color = '🟢' if delta > 0 else '🔴' if delta < 0 else '⚪'
                  sign = '+' if delta > 0 else ''
                  msg = '⚠️ Coverage decreased' if delta < 0 else '✅ Coverage improved!' if delta > 0 else 'ℹ️ Coverage unchanged'

                  comment_body = (
                      f"## {emoji} Unit Test Coverage Delta vs Main Branch\n\n"
                      f"| Metric | Value |\n"
                      f"|--------|-------|\n"
                      f"| **Main Branch** | {baseline:.2f}% |\n"
                      f"| **This PR** | {current:.2f}% |\n"
                      f"| **Delta** | {color} {sign}{delta:.2f}% |\n"
                      f"{msg}\n"
                  )

                  # Find and update existing comment or create new one
                  pr_number = os.environ['PR_NUMBER']
                  comments = github_api('GET', f'issues/{pr_number}/comments')
                  existing = next((c for c in comments if c.get('user', {}).get('login') == 'github-actions[bot]'
                                   and 'Unit Test Coverage Delta vs Main Branch' in c.get('body', '')), None)

                  if existing:
                      github_api('PATCH', f"issues/comments/{existing['id']}", {'body': comment_body})
                      print(f'Updated comment {existing["id"]}')
                  else:
                      github_api('POST', f'issues/{pr_number}/comments', {'body': comment_body})
                      print('Created new comment')
              else:
                  f.write(f'has_baseline=false\n')
                  print('⚠️  No baseline coverage (expected on first PR after setup)')
          EOF

      - name: Add coverage to PR
        id: jacoco
        uses: madrapps/jacoco-report@v1.8.0
        if: github.event_name == 'pull_request'
        with:
          paths: |
            ${{ github.workspace }}/target/site/jacoco/jacoco.xml
          token: ${{ secrets.GITHUB_TOKEN }}
          min-coverage-overall: 40
          min-coverage-changed-files: 60
          title: 'Unit Test Coverage Report'
          update-comment: true

      - name: Generate JaCoCo Badge
        id: jacoco-badge
        uses: cicirello/jacoco-badge-generator@v2
        with:
          badges-directory: .github/badges
          generate-branches-badge: true
          generate-summary: true

      - name: Add unit test coverage to Job Summary
        run: |
          COVERAGE=$(printf "%.2f" $(echo "${{ steps.jacoco-badge.outputs.coverage }} * 100" | bc -l))
          BRANCHES=$(printf "%.2f" $(echo "${{ steps.jacoco-badge.outputs.branches }} * 100" | bc -l))
          cat >> $GITHUB_STEP_SUMMARY << EOF
          ## 📊 Unit Test Code Coverage Report

          | Metric | Coverage |
          |--------|----------|
          | **Instructions** | ${COVERAGE}% |
          | **Branches** | ${BRANCHES}% |
          EOF

          # Add delta for PRs with baseline
          if [ "${{ github.event_name }}" == "pull_request" ] && [ "${{ steps.coverage-delta.outputs.has_baseline }}" == "true" ]; then
            delta="${{ steps.coverage-delta.outputs.delta }}"
            cat >> $GITHUB_STEP_SUMMARY << EOF

          ### Coverage vs Main: ${{ steps.coverage-delta.outputs.baseline_coverage }}% → ${{ steps.coverage-delta.outputs.current_coverage }}% (${delta}%)
          EOF
          elif [ "${{ github.event_name }}" == "pull_request" ]; then
            cat >> $GITHUB_STEP_SUMMARY << EOF

          _Baseline coverage will be available after first merge to main_
          EOF
          fi

          cat >> $GITHUB_STEP_SUMMARY << EOF

          📥 [Download Full HTML Report](https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }})
          EOF

      - name: Upload JaCoCo coverage report
        uses: actions/upload-artifact@v7
        with:
          name: jacoco-report
          path: target/site/jacoco/

  # runs int tests
  int-tests:
    name: Integration tests
    runs-on: ubuntu-latest

    # max run time 40 minutes
    timeout-minutes: 40

    strategy:

      # let all tests run, can find multiple failures in different setup
      fail-fast: false

      # matrix props:
      matrix:
        type: [ docker, dse69-it, hcd-it ]
        include:
          - type: dse69-it
            profile: '-Pdse69-it'
          - type: hcd-it
            profile: '-Phcd-it'
        exclude:
          - type: docker

    steps:
      - uses: actions/checkout@v7

      - name: Set up JDK 21
        uses: actions/setup-java@v5
        with:
          distribution: 'temurin'
          java-version: '21'
          cache: maven

      - name: Setup Maven
        run: |
          mkdir -p ~/.m2
          cat <<EOF > ~/.m2/settings.xml
          <settings>
            <servers>
              <server>
                <id>stargate-central</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
              </server>
              <server>
                <id>stargate-snapshots</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
             </server>
              <server>
                <id>artifactory</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
              </server>
              <server>
                <id>artifactory-snapshots</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
              </server>
              <server>
                <id>artifactory-releases</id>
                <username>${DS_ARTIFACTORY_USERNAME}</username>
                <password>${DS_ARTIFACTORY_PASSWORD}</password>
             </server>
           </servers>
          </settings>
          EOF

      - name: Configure AWS credentials via OIDC
        uses: aws-actions/configure-aws-credentials@v6
        with:
          role-to-assume: ${{ secrets.US_WEST_2_559669398656_DATA_API_ECR_HCD }}
          aws-region: us-west-2

      - name: Login to Amazon ECR
        id: login-ecr
        uses: aws-actions/amazon-ecr-login@v2
        with:
          mask-password: 'true'

      # run the int tests
      - name: Integration Test
        # -DRERANKING_CONFIG_RESOURCE=test-reranking-providers-config.yaml is to override the reranking config to customized one
        # -DEMBEDDING_CONFIG_RESOURCE=test-embedding-providers-config.yaml is to override the embedding config to customized one
        # -Pjacoco-it enables the quarkus-jacoco extension for IT coverage
        # -Dquarkus.package.write-transformed-bytecode-to-build-output=true is required for JaCoCo IT coverage
        # Note: Container image build is disabled for coverage (JaCoCo only works with JAR artifacts)
        run: |
          ./mvnw -B -ntp clean verify -Dstyle.color=always -DskipUnitTests -DRERANKING_CONFIG_RESOURCE=test-reranking-providers-config.yaml -DEMBEDDING_CONFIG_RESOURCE=test-embedding-providers-config.yaml -Drun-create-index-parallel=true -Pjacoco-it -Dquarkus.package.write-transformed-bytecode-to-build-output=true ${{ matrix.profile }}

      # Publish baseline IT coverage for main branch (for future PR comparisons)
      - name: Publish baseline IT coverage
        if: github.ref == 'refs/heads/main' && github.event_name == 'push'
        uses: actions/upload-artifact@v7
        with:
          name: baseline-it-coverage-${{ matrix.type }}
          path: target/site/jacoco-it/jacoco.xml
          retention-days: 30

      # Download baseline IT coverage from main branch for PR comparison
      - name: Download baseline IT coverage
        if: github.event_name == 'pull_request'
        continue-on-error: true
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          # Get the latest successful run on main that has the baseline IT coverage artifact
          run_id=$(gh run list --workflow=continuous-integration.yaml --branch=main --status=success --limit=1 --json databaseId --jq='.[0].databaseId')
          if [ -n "$run_id" ]; then
            echo "Downloading baseline IT coverage for ${{ matrix.type }} from run $run_id"
            gh run download $run_id --name baseline-it-coverage-${{ matrix.type }} --dir baseline-it-coverage-${{ matrix.type }} || echo "No baseline IT coverage artifact found for ${{ matrix.type }}"
          else
            echo "No successful workflow run found on main branch with IT coverage artifact"
          fi

      - name: Calculate and report IT coverage delta
        if: github.event_name == 'pull_request'
        id: it-coverage-delta
        continue-on-error: true
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          PR_NUMBER: ${{ github.event.pull_request.number }}
          REPO: ${{ github.repository }}
          MATRIX_TYPE: ${{ matrix.type }}
        run: |
          python3 << 'EOF'
          import os
          import json
          import urllib.request

          def get_coverage(jacoco_dir):
              """Calculate coverage from JaCoCo XML file"""
              import xml.etree.ElementTree as ET

              xml_file = os.path.join(jacoco_dir, 'jacoco.xml')
              if os.path.exists(xml_file):
                  try:
                      tree = ET.parse(xml_file)
                      root = tree.getroot()
                      # Find report-level counter (direct child of root with largest total)
                      # Strip namespace if present
                      max_total = 0
                      best_covered = 0
                      best_missed = 0
                      for child in root:
                          tag = child.tag.split('}')[-1] if '}' in child.tag else child.tag
                          if tag == 'counter' and child.get('type') == 'INSTRUCTION':
                              covered = int(child.get('covered', 0))
                              missed = int(child.get('missed', 0))
                              total = covered + missed
                              if total > max_total:
                                  max_total = total
                                  best_covered = covered
                                  best_missed = missed
                      total = best_covered + best_missed
                      pct = (best_covered / total * 100) if total > 0 else 0
                      print(f'XML parse ({xml_file}): covered={best_covered}, missed={best_missed}, total={total}, pct={pct:.2f}%')
                      return pct
                  except Exception as e:
                      print(f'XML parse error: {e}')
                      pass

              return 0

          def github_api(method, endpoint, data=None):
              url = f"https://api.github.com/repos/{os.environ['REPO']}/{endpoint}"
              headers = {
                  'Authorization': f"token {os.environ['GH_TOKEN']}",
                  'Accept': 'application/vnd.github.v3+json'
              }
              req = urllib.request.Request(url, headers=headers, method=method)
              if data:
                  req.data = json.dumps(data).encode()
              return json.loads(urllib.request.urlopen(req).read())

          matrix_type = os.environ['MATRIX_TYPE']
          current = get_coverage('target/site/jacoco-it')
          baseline = get_coverage(f'baseline-it-coverage-{matrix_type}')
          has_baseline = os.path.exists(f'baseline-it-coverage-{matrix_type}/jacoco.xml')

          # Write outputs for job summary
          with open(os.environ['GITHUB_OUTPUT'], 'a') as f:
              f.write(f'current_coverage={current:.2f}\n')
              if has_baseline:
                  delta = current - baseline
                  f.write(f'baseline_coverage={baseline:.2f}\n')
                  f.write(f'delta={delta:.2f}\n')
                  f.write(f'has_baseline=true\n')
                  print(f'IT Coverage ({matrix_type}): {baseline:.2f}% (main) → {current:.2f}% (PR) = {delta:+.2f}%')

                  # Post PR comment with delta
                  emoji = '📈' if delta > 0 else '📉' if delta < 0 else '➡️'
                  color = '🟢' if delta > 0 else '🔴' if delta < 0 else '⚪'
                  sign = '+' if delta > 0 else ''
                  msg = '⚠️ Coverage decreased' if delta < 0 else '✅ Coverage improved!' if delta > 0 else 'ℹ️ Coverage unchanged'

                  comment_body = (
                      f"## {emoji} Integration Test Coverage Delta vs Main Branch ({matrix_type})\n\n"
                      f"| Metric | Value |\n"
                      f"|--------|-------|\n"
                      f"| **Main Branch** | {baseline:.2f}% |\n"
                      f"| **This PR** | {current:.2f}% |\n"
                      f"| **Delta** | {color} {sign}{delta:.2f}% |\n"
                      f"{msg}\n"
                  )

                  # Find and update existing comment or create new one
                  pr_number = os.environ['PR_NUMBER']
                  comments = github_api('GET', f'issues/{pr_number}/comments')
                  existing = next((c for c in comments if c.get('user', {}).get('login') == 'github-actions[bot]'
                                   and f'Integration Test Coverage Delta vs Main Branch ({matrix_type})' in c.get('body', '')), None)

                  if existing:
                      github_api('PATCH', f"issues/comments/{existing['id']}", {'body': comment_body})
                      print(f'Updated comment {existing["id"]}')
                  else:
                      github_api('POST', f'issues/{pr_number}/comments', {'body': comment_body})
                      print('Created new comment')
              else:
                  f.write(f'has_baseline=false\n')
                  print(f'⚠️  No baseline IT coverage for {matrix_type} (expected on first PR after setup)')
          EOF

      - name: Add integration test coverage to PR
        id: jacoco-it
        uses: madrapps/jacoco-report@v1.8.0
        if: github.event_name == 'pull_request'
        with:
          paths: |
            ${{ github.workspace }}/target/site/jacoco-it/jacoco.xml
          token: ${{ secrets.GITHUB_TOKEN }}
          min-coverage-overall: 40
          min-coverage-changed-files: 60
          title: 'Integration Test Coverage Report (${{ matrix.type }})'
          update-comment: true
        continue-on-error: true

      - name: Generate JaCoCo IT Badge
        id: jacoco-it-badge
        uses: cicirello/jacoco-badge-generator@v2
        with:
          jacoco-csv-file: target/site/jacoco-it/jacoco.csv
          badges-directory: .github/badges/it-${{ matrix.type }}
          generate-branches-badge: true
          generate-summary: true
        continue-on-error: true

      - name: Add IT coverage to Job Summary
        run: |
          COVERAGE=$(printf "%.2f" $(echo "${{ steps.jacoco-it-badge.outputs.coverage }} * 100" | bc -l))
          BRANCHES=$(printf "%.2f" $(echo "${{ steps.jacoco-it-badge.outputs.branches }} * 100" | bc -l))
          echo "## 📊 Integration Test Code Coverage Report (${{ matrix.type }})" >> $GITHUB_STEP_SUMMARY
          echo "" >> $GITHUB_STEP_SUMMARY
          echo "| Metric | Coverage |" >> $GITHUB_STEP_SUMMARY
          echo "|--------|----------|" >> $GITHUB_STEP_SUMMARY
          echo "| **Instructions** | ${COVERAGE}% |" >> $GITHUB_STEP_SUMMARY
          echo "| **Branches** | ${BRANCHES}% |" >> $GITHUB_STEP_SUMMARY
          echo "" >> $GITHUB_STEP_SUMMARY

          # Add delta for PRs with baseline
          if [ "${{ github.event_name }}" == "pull_request" ] && [ "${{ steps.it-coverage-delta.outputs.has_baseline }}" == "true" ]; then
            delta="${{ steps.it-coverage-delta.outputs.delta }}"
            echo "### Coverage vs Main: ${{ steps.it-coverage-delta.outputs.baseline_coverage }}% → ${{ steps.it-coverage-delta.outputs.current_coverage }}% (${delta}%)" >> $GITHUB_STEP_SUMMARY
            echo "" >> $GITHUB_STEP_SUMMARY
          elif [ "${{ github.event_name }}" == "pull_request" ]; then
            echo "_Baseline IT coverage will be available after first merge to main_" >> $GITHUB_STEP_SUMMARY
            echo "" >> $GITHUB_STEP_SUMMARY
          fi

          echo "📥 [Download Full HTML Report](https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }})" >> $GITHUB_STEP_SUMMARY
          echo "" >> $GITHUB_STEP_SUMMARY
          echo "The complete JaCoCo HTML report is available as an artifact named \`jacoco-it-report-${{ matrix.type }}\`" >> $GITHUB_STEP_SUMMARY

      - name: Upload JaCoCo integration test code coverage report
        uses: actions/upload-artifact@v7
        with:
          name: jacoco-it-report-${{ matrix.type }}
          path: target/site/jacoco-it/