python-integration-tests.yml

#
# This workflow will run all python integrations tests.
#

name: Python Integration Tests

on:
  workflow_dispatch:
  pull_request:
    branches: ["main"]
  merge_group:
    branches: ["main"]
  schedule:
    - cron: "0 0 * * *" # Run at midnight UTC daily

permissions:
  contents: read
  id-token: "write"

env:
  # Configure a constant location for the uv cache
  UV_CACHE_DIR: /tmp/.uv-cache
  HNSWLIB_NO_NATIVE: 1
  Python_Integration_Tests: Python_Integration_Tests
  AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME: ${{ vars.AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME }} # azure-text-embedding-ada-002
  AZURE_OPENAI_CHAT_DEPLOYMENT_NAME: ${{ vars.AZURE_OPENAI_CHAT_DEPLOYMENT_NAME }}
  AZURE_OPENAI_TEXT_DEPLOYMENT_NAME: ${{ vars.AZURE_OPENAI_TEXT_DEPLOYMENT_NAME }}
  AZURE_OPENAI_AUDIO_TO_TEXT_DEPLOYMENT_NAME: ${{ vars.AZURE_OPENAI_AUDIO_TO_TEXT_DEPLOYMENT_NAME }}
  AZURE_OPENAI_TEXT_TO_AUDIO_DEPLOYMENT_NAME: ${{ vars.AZURE_OPENAI_TEXT_TO_AUDIO_DEPLOYMENT_NAME }}
  AZURE_OPENAI_TEXT_TO_IMAGE_DEPLOYMENT_NAME: ${{ vars.AZURE_OPENAI_TEXT_TO_IMAGE_DEPLOYMENT_NAME }}
  AZURE_OPENAI_API_VERSION: ${{ vars.AZURE_OPENAI_API_VERSION }}
  AZURE_OPENAI_ENDPOINT: ${{ secrets.AZURE_OPENAI_ENDPOINT }}
  AZURE_OPENAI_AUDIO_TO_TEXT_ENDPOINT: ${{ secrets.AZURE_OPENAI_AUDIO_TO_TEXT_ENDPOINT }}
  AZURE_OPENAI_TEXT_TO_AUDIO_ENDPOINT: ${{ secrets.AZURE_OPENAI_TEXT_TO_AUDIO_ENDPOINT }}
  BING_API_KEY: ${{ secrets.BING_API_KEY }}
  OPENAI_CHAT_MODEL_ID: ${{ vars.OPENAI_CHAT_MODEL_ID }}
  OPENAI_TEXT_MODEL_ID: ${{ vars.OPENAI_TEXT_MODEL_ID }}
  OPENAI_EMBEDDING_MODEL_ID: ${{ vars.OPENAI_EMBEDDING_MODEL_ID }}
  OPENAI_AUDIO_TO_TEXT_MODEL_ID: ${{ vars.OPENAI_AUDIO_TO_TEXT_MODEL_ID }}
  OPENAI_TEXT_TO_AUDIO_MODEL_ID: ${{ vars.OPENAI_TEXT_TO_AUDIO_MODEL_ID }}
  OPENAI_TEXT_TO_IMAGE_MODEL_ID: ${{ vars.OPENAI_TEXT_TO_IMAGE_MODEL_ID }}
  OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
  PINECONE_API_KEY: ${{ secrets.PINECONE__APIKEY }}
  POSTGRES_CONNECTION_STRING: ${{secrets.POSTGRES__CONNECTIONSTR}}
  POSTGRES_MAX_POOL: ${{ vars.POSTGRES_MAX_POOL }}
  AZURE_AI_SEARCH_API_KEY: ${{secrets.AZURE_AI_SEARCH_API_KEY}}
  AZURE_AI_SEARCH_ENDPOINT: ${{secrets.AZURE_AI_SEARCH_ENDPOINT}}
  MONGODB_ATLAS_CONNECTION_STRING: ${{secrets.MONGODB_ATLAS_CONNECTION_STRING}}
  AZURE_KEY_VAULT_ENDPOINT: ${{secrets.AZURE_KEY_VAULT_ENDPOINT}}
  AZURE_KEY_VAULT_CLIENT_ID: ${{secrets.AZURE_KEY_VAULT_CLIENT_ID}}
  AZURE_KEY_VAULT_CLIENT_SECRET: ${{secrets.AZURE_KEY_VAULT_CLIENT_SECRET}}
  ACA_POOL_MANAGEMENT_ENDPOINT: ${{secrets.ACA_POOL_MANAGEMENT_ENDPOINT}}
  MISTRALAI_API_KEY: ${{secrets.MISTRALAI_API_KEY}}
  MISTRALAI_CHAT_MODEL_ID: ${{ vars.MISTRALAI_CHAT_MODEL_ID }}
  MISTRALAI_EMBEDDING_MODEL_ID: ${{ vars.MISTRALAI_EMBEDDING_MODEL_ID }}
  ANTHROPIC_API_KEY: ${{secrets.ANTHROPIC_API_KEY}}
  ANTHROPIC_CHAT_MODEL_ID: ${{ vars.ANTHROPIC_CHAT_MODEL_ID }}
  OLLAMA_CHAT_MODEL_ID: "${{ vars.OLLAMA_CHAT_MODEL_ID || '' }}" # llava-phi3
  OLLAMA_CHAT_MODEL_ID_IMAGE: "${{ vars.OLLAMA_CHAT_MODEL_ID_IMAGE || '' }}" # llava-phi3
  OLLAMA_CHAT_MODEL_ID_TOOL_CALL: "${{ vars.OLLAMA_CHAT_MODEL_ID_TOOL_CALL || '' }}" # llama3.2
  OLLAMA_TEXT_MODEL_ID: "${{ vars.OLLAMA_TEXT_MODEL_ID || '' }}" # llava-phi3
  OLLAMA_EMBEDDING_MODEL_ID: "${{ vars.OLLAMA_EMBEDDING_MODEL_ID || '' }}" # nomic-embed-text
  GOOGLE_AI_GEMINI_MODEL_ID: ${{ vars.GOOGLE_AI_GEMINI_MODEL_ID }}
  GOOGLE_AI_EMBEDDING_MODEL_ID: ${{ vars.GOOGLE_AI_EMBEDDING_MODEL_ID }}
  GOOGLE_AI_API_KEY: ${{ secrets.GOOGLE_AI_API_KEY }}
  VERTEX_AI_PROJECT_ID: ${{ vars.VERTEX_AI_PROJECT_ID }}
  VERTEX_AI_GEMINI_MODEL_ID: ${{ vars.VERTEX_AI_GEMINI_MODEL_ID }}
  VERTEX_AI_EMBEDDING_MODEL_ID: ${{ vars.VERTEX_AI_EMBEDDING_MODEL_ID }}
  REDIS_CONNECTION_STRING: ${{ vars.REDIS_CONNECTION_STRING }}
  AZURE_COSMOS_DB_NO_SQL_URL: ${{ vars.AZURE_COSMOS_DB_NO_SQL_URL }}
  AZURE_COSMOS_DB_NO_SQL_KEY: ${{ secrets.AZURE_COSMOS_DB_NO_SQL_KEY }}

jobs:
  paths-filter:
    runs-on: ubuntu-latest
    outputs:
      pythonChanges: ${{ steps.filter.outputs.python}}
    steps:
      - uses: actions/checkout@v4
      - uses: dorny/paths-filter@v2
        id: filter
        with:
          filters: |
            python:
              - 'python/**'
      # run only if 'python' files were changed
      - name: python tests
        if: steps.filter.outputs.python == 'true'
        run: echo "Python file"
      # run only if not 'python' files were changed
      - name: not python tests
        if: steps.filter.outputs.python != 'true'
        run: echo "NOT python file"

  python-merge-gate:
    name: Python Pre-Merge Integration Tests
    needs: paths-filter
    if: github.event_name != 'pull_request' && github.event_name != 'schedule' && needs.paths-filter.outputs.pythonChanges == 'true'
    strategy:
      max-parallel: 1
      fail-fast: false
      matrix:
        python-version: ["3.11"]
        os: [ubuntu-latest]
    defaults:
      run:
        working-directory: python
    runs-on: ${{ matrix.os }}
    environment: "integration"
    steps:
      - uses: actions/checkout@v4
      - name: Set up uv
        uses: astral-sh/setup-uv@v3
        with:
          version: "0.4.30"
          enable-cache: true
          cache-suffix: ${{ runner.os }}-${{ matrix.python-version }}
      - name: Install dependencies with hnswlib native disabled
        if: matrix.os == 'macos-latest' && matrix.python-version == '3.11'
        run: |
          export HNSWLIB_NO_NATIVE=1
          uv sync --all-extras --dev
      - name: Install dependencies with hnswlib native enabled
        if: matrix.os != 'macos-latest' || matrix.python-version != '3.11'
        run: |
          uv sync --all-extras --dev
      - name: Install Ollama
        if: matrix.os == 'ubuntu-latest'
        run: |
          curl -fsSL https://ollama.com/install.sh | sh
          ollama serve &
          sleep 5
      - name: Pull model in Ollama
        if: matrix.os == 'ubuntu-latest'
        run: |
          ollama pull ${{ vars.OLLAMA_CHAT_MODEL_ID }}
          ollama pull ${{ vars.OLLAMA_CHAT_MODEL_ID_IMAGE }}
          ollama pull ${{ vars.OLLAMA_CHAT_MODEL_ID_TOOL_CALL }}
          ollama pull ${{ vars.OLLAMA_TEXT_MODEL_ID }}
          ollama pull ${{ vars.OLLAMA_EMBEDDING_MODEL_ID }}
          ollama list
      - name: Google auth
        uses: google-github-actions/auth@v2
        with:
          project_id: ${{ vars.VERTEX_AI_PROJECT_ID }}
          credentials_json: ${{ secrets.VERTEX_AI_SERVICE_ACCOUNT_KEY }}
      - name: Set up gcloud
        uses: google-github-actions/setup-gcloud@v2
      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@v4
        with:
          aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
          aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
          aws-region: ${{ vars.AWS_REGION }}
      - name: Setup Redis Stack Server
        if: matrix.os == 'ubuntu-latest'
        run: docker run -d --name redis-stack-server -p 6379:6379 redis/redis-stack-server:latest
      - name: Setup Weaviate docker deployment
        if: matrix.os == 'ubuntu-latest'
        run: docker run -d -p 8080:8080 -p 50051:50051 cr.weaviate.io/semitechnologies/weaviate:1.26.6
      - name: Start Azure Cosmos DB emulator
        if: matrix.os == 'windows-latest'
        run: |
          Write-Host "Launching Cosmos DB Emulator"
          Import-Module "$env:ProgramFiles\Azure Cosmos DB Emulator\PSModules\Microsoft.Azure.CosmosDB.Emulator"
          Start-CosmosDbEmulator
      - name: Azure CLI Login
        if: github.event_name != 'pull_request'
        uses: azure/login@v2
        with:
          client-id: ${{ secrets.AZURE_CLIENT_ID }}
          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
      - name: Run Integration Tests - Completions
        id: run_tests_completions
        timeout-minutes: 15
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/completions -v --junitxml=pytest-completions.xml
      - name: Run Integration Tests - Embeddings
        id: run_tests_embeddings
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/embeddings -v --junitxml=pytest-embeddings.xml
      - name: Run Integration Tests - Memory
        id: run_tests_memory
        timeout-minutes: 10
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/memory -v --junitxml=pytest-memory.xml
      - name: Run Integration Tests - Cross Language
        id: run_tests_cross_language
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/cross_language -v --junitxml=pytest-cross.xml
      - name: Run Integration Tests - Planning
        id: run_tests_planning
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/planning -v --junitxml=pytest-planning.xml
      - name: Run Integration Tests - Samples
        id: run_tests_samples
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/samples -v --junitxml=pytest-samples.xml
      - name: Surface failing tests
        if: always()
        uses: pmeier/pytest-results-action@main
        with:
          path: python/pytest-*.xml
          summary: true
          display-options: fEX
          fail-on-empty: true
          title: Test results
      - name: Minimize uv cache
        run: uv cache prune --ci

  python-integration-tests:
    needs: paths-filter
    if: (github.event_name == 'schedule' || github.event_name == 'workflow_dispatch') && needs.paths-filter.outputs.pythonChanges == 'true'
    strategy:
      max-parallel: 1
      fail-fast: false
      matrix:
        python-version: ["3.10", "3.11", "3.12"]
        os: [ubuntu-latest, windows-latest, macos-latest]
    defaults:
      run:
        working-directory: python
    runs-on: ${{ matrix.os }}
    environment: "integration"
    steps:
      - uses: actions/checkout@v4
      - name: Set up uv
        uses: astral-sh/setup-uv@v3
        with:
          version: "0.4.30"
          enable-cache: true
          cache-suffix: ${{ runner.os }}-${{ matrix.python-version }}
      - name: Install dependencies with hnswlib native disabled
        if: matrix.os == 'macos-latest' && matrix.python-version == '3.11'
        run: |
          export HNSWLIB_NO_NATIVE=1
          uv sync --all-extras --dev
      - name: Install dependencies with hnswlib native enabled
        if: matrix.os != 'macos-latest' || matrix.python-version != '3.11'
        run: |
          uv sync --all-extras --dev
      - name: Install Ollama
        if: matrix.os == 'ubuntu-latest'
        run: |
          curl -fsSL https://ollama.com/install.sh | sh
          ollama serve &
          sleep 5
      - name: Pull model in Ollama
        if: matrix.os == 'ubuntu-latest'
        run: |
          ollama pull ${{ vars.OLLAMA_CHAT_MODEL_ID }}
          ollama pull ${{ vars.OLLAMA_CHAT_MODEL_ID_IMAGE }}
          ollama pull ${{ vars.OLLAMA_CHAT_MODEL_ID_TOOL_CALL }}
          ollama pull ${{ vars.OLLAMA_TEXT_MODEL_ID }}
          ollama pull ${{ vars.OLLAMA_EMBEDDING_MODEL_ID }}
          ollama list
      - name: Google auth
        uses: google-github-actions/auth@v2
        with:
          project_id: ${{ vars.VERTEX_AI_PROJECT_ID }}
          credentials_json: ${{ secrets.VERTEX_AI_SERVICE_ACCOUNT_KEY }}
      - name: Set up gcloud
        uses: google-github-actions/setup-gcloud@v2
      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@v4
        with:
          aws-access-key-id: ${{ secrets.AWS_ACCESS_KEY_ID }}
          aws-secret-access-key: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
          aws-region: ${{ vars.AWS_REGION }}
      - name: Setup Redis Stack Server
        if: matrix.os == 'ubuntu-latest'
        run: docker run -d --name redis-stack-server -p 6379:6379 redis/redis-stack-server:latest
      - name: Setup Weaviate docker deployment
        if: matrix.os == 'ubuntu-latest'
        run: docker run -d -p 8080:8080 -p 50051:50051 cr.weaviate.io/semitechnologies/weaviate:1.26.6
      - name: Start Azure Cosmos DB emulator
        if: matrix.os == 'windows-latest'
        run: |
          Write-Host "Launching Cosmos DB Emulator"
          Import-Module "$env:ProgramFiles\Azure Cosmos DB Emulator\PSModules\Microsoft.Azure.CosmosDB.Emulator"
          Start-CosmosDbEmulator
      - name: Azure CLI Login
        if: github.event_name != 'pull_request'
        uses: azure/login@v2
        with:
          client-id: ${{ secrets.AZURE_CLIENT_ID }}
          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}
      - name: Run Integration Tests - Completions
        id: run_tests_completions
        timeout-minutes: 10
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/completions -v --junitxml=pytest-completions.xml
      - name: Run Integration Tests - Embeddings
        id: run_tests_embeddings
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/embeddings -v --junitxml=pytest-embeddings.xml
      - name: Run Integration Tests - Memory
        id: run_tests_memory
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/memory -v --junitxml=pytest-memory.xml
      - name: Run Integration Tests - Cross Language
        id: run_tests_cross_language
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/cross_language -v --junitxml=pytest-cross.xml
      - name: Run Integration Tests - Planning
        id: run_tests_planning
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/integration/planning -v --junitxml=pytest-planning.xml
      - name: Run Integration Tests - Samples
        id: run_tests_samples
        timeout-minutes: 5
        shell: bash
        run: |
          uv run pytest -n logical --dist loadfile --dist worksteal ./tests/samples -v --junitxml=pytest-samples.xml
      - name: Surface failing tests
        if: always()
        uses: pmeier/pytest-results-action@main
        with:
          path: python/pytest-*.xml
          summary: true
          display-options: fEX
          fail-on-empty: true
          title: Test results
      - name: Minimize uv cache
        run: uv cache prune --ci

  # This final job is required to satisfy the merge queue. It must only run (or succeed) if no tests failed
  python-integration-tests-check:
    if: always()
    runs-on: ubuntu-latest
    strategy:
      max-parallel: 1
      fail-fast: false
    needs: [python-merge-gate, python-integration-tests]
    steps:
      - name: Get Date
        shell: bash
        run: |
          echo "date=$(date +'%m/%d/%Y %H:%M:%S')" >> "$GITHUB_ENV"

      - name: Run Type is Daily
        if: ${{ github.event_name == 'schedule' }}
        shell: bash
        run: |
          echo "run_type=Daily" >> "$GITHUB_ENV"

      - name: Run Type is Manual
        if: ${{ github.event_name == 'workflow_dispatch' }}
        shell: bash
        run: |
          echo "run_type=Manual" >> "$GITHUB_ENV"

      - name: Run Type is ${{ github.event_name }}
        if: ${{ github.event_name != 'schedule' && github.event_name != 'workflow_dispatch'}}
        shell: bash
        run: |
          echo "run_type=${{ github.event_name }}" >> "$GITHUB_ENV"

      - name: Fail workflow if tests failed
        id: check_tests_failed
        if: contains(join(needs.*.result, ','), 'failure')
        uses: actions/github-script@v6
        with:
          script: core.setFailed('Integration Tests Failed!')

      - name: Fail workflow if tests cancelled
        id: check_tests_cancelled
        if: contains(join(needs.*.result, ','), 'cancelled')
        uses: actions/github-script@v6
        with:
          script: core.setFailed('Integration Tests Cancelled!')

      - name: Microsoft Teams Notification
        uses: skitionek/notify-microsoft-teams@master
        if: github.ref == 'refs/heads/main' && github.event_name != 'pull_request'
        with:
          webhook_url: ${{ secrets.MSTEAMS_WEBHOOK }}
          dry_run: ${{ env.run_type != 'Daily' && env.run_type != 'Manual'}}
          job: ${{ toJson(job) }}
          steps: ${{ toJson(steps) }}
          overwrite: "{title: ` ${{ env.run_type }}: ${{ env.date }} `, text: ` ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}`}"

      - name: Microsoft Teams Notification (Dry Run)
        uses: skitionek/notify-microsoft-teams@master
        if: github.ref != 'refs/heads/main'
        with:
          webhook_url: NONE
          dry_run: ${{ env.run_type != 'Daily' && env.run_type != 'Manual'}}
          job: ${{ toJson(job) }}
          steps: ${{ toJson(steps) }}
          overwrite: "{title: ` ${{ env.run_type }}: ${{ env.date }} `, text: ` ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}`}"