buildspec-unified.yml

version: 0.2

phases:
  install:
    runtime-versions:
      python: 3.12
      nodejs: 18
    commands:
      - echo "Installing dependencies..."
      - npm install -g aws-cdk@latest
      - pip install --upgrade pip
      - echo "Deployment Type is $DEPLOYMENT_TYPE"
      
  pre_build:
    commands:
      - echo "Pre-build phase started on $(date)"
      - echo "Deployment Type is $DEPLOYMENT_TYPE"
      - aws --version
      - cdk --version
      - |
        if [ "$DEPLOYMENT_TYPE" = "pdf2html" ]; then
          echo "PDF-to-HTML pre-build steps..."
          echo "Environment variables:"
          echo "  ACCOUNT_ID=$ACCOUNT_ID"
          echo "  REGION=$REGION"
          echo "  BUCKET_NAME=$BUCKET_NAME"
          echo "  BDA_PROJECT_ARN=$BDA_PROJECT_ARN"
          docker --version
          echo "Current directory contents:"
          ls -la
          echo "Navigating to pdf2html directory:"
          cd pdf2html
          ls -la
          echo "Logging in to Amazon ECR..."
          if ! aws ecr get-login-password --region $REGION | docker login --username AWS --password-stdin $ACCOUNT_ID.dkr.ecr.$REGION.amazonaws.com; then
            echo "ERROR: Initial ECR login failed!"
            echo "AWS CLI version:"
            aws --version
            echo "Region: $REGION"
            echo "Account ID: $ACCOUNT_ID"
            exit 1
          fi
          echo "Initial ECR login successful."
          cd ..
        else
          echo "PDF-to-PDF pre-build steps..."
          echo "Installing Python dependencies..."
          pip install -r requirements.txt
          echo "Bootstrapping CDK environment with retry logic..."
          for i in {1..3}; do
            echo "CDK bootstrap attempt $i/3..."
            if cdk bootstrap; then
              echo "CDK bootstrap successful on attempt $i"
              break
            else
              echo "CDK bootstrap failed on attempt $i"
              if [ $i -eq 3 ]; then
                echo "All CDK bootstrap attempts failed, continuing anyway..."
              else
                echo "Waiting 15 seconds before retry..."
                sleep 15
              fi
            fi
          done
        fi
      
  build:
    commands:
      - echo "Build phase started on $(date)"
      - |
        if [ "$DEPLOYMENT_TYPE" = "pdf2html" ]; then
          echo "Deploying PDF-to-HTML Remediation Solution..."
          cd pdf2html
          
          # Create S3 bucket if it doesn't exist
          if ! aws s3api head-bucket --bucket $BUCKET_NAME 2>/dev/null; then
            echo "Creating S3 bucket $BUCKET_NAME..."
            if [ "$REGION" = "us-east-1" ]; then
              aws s3api create-bucket --bucket $BUCKET_NAME --region $REGION
            else
              aws s3api create-bucket --bucket $BUCKET_NAME --region $REGION --create-bucket-configuration LocationConstraint=$REGION
            fi
            aws s3api put-bucket-versioning --bucket $BUCKET_NAME --versioning-configuration Status=Enabled
            
            aws s3api put-object --bucket $BUCKET_NAME --key uploads/
            aws s3api put-object --bucket $BUCKET_NAME --key output/
            aws s3api put-object --bucket $BUCKET_NAME --key remediated/
            echo "S3 bucket created successfully."
          else
            echo "S3 bucket $BUCKET_NAME already exists."
          fi
          
          # Create ECR repository if it doesn't exist
          REPO_NAME="pdf2html-lambda"
          if ! aws ecr describe-repositories --repository-names $REPO_NAME --region $REGION 2>/dev/null; then
            echo "Creating ECR repository $REPO_NAME..."
            aws ecr create-repository --repository-name $REPO_NAME --region $REGION
            echo "ECR repository created successfully."
          else
            echo "ECR repository $REPO_NAME already exists."
          fi
          
          # Build and push Docker image with proper error handling
          echo "Building Docker image..."
          REPO_URI="$ACCOUNT_ID.dkr.ecr.$REGION.amazonaws.com/pdf2html-lambda"
          
          # Build Docker image with error checking
          if ! docker build --platform linux/amd64 --no-cache -t $REPO_URI:latest .; then
            echo "ERROR: Docker build failed!"
            echo "Build context contents:"
            ls -la
            echo "Docker version:"
            docker --version
            exit 1
          fi
          echo "Docker build completed successfully."
          
          # Refresh ECR login right before push to prevent token expiry
          echo "Refreshing ECR login..."
          if ! aws ecr get-login-password --region $REGION | docker login --username AWS --password-stdin $ACCOUNT_ID.dkr.ecr.$REGION.amazonaws.com; then
            echo "ERROR: ECR login failed!"
            exit 1
          fi
          echo "ECR login successful."
          
          # Push Docker image with error checking and retry logic
          echo "Pushing Docker image to ECR..."
          if ! docker push $REPO_URI:latest; then
            echo "ERROR: Docker push failed on first attempt!"
            echo "Checking Docker images:"
            docker images | grep pdf2html-lambda
            echo "Retrying ECR login and push..."
            aws ecr get-login-password --region $REGION | docker login --username AWS --password-stdin $ACCOUNT_ID.dkr.ecr.$REGION.amazonaws.com
            if ! docker push $REPO_URI:latest; then
              echo "ERROR: Docker push failed after retry!"
              echo "ECR repository status:"
              aws ecr describe-repositories --repository-names pdf2html-lambda --region $REGION || echo "Repository not found"
              exit 1
            fi
          fi
          echo "Docker image pushed successfully."
          
          # Verify image exists in ECR
          echo "Verifying image in ECR..."
          if ! aws ecr describe-images --repository-name pdf2html-lambda --region $REGION --image-ids imageTag=latest; then
            echo "ERROR: Image not found in ECR after push!"
            echo "Available images in repository:"
            aws ecr list-images --repository-name pdf2html-lambda --region $REGION || echo "No images found"
            exit 1
          fi
          echo "Image verified successfully in ECR."
          
          # Bootstrap CDK and deploy
          echo "Bootstrapping CDK..."
          cd cdk
          npm install
          
          # Set CDK environment variables to ensure region consistency
          export CDK_DEFAULT_ACCOUNT=$ACCOUNT_ID
          export CDK_DEFAULT_REGION=$REGION
          
          npx cdk bootstrap aws://$ACCOUNT_ID/$REGION
          echo "Deploying CDK stack..."
          npx cdk deploy --app "node bin/app.js" --parameters BdaProjectArn=$BDA_PROJECT_ARN --parameters BucketName=$BUCKET_NAME --require-approval never
          
        elif [ "$DEPLOYMENT_TYPE" = "pdf2pdf" ]; then
          echo "Deploying PDF-to-PDF Remediation Solution..."
          echo "Deploying CDK stacks with retry logic..."
          for i in {1..3}; do
            echo "CDK deploy attempt $i/3..."
            if cdk deploy --all --require-approval never; then
              echo "CDK deploy successful on attempt $i"
              break
            else
              echo "CDK deploy failed on attempt $i"
              if [ $i -eq 3 ]; then
                echo "All CDK deploy attempts failed"
                exit 1
              fi
              echo "Waiting 30 seconds before retry..."
              sleep 30
            fi
          done
          
        else
          echo "Invalid DEPLOYMENT_TYPE: $DEPLOYMENT_TYPE"
          echo "Valid options are: pdf2pdf, pdf2html"
          exit 1
        fi
      
  post_build:
    commands:
      - echo "Post-build phase completed on $(date)"
      - |
        if [ "$DEPLOYMENT_TYPE" = "pdf2html" ]; then
          echo "PDF-to-HTML deployment completed successfully!"
          echo "S3 Bucket: $BUCKET_NAME"
          echo "BDA Project: $BDA_PROJECT_ARN"
        else
          echo "PDF-to-PDF deployment completed successfully!"
          echo "Check the CloudWatch dashboard for monitoring."
        fi