update-vector-stores.yml

name: Update Vector Stores

on:
    schedule:
        # Run every Sunday at midnight UTC (0 0 * * 0)
        - cron: "0 0 * * 0"

jobs:
    update-vector-stores:
        runs-on: ubuntu-latest

        # Lists of directories to exclude for each repository
        # Add comma-separated directory names to exclude them from indexing
        # Example: EXCLUDE_AQA_TESTS: 'node_modules,dist,build'
        env:
            EXCLUDE_AQA_TESTS: ""
            EXCLUDE_TKG: ""
            EXCLUDE_AQA_SYSTEMTEST: ""
            EXCLUDE_AQA_TEST_TOOLS: "VitAI/RAG/VectorStore"
            EXCLUDE_STF: ""
            EXCLUDE_BUMBLEBENCH: ""
            EXCLUDE_RUN_AQA: ""
            EXCLUDE_OPENJ9_SYSTEMTEST: ""
            EXCLUDE_OPENJ9: ""
            EXCLUDE_OMR: ""
            EXCLUDE_OPENJ9_OPENJDK_JDK: "src/hostpot"

        steps:
            - name: Checkout repository
              uses: actions/checkout@v4

            - name: Set up Python
              uses: actions/setup-python@v5
              with:
                  python-version: "3.x"
                  cache: "pip"
                  cache-dependency-path: "VitAI/RAG/requirements.txt"

            - name: Install dependencies
              run: |
                  cd VitAI/RAG
                  pip install -r requirements.txt

            - name: Create .env file
              run: |
                  cd VitAI/RAG
                  echo "GITHUB_TOKEN=${{ secrets.MODELS_TOKEN }}" > .env

            - name: Update adoptium/aqa-tests
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_AQA_TESTS }}" ]; then
                    python main.py --repo adoptium/aqa-tests --exclude "${{ env.EXCLUDE_AQA_TESTS }}"
                  else
                    python main.py --repo adoptium/aqa-tests
                  fi

            - name: Update adoptium/TKG
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_TKG }}" ]; then
                    python main.py --repo adoptium/TKG --exclude "${{ env.EXCLUDE_TKG }}"
                  else
                    python main.py --repo adoptium/TKG
                  fi

            - name: Update adoptium/aqa-systemtest
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_AQA_SYSTEMTEST }}" ]; then
                    python main.py --repo adoptium/aqa-systemtest --exclude "${{ env.EXCLUDE_AQA_SYSTEMTEST }}"
                  else
                    python main.py --repo adoptium/aqa-systemtest
                  fi

            - name: Update adoptium/aqa-test-tools
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_AQA_TEST_TOOLS }}" ]; then
                    python main.py --repo adoptium/aqa-test-tools --exclude "${{ env.EXCLUDE_AQA_TEST_TOOLS }}"
                  else
                    python main.py --repo adoptium/aqa-test-tools
                  fi

            - name: Update adoptium/STF
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_STF }}" ]; then
                    python main.py --repo adoptium/STF --exclude "${{ env.EXCLUDE_STF }}"
                  else
                    python main.py --repo adoptium/STF
                  fi

            - name: Update adoptium/bumblebench
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_BUMBLEBENCH }}" ]; then
                    python main.py --repo adoptium/bumblebench --exclude "${{ env.EXCLUDE_BUMBLEBENCH }}"
                  else
                    python main.py --repo adoptium/bumblebench
                  fi

            - name: Update adoptium/run-aqa
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_RUN_AQA }}" ]; then
                    python main.py --repo adoptium/run-aqa --exclude "${{ env.EXCLUDE_RUN_AQA }}"
                  else
                    python main.py --repo adoptium/run-aqa
                  fi

            - name: Update adoptium/openj9-systemtest
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_OPENJ9_SYSTEMTEST }}" ]; then
                    python main.py --repo adoptium/openj9-systemtest --exclude "${{ env.EXCLUDE_OPENJ9_SYSTEMTEST }}"
                  else
                    python main.py --repo adoptium/openj9-systemtest
                  fi

            - name: Update eclipse-openj9/openj9
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_OPENJ9 }}" ]; then
                    python main.py --repo eclipse-openj9/openj9 --exclude "${{ env.EXCLUDE_OPENJ9 }}"
                  else
                    python main.py --repo eclipse-openj9/openj9
                  fi

            - name: Update eclipse-omr/omr
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_OMR }}" ]; then
                    python main.py --repo eclipse-omr/omr --exclude "${{ env.EXCLUDE_OMR }}"
                  else
                    python main.py --repo eclipse-omr/omr
                  fi

            - name: Update ibmruntimes/openj9-openjdk-jdk
              run: |
                  cd VitAI/RAG
                  if [ -n "${{ env.EXCLUDE_OPENJ9_OPENJDK_JDK }}" ]; then
                    python main.py --repo ibmruntimes/openj9-openjdk-jdk --exclude "${{ env.EXCLUDE_OPENJ9_OPENJDK_JDK }}"
                  else
                    python main.py --repo ibmruntimes/openj9-openjdk-jdk
                  fi

            - name: Check for changes
              id: check_changes
              run: |
                  git config --global user.name 'github-actions[bot]'
                  git config --global user.email 'github-actions[bot]@users.noreply.github.com'

                  # Check if there are any changes to commit
                  if git diff --quiet && git diff --staged --quiet; then
                    echo "has_changes=false" >> $GITHUB_OUTPUT
                    echo "No changes detected in vector stores"
                  else
                    echo "has_changes=true" >> $GITHUB_OUTPUT
                    echo "Changes detected in vector stores"
                  fi

            - name: Commit and push changes
              if: steps.check_changes.outputs.has_changes == 'true'
              run: |
                  git add VitAI/RAG/VectorStore/
                  git add VitAI/RAG/.checkpoints/
                  git commit -m "chore: update vector stores (automated weekly update)"
                  git push

            - name: Summary
              run: |
                  if [ "${{ steps.check_changes.outputs.has_changes }}" == "true" ]; then
                    echo "✅ Vector stores updated and committed successfully"
                  else
                    echo "ℹ️ No updates needed - vector stores are up to date"
                  fi