ci.yml

name: MLOps Pipeline (Langfuse v3)
run-name: "CI Check - wandb-integration"
on:
  push:
    branches: [ "main" ]
  pull_request:
    branches: [ "main" ]

jobs:
  build-and-test:
    runs-on: ubuntu-latest

    services:
      qdrant:
        image: qdrant/qdrant:latest
        ports:
          - 6333:6333

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

    - name: Set up Python 3.11
      uses: actions/setup-python@v5
      with:
        python-version: "3.11"
        cache: 'pip'

    - name: Install System Dependencies
      run: |
        sudo apt-get update
        sudo apt-get install -y curl build-essential

    - name: Install Python Dependencies
      run: |
        python -m pip install --upgrade pip
        # Install PyTorch CPU version first (faster, no GPU needed in CI)
        pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
        # Install all other dependencies
        pip install -r requirements.txt
        # Install dev dependencies
        pip install ruff pytest pytest-cov

    - name: Lint with Ruff
      run: |
        ruff check . --fix
      continue-on-error: true

    - name: Component Initialization Check
      env:
        GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}
        LANGFUSE_PUBLIC_KEY: ${{ secrets.LANGFUSE_PUBLIC_KEY }}
        LANGFUSE_SECRET_KEY: ${{ secrets.LANGFUSE_SECRET_KEY }}
        LANGFUSE_HOST: "https://cloud.langfuse.com"
        QDRANT_URL: "http://localhost:6333"
        PYTHONPATH: .
      run: |
        echo "🔍 Testing RAG Engine initialization..."
        # Невелика пауза для впевненості, що контейнер прокинувся
        sleep 5
        python -c "
        import sys
        try:
            from src.rag import engine
            print('✅ RAG Engine initialized successfully')
            print(f'✅ Collection: {engine.client.get_collection(\"rag_documents\")}')
        except Exception as e:
            print(f'❌ Initialization failed: {e}')
            sys.exit(1)
        "

    - name: Ingest Test Data
      env:
        GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}
        QDRANT_URL: "http://localhost:6333"
        PYTHONPATH: .
      run: |
        echo "📄 Ingesting test data..."
        python -c "
        from src.rag import engine
        from langchain.schema import Document
        
        test_docs = [
            Document(
                page_content='PDF was created by Dr. John Warnock at Adobe. The ISO standard is ISO 32000.',
                metadata={'source': 'test_doc.pdf', 'page': 1, 'doc_hash': 'test123'}
            ),
            Document(
                page_content='PostScript is the programming language that PDF is based on.',
                metadata={'source': 'test_doc.pdf', 'page': 2, 'doc_hash': 'test456'}
            )
        ]
        
        engine.vector_store.add_documents(test_docs)
        
        # Verify ingestion
        count = engine.client.count('rag_documents').count
        print(f'✅ Test data ingested successfully. Total documents: {count}')
        
        if count < 2:
            raise Exception('Expected at least 2 documents in the collection')
        "

    - name: Test RAG Query (Smoke Test)
      env:
        GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}
        LANGFUSE_PUBLIC_KEY: ${{ secrets.LANGFUSE_PUBLIC_KEY }}
        LANGFUSE_SECRET_KEY: ${{ secrets.LANGFUSE_SECRET_KEY }}
        QDRANT_URL: "http://localhost:6333"
        PYTHONPATH: .
      run: |
        echo "🧪 Running smoke test query..."
        python -c "
        from src.rag import engine
        
        # Simple query test (skip if API key missing)
        try:
            answer, sources, trace_id = engine.get_answer_with_sources('Who created PDF?')
            print(f'✅ Query test passed')
            print(f'Answer preview: {answer[:100]}...')
            print(f'Sources found: {len(sources)}')
            print(f'Trace ID: {trace_id}')
        except Exception as e:
            print(f'⚠️ Query test skipped or failed: {e}')
        "
      continue-on-error: true

    - name: Run RAG Evaluation
      if: github.event_name == 'push' && github.ref == 'refs/heads/main'
      env:
        GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}
        LANGFUSE_PUBLIC_KEY: ${{ secrets.LANGFUSE_PUBLIC_KEY }}
        LANGFUSE_SECRET_KEY: ${{ secrets.LANGFUSE_SECRET_KEY }}
        QDRANT_URL: "http://localhost:6333"
        PYTHONPATH: .
      run: |
        if [ -z "$GROQ_API_KEY" ]; then
          echo "⚠️ GROQ_API_KEY not set. Skipping evaluation."
          exit 0
        fi
        
        echo "📊 Running Ragas evaluation..."
        
        # Try to find evaluation script
        if [ -f "evaluation/evaluate.py" ]; then
            python evaluation/evaluate.py
        elif [ -f "evaluation/run_eval.py" ]; then
            python evaluation/run_eval.py
        else
            echo "⚠️ Evaluation script not found. Skipping."
            exit 0
        fi
      continue-on-error: true

    - name: Upload Evaluation Report
      uses: actions/upload-artifact@v4
      if: always()
      with:
        name: rag-evaluation-report
        path: |
          evaluation/report.csv
          evaluation/*.csv
        retention-days: 30

    - name: Test FastAPI Endpoints (Integration Test)
      env:
        GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}
        LANGFUSE_PUBLIC_KEY: ${{ secrets.LANGFUSE_PUBLIC_KEY }}
        LANGFUSE_SECRET_KEY: ${{ secrets.LANGFUSE_SECRET_KEY }}
        QDRANT_URL: "http://localhost:6333"
        PYTHONPATH: .
      run: |
        echo "🧪 Testing FastAPI endpoints..."
        
        # Start API in background
        uvicorn src.main:app --host 0.0.0.0 --port 8000 &
        API_PID=$!
        
        # Wait for API to start
        sleep 10
        
        # Test health endpoint
        curl -f http://localhost:8000/health || {
          echo "❌ Health check failed"
          kill $API_PID
          exit 1
        }
        
        echo "✅ FastAPI health check passed"
        
        # Cleanup
        kill $API_PID
      continue-on-error: true

    - name: Cleanup Test Data
      if: always()
      env:
        QDRANT_URL: "http://localhost:6333"
      run: |
        echo "🧹 Cleaning up test data..."
        curl -X DELETE "http://localhost:6333/collections/rag_documents" || true
        echo "✅ Cleanup complete"

    - name: Summary
      if: always()
      run: |
        echo "================================"
        echo "📋 CI Pipeline Summary"
        echo "================================"
        echo "✅ CI process finished"
        echo "================================"