combined-docs-md-ci.yml

name: Generate PDF Documentation

on:
  pull_request:
    branches:
      - "main"
      - "pdf-documentation"
  workflow_dispatch:
    inputs:
      folder_path:
        description: "Path to the documentation folder"
        required: true
        default: "./docs"
      sidebar_path:
        description: "Path to the sidebar.js file"
        required: true
        default: "./sidebars.js"
      combined_md_file_path:
        description: "Path to the combined docs markdown file"
        required: true
        default: "./combined_docs.md"

permissions:
  id-token: write
  contents: read

jobs:
  generate_docs:
    runs-on: ubuntu-latest
    outputs:
      pdf_url: ${{ steps.s3_upload.outputs.s3_url }} # Output the S3 bucket URL where the generated PDF is stored

    steps:
      - name: Checkout repository
        uses: actions/checkout@v6

      - name: Install Node.js
        uses: actions/setup-node@v4
        with:
          node-version: "20" # Specify the Node.js version you need

      - name: Install Python
        run: |
          sudo apt-get update
          sudo apt-get install -y python3 python3-pip

      - name: Install LaTeX, Pandoc, and Required Packages
        run: |
          sudo apt-get install -y pandoc
          sudo apt-get install texlive-latex-base
          sudo apt-get install texlive-luatex
          sudo apt-get install lmodern
          sudo apt-get install texlive-fonts-recommended
          sudo apt-get install texlive-fonts-extra
          sudo apt-get install texlive-latex-extra

      - name: Install Ghostscript
        run: |
          sudo apt-get install -y ghostscript

      - name: Verify Ghostscript Installation
        run: |
          gs --version
          which gs
          echo "Ghostscript installation verified"

      - name: Set environment variables
        env:
          FOLDER_PATH: ${{ github.event.inputs.folder_path || './docs' }} # For manual runs or PR body
          SIDEBAR_PATH: ${{ github.event.inputs.sidebar_path || './sidebars.js' }} # For manual runs or PR body
          COMBINED_DOC_PATH: ${{ github.event.inputs.combined_md_file_path || './combined_docs.md' }} # For manual runs or PR body
          BRANCH_NAME: ${{ github.head_ref || github.ref }}
        run: |
          echo "FOLDER_PATH=${FOLDER_PATH}" >> $GITHUB_ENV
          echo "SIDEBAR_PATH=${SIDEBAR_PATH}" >> $GITHUB_ENV
          echo "COMBINED_DOC_PATH=${COMBINED_DOC_PATH}" >> $GITHUB_ENV
          echo "BRANCH_NAME=${BRANCH_NAME}" >> $GITHUB_ENV

      # Fallback: Set the repository name as the documentation name if no title is found
      - name: Set repository name as documentation name fallback
        run:
          | # Extracts the repository name (everything after the slash in org/repo)
          echo "REPO_NAME=$(echo $GITHUB_REPOSITORY | cut -d'/' -f2)" >> $GITHUB_ENV

      # Extract documentation title and determine the version number dynamically
      - name: Extract title and version
        run: |
          DOC_TITLE=$(node -p "require('./documentation/makersaurus.config.js').title || ''")

          if [[ "${{ env.FOLDER_PATH }}" == "./docs" ]]; then
            VERSION=$(node -p "require('./documentation/makersaurus.config.js').versions?.current?.label || '0.0.0'")
          else
            VERSION=$(basename "${{ env.FOLDER_PATH }}" | sed 's/version-//')
          fi

          SAFE_TITLE=$(echo "$DOC_TITLE" | tr ' ' '_' | tr -dc 'A-Za-z0-9._-')

          echo "DOC_TITLE=${DOC_TITLE}" >> $GITHUB_ENV
          echo "VERSION=${VERSION}" >> $GITHUB_ENV
          echo "SAFE_TITLE=${SAFE_TITLE}" >> $GITHUB_ENV

      # Download the script from another repository (replace with correct repo and script path)
      - name: Download script and make it executable
        run: |
          curl -H "Authorization: bearer ${{ secrets.GH_ACTIONS_RUNNERS_H2O_OPS_TOKEN }}" -L -o ./documentation/generate_combined_md.py "https://raw.githubusercontent.com/h2oai/makersaurus/refs/heads/main/documentation/generate_combined_md.py"
          curl -H "Authorization: bearer ${{ secrets.GH_ACTIONS_RUNNERS_H2O_OPS_TOKEN }}" -L -o ./documentation/generate_pdf_doc.py "https://raw.githubusercontent.com/h2oai/makersaurus/refs/heads/main/documentation/generate_pdf_doc.py"
          curl -H "Authorization: bearer ${{ secrets.GH_ACTIONS_RUNNERS_H2O_OPS_TOKEN }}" -L -o ./documentation/parse_sidebar.js "https://raw.githubusercontent.com/h2oai/makersaurus/refs/heads/main/documentation/parse_sidebar.js"
          chmod +x ./documentation/generate_combined_md.py
          chmod +x ./documentation/generate_pdf_doc.py
          chmod +x ./documentation/parse_sidebar.js

      # Generate either a combined markdown file or a PDF depending on the branch
      - name: Generate combined markdown or PDF
        run: |
          cd documentation
          if [[ "${{ env.BRANCH_NAME }}" =~ ^refs/heads/pdf-documentation ]]; then
            python3 generate_pdf_doc.py "${{ env.FOLDER_PATH }}" "${{ env.SIDEBAR_PATH }}" "${{ env.COMBINED_DOC_PATH }}"
          else
            python3 generate_combined_md.py "${{ env.FOLDER_PATH }}" "${{ env.SIDEBAR_PATH }}"
          fi

      # Rename the generated output files to include the documentation title and version
      - name: Rename output files
        run: |
          cd documentation
          if [[ -f documentation.pdf ]]; then
            mv documentation.pdf "${{ env.SAFE_TITLE }}_${{ env.VERSION }}.pdf"
          fi
          if [[ -f combined_docs.md ]]; then
            mv combined_docs.md "${{ env.SAFE_TITLE }}_${{ env.VERSION }}.md"
          fi

      - name: Upload generated documentation
        uses: actions/upload-artifact@v4
        with:
          name: documentation_artifacts
          path: |
            documentation/${{ env.SAFE_TITLE }}_${{ env.VERSION }}.md
            documentation/${{ env.SAFE_TITLE }}_${{ env.VERSION }}.pdf

      - name: Configure AWS credentials
        uses: aws-actions/configure-aws-credentials@v4
        with:
          role-to-assume: arn:aws:iam::905418351423:role/Github-OIDC-Role-h2o-llmstudio
          role-session-name: h2o-llm-studio
          aws-region: us-east-1

      - name: Publish documentation to S3
        id: s3_upload
        run: |
          if [[ "${{ env.BRANCH_NAME }}" =~ ^refs/heads/pdf-documentation ]]; then
            aws s3 cp "documentation/${{ env.SAFE_TITLE }}_${{ env.VERSION }}.pdf" s3://pdf-documentation/h2o-llm-studio/
            
            S3_BUCKET="pdf-documentation"
            S3_KEY="h2o-llm-studio/${{ env.SAFE_TITLE }}_${{ env.VERSION }}.pdf"
            S3_URL="https://${S3_BUCKET}.s3.amazonaws.com/${S3_KEY}"
            
            echo "s3_url=${S3_URL}" >> $GITHUB_OUTPUT
            
            echo "PDF uploaded successfully to: ${S3_URL}"
          else
            echo "Skipping S3 upload - not on pdf-documentation branch"
          fi

      - name: Display PDF URL
        if: steps.s3_upload.outputs.s3_url != ''
        run: |
          echo "PDF is available at: ${{ steps.s3_upload.outputs.s3_url }}"
          echo "pdf_url=${{ steps.s3_upload.outputs.s3_url }}" >> $GITHUB_ENV