sync-S3-KB.yml

name: syncS3andKB
on:
  push:
    branches: ["main"]
  workflow_dispatch:
    inputs:
      sdk_name:
        description: 'SDK Name'
        required: true
        default: 'python'
        type: choice
        options:
          - javascriptv3
          - dotnetv4
          - javav2
          - rustv1
          - gov2
          - swift
          - python
          - ruby
          - php
          - cpp
          - kotlin
          - steering_docs
          - specs
          - coding-standards

permissions:
  id-token: write

jobs:
  run_job_with_aws:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        sdk_name: ${{ github.event_name == 'push' && fromJSON('["javascriptv3","dotnetv4","javav2","rustv1","gov2","swift","python","ruby","php","cpp","kotlin","steering_docs","specs","coding-standards"]') || fromJSON(format('["{0}"]', github.event.inputs.sdk_name)) }}
    env:
      sdk_name: ${{ matrix.sdk_name }}
    
    steps:
      - name: Checkout
        uses: actions/checkout@v4

      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@v5 
        with:
          role-to-assume: ${{ secrets.AWS_ASSUME_ROLE }}     # once merged, update trust policy of the role to point to main branch
          aws-region: us-west-2

      - name: Set SDK and language mapping for S3
        run: |
          if [ "$sdk_name" == "javascriptv3" ]; then
              echo "S3_LANGUAGE=javascript" >> $GITHUB_ENV
          elif [ "$sdk_name" == "dotnetv4" ]; then
              echo "S3_LANGUAGE=dotnet" >> $GITHUB_ENV
          elif [ "$sdk_name" == "javav2" ]; then
              echo "S3_LANGUAGE=java" >> $GITHUB_ENV
          elif [ "$sdk_name" == "rustv1" ]; then
              echo "S3_LANGUAGE=rust" >> $GITHUB_ENV
          elif [ "$sdk_name" == "gov2" ]; then
              echo "S3_LANGUAGE=go" >> $GITHUB_ENV
          elif [ "$sdk_name" == "steering_docs" ]; then
              echo "S3_LANGUAGE=steering-docs" >> $GITHUB_ENV
          elif [ "$sdk_name" == "coding-standards" ]; then
              echo "S3_LANGUAGE=coding-standards" >> $GITHUB_ENV
          elif [ "$sdk_name" == "specs" ]; then
              echo "S3_LANGUAGE=final-specs" >> $GITHUB_ENV
          else
              echo "S3_LANGUAGE=$sdk_name" >> $GITHUB_ENV
          fi
    
      - name: Filter SPECIFICATION.md files for specs
        if: ${{ github.event.inputs.sdk_name == 'specs' }}
        run: |
          find ./scenarios -name "SPECIFICATION.md" | while read file; do
            mkdir -p "./filtered_specs/$(dirname "$file")"
            cp "$file" "./filtered_specs/$file"
          done
      
      - name: Clone and filter for coding standards
        if: ${{ github.event.inputs.sdk_name == 'coding-standards' }}
        run: |
          git clone https://x-access-token:${{ secrets.GITHUB_TOKEN }}@github.com/${{ github.repository }}.wiki.git wiki-repo
          mkdir -p ./filtered-wiki
          find ./wiki-repo -type f -name "*[Gg]uidelines*.md" -o -name "*[Ss]tandards*.md" | while read file; do
            cp "$file" ./filtered-wiki/
          done
      
      - name: Extract and copy premium examples in temp. dir.
        if: ${{ contains(fromJSON('["javascriptv3","dotnetv4","javav2","rustv1","gov2","swift","python","ruby","php","cpp","kotlin"]'), github.event.inputs.sdk_name) }}
        run: |
          MARKDOWN_FILE="./$sdk_name/premium-ex.md"
          
          if [ ! -f "$MARKDOWN_FILE" ]; then
            echo "Premium examples file not found: $MARKDOWN_FILE"
            exit 1
          fi

          extract_paths() {
            local level="$1"
            local section_found=false
            
            while IFS= read -r line; do
              if [[ "$line" =~ ^##[[:space:]]*${level}:[[:space:]]*$ ]]; then
                section_found=true
                continue
              elif [[ "$line" =~ ^##[[:space:]] ]] && [ "$section_found" = true ]; then
                break
              elif [ "$section_found" = true ] && [[ "$line" =~ ^/ ]]; then
                echo "$line"
              fi
            done < "$MARKDOWN_FILE"
          }
          
          for level in "basics" "feature-scenario" "complex-feature-scenario"; do
            paths=$(extract_paths "$level")
            
            if [ -n "$paths" ]; then
              mkdir -p "./extracted_snippets/$level"
              
              while IFS= read -r path; do
                if [ -n "$path" ]; then
                  source_path="./$sdk_name$path"
                  if [ -e "$source_path" ]; then
                    cp -r "$source_path" "./extracted_snippets/$level/"
                  fi
                fi
              done <<< "$paths"
            fi
          done

      - name: Upload/Sync to S3 (SDK languages)
        if: ${{ contains(fromJSON('["javascriptv3","dotnetv4","javav2","rustv1","gov2","swift","python","ruby","php","cpp","kotlin"]'), github.event.inputs.sdk_name) }}
        run: |
          for level in "basics" "feature-scenario" "complex-feature-scenario"; do
            if [ -d "./extracted_snippets/$level" ]; then
              aws s3 sync "./extracted_snippets/$level/" "s3://$S3_LANGUAGE-premium-bucket/$level/" --delete
              echo "Uploaded $level examples to S3"
            fi
          done
      
      - name: Upload/Sync to S3 (Other directories)
        if: ${{ contains(fromJSON('["steering_docs","coding-standards","specs"]'), github.event.inputs.sdk_name) }}
        run: |
          if [ "$sdk_name" == "steering_docs" ]; then
            aws s3 sync "./$sdk_name/" "s3://$S3_LANGUAGE-bucket/" --delete
          elif [ "$sdk_name" == "coding-standards" ]; then
            aws s3 sync "./filtered-wiki/" "s3://$S3_LANGUAGE-bucket/" --delete
          else
            aws s3 sync "./filtered_specs/" "s3://$S3_LANGUAGE-bucket/" --delete
          fi

      - name: Sync Knowledge Base Data Source
        run: |
          aws lambda invoke \
            --function-name KB_Updater \
            --payload "{\"language\":\"$S3_LANGUAGE\",\"region\":\"us-west-2\"}" \
            --cli-binary-format raw-in-base64-out \
            response.json
          
          echo "Knowledge Base sync initiated"
          cat response.json