book-preview-dev.yml

name: '📚 Book · 👁️ Preview (Dev)'

# =============================================================================
# 👁️ Preview (Dev) — Deploy dev build artifacts to staging site
# =============================================================================
#
# Downloads HTML/PDF/EPUB artifacts from a successful Validate run and deploys
# them to the dev preview site via SSH push to the target repository.
#
# Flow:
#   1. RESOLVE_RUN   — Find the correct Validate workflow run (auto or manual)
#   2. DOWNLOAD       — Pull Vol1 + Vol2 artifacts (HTML, PDF, EPUB)
#   3. ASSEMBLE_SITE  — Build preview site with volume chooser and downloads
#   4. DEPLOY         — Push to dev preview repo via SSH deploy key
#
# Triggers:
#   - workflow_run: on successful completion of Validate (Dev)
#   - workflow_dispatch: manual deploy of a specific commit or run ID
#
# Deploys to: Dev preview site (via SSH to DEV_REPO)
# Secrets:    GITHUB_TOKEN, SSH_DEPLOY_KEY
# Vars:       BOOK_TOOLS, DEV_REPO, DEV_REPO_URL,
#             VOL1_DEPLOY_PATH (top-level dir for Vol I, mirrors live),
#             VOL2_DEPLOY_PATH (top-level dir for Vol II, mirrors live)
#
# Related:
#   - book-validate-dev.yml — Produces the build artifacts this workflow deploys
#   - book-publish-live.yml — Production equivalent (deploys to gh-pages)
#
# =============================================================================

on:
  workflow_dispatch:
    inputs:
      commit_sha:
        description: 'Optional: Specific commit SHA to deploy. Deploys the latest successful `dev` build if empty. IMPORTANT: Select the `dev` branch before running.'
        required: false
        type: string
      run_id:
        description: 'Optional: Specific validate workflow run ID to pull artifacts from (bypasses success check).'
        required: false
        type: string
  workflow_run:
    workflows: ["📚 Book · ✅ Validate (Dev)"]
    types:
      - completed

concurrency:
  group: deploy-preview-${{ github.ref }}
  cancel-in-progress: true

jobs:
  deploy:
    name: '📺 Deploy to GitHub Pages (Dev Preview)'
    if: (github.event.workflow_run.conclusion == 'success' && github.event.workflow_run.head_branch == 'dev') || github.event_name == 'workflow_dispatch'
    runs-on: ubuntu-latest
    # Deploys dev branch to: https://harvard-edge.github.io/cs249r_book_dev/
    permissions:
      contents: write # Allow write access to repository
      pages: write
      actions: read # Allow reading of workflow runs

    steps:
      - name: ⬇️ Get workflow run ID
        id: run_info
        run: |
          if [[ "${{ github.event_name }}" == "workflow_run" ]]; then
            echo "run_id=${{ github.event.workflow_run.id }}" >> $GITHUB_OUTPUT
            echo "head_sha=${{ github.event.workflow_run.head_sha }}" >> $GITHUB_OUTPUT
            exit 0
          fi

          # If a specific run_id is provided, use it directly (bypasses success check)
          RUN_ID_INPUT=$(echo "${{ github.event.inputs.run_id }}" | xargs)
          if [[ -n "$RUN_ID_INPUT" ]]; then
            echo "📌 Using provided run ID: $RUN_ID_INPUT"
            echo "run_id=$RUN_ID_INPUT" >> $GITHUB_OUTPUT
            RUN_SHA=$(gh api repos/${{ github.repository }}/actions/runs/${RUN_ID_INPUT} --jq '.head_sha')
            echo "head_sha=$RUN_SHA" >> $GITHUB_OUTPUT
            exit 0
          fi

          COMMIT_SHA=$(echo "${{ github.event.inputs.commit_sha }}" | xargs)
          API_URL="repos/${{ github.repository }}/actions/workflows/book-validate-dev.yml/runs"

          if [[ -n "$COMMIT_SHA" ]]; then
            echo "🔎 Searching for successful 'Validate Dev' run for commit starting with: $COMMIT_SHA"
            JQ_QUERY=".workflow_runs[] | select((.head_sha | startswith(\"$COMMIT_SHA\")) and .status == \"completed\" and .conclusion == \"success\") | .id"
          else
            echo "🔎 No commit SHA provided. Searching for the latest successful 'Validate Dev' run on 'dev' branch."
            JQ_QUERY=".workflow_runs[] | select(.head_branch == \"dev\" and .status == \"completed\" and .conclusion == \"success\") | .id"
          fi

          latest_run_id=$(gh api "$API_URL" --jq "$JQ_QUERY" | head -n 1)

          if [ -z "$latest_run_id" ]; then
              if [[ -n "$COMMIT_SHA" ]]; then
                echo "::error::Could not find a successful 'Validate Dev' run for commit ${COMMIT_SHA}."
              else
                echo "::error::Could not find any successful 'Validate Dev' run on the 'dev' branch."
              fi
              exit 1
          fi

          echo "Found run ID: ${latest_run_id}"
          echo "run_id=${latest_run_id}" >> $GITHUB_OUTPUT

          latest_run_sha=$(gh api repos/${{ github.repository }}/actions/runs/${latest_run_id} --jq '.head_sha')
          echo "head_sha=${latest_run_sha}" >> $GITHUB_OUTPUT
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

      - name: 📥 Download Vol1 HTML Artifact
        uses: actions/download-artifact@v8
        with:
          name: dev-html-vol1-linux
          run-id: ${{ steps.run_info.outputs.run_id }}
          path: ./html-vol1-artifact
          github-token: ${{ secrets.GITHUB_TOKEN }}

      - name: 📥 Download Vol2 HTML Artifact
        uses: actions/download-artifact@v8
        with:
          name: dev-html-vol2-linux
          run-id: ${{ steps.run_info.outputs.run_id }}
          path: ./html-vol2-artifact
          github-token: ${{ secrets.GITHUB_TOKEN }}

      - name: 📥 Download Vol1 PDF Artifact
        uses: actions/download-artifact@v8
        continue-on-error: true
        with:
          name: dev-pdf-vol1-linux
          run-id: ${{ steps.run_info.outputs.run_id }}
          path: ./pdf-vol1-artifact
          github-token: ${{ secrets.GITHUB_TOKEN }}

      - name: 📥 Download Vol2 PDF Artifact
        uses: actions/download-artifact@v8
        continue-on-error: true
        with:
          name: dev-pdf-vol2-linux
          run-id: ${{ steps.run_info.outputs.run_id }}
          path: ./pdf-vol2-artifact
          github-token: ${{ secrets.GITHUB_TOKEN }}

      - name: 📥 Download Vol1 EPUB Artifact
        uses: actions/download-artifact@v8
        continue-on-error: true
        with:
          name: dev-epub-vol1-linux
          run-id: ${{ steps.run_info.outputs.run_id }}
          path: ./epub-vol1-artifact
          github-token: ${{ secrets.GITHUB_TOKEN }}

      - name: 📥 Download Vol2 EPUB Artifact
        uses: actions/download-artifact@v8
        continue-on-error: true
        with:
          name: dev-epub-vol2-linux
          run-id: ${{ steps.run_info.outputs.run_id }}
          path: ./epub-vol2-artifact
          github-token: ${{ secrets.GITHUB_TOKEN }}

      - name: 📦 Build volume preview site
        run: |
          echo "📦 Preparing volume preview site..."
          mkdir -p ./preview-site/vol1 ./preview-site/vol2

          echo "📋 Vol1 artifact contents:"
          ls -la ./html-vol1-artifact/ 2>/dev/null || echo "  (empty or missing)"
          echo "📋 Vol2 artifact contents:"
          ls -la ./html-vol2-artifact/ 2>/dev/null || echo "  (empty or missing)"

          # Use rsync-style copy that handles both flat and nested artifact layouts
          cp -r ./html-vol1-artifact/. ./preview-site/vol1/
          cp -r ./html-vol2-artifact/. ./preview-site/vol2/

          mkdir -p ./preview-site/vol1/assets/downloads ./preview-site/vol2/assets/downloads

          # Copy PDF/EPUB downloads by exact name (artifact dirs contain many files;
          # a wildcard find can pick up a tiny figure PDF instead of the book)
          VOL1_PDF=$(find ./pdf-vol1-artifact -name "Machine-Learning-Systems-Vol1.pdf" -type f 2>/dev/null | head -1)
          [ -n "$VOL1_PDF" ] && cp "$VOL1_PDF" "./preview-site/vol1/assets/downloads/Machine-Learning-Systems-Vol1.pdf" || true

          VOL1_EPUB=$(find ./epub-vol1-artifact -name "Machine-Learning-Systems-Vol1.epub" -type f 2>/dev/null | head -1)
          [ -n "$VOL1_EPUB" ] && cp "$VOL1_EPUB" "./preview-site/vol1/assets/downloads/Machine-Learning-Systems-Vol1.epub" || true

          VOL2_PDF=$(find ./pdf-vol2-artifact -name "Machine-Learning-Systems-Vol2.pdf" -type f 2>/dev/null | head -1)
          [ -n "$VOL2_PDF" ] && cp "$VOL2_PDF" "./preview-site/vol2/assets/downloads/Machine-Learning-Systems-Vol2.pdf" || true

          VOL2_EPUB=$(find ./epub-vol2-artifact -name "Machine-Learning-Systems-Vol2.epub" -type f 2>/dev/null | head -1)
          [ -n "$VOL2_EPUB" ] && cp "$VOL2_EPUB" "./preview-site/vol2/assets/downloads/Machine-Learning-Systems-Vol2.epub" || true

      - name: 📝 Create .nojekyll file
        run: touch ./preview-site/.nojekyll

      - name: 📦 Inspect final site contents
        run: |
          echo "::group::Site Contents"
          ls -la ./preview-site
          echo "::endgroup::"

          echo "::group::Vol1 Downloads"
          if [ -d "./preview-site/vol1/assets/downloads" ]; then
            ls -lh ./preview-site/vol1/assets/downloads
          else
            echo "⚠️ No vol1/assets/downloads directory"
          fi
          echo "::endgroup::"

          echo "::group::Vol2 Downloads"
          if [ -d "./preview-site/vol2/assets/downloads" ]; then
            ls -lh ./preview-site/vol2/assets/downloads
          else
            echo "⚠️ No vol2/assets/downloads directory"
          fi
          echo "::endgroup::"

          # Sanity check: PDFs should be >1MB (a figure PDF would be <100KB)
          for vol in vol1 vol2; do
            PDF="./preview-site/$vol/assets/downloads/Machine-Learning-Systems-$(echo $vol | sed 's/vol/Vol/').pdf"
            if [ -f "$PDF" ]; then
              SIZE=$(stat -c%s "$PDF" 2>/dev/null || stat -f%z "$PDF" 2>/dev/null)
              if [ "$SIZE" -lt 1000000 ]; then
                echo "::error::$vol PDF is only ${SIZE} bytes — likely a figure, not the book!"
              else
                echo "✅ $vol PDF looks correct: $(du -h "$PDF" | cut -f1)"
              fi
            fi
          done

      - name: ⬇️ Checkout repository (for scripts)
        uses: actions/checkout@v6
        with:
          ref: dev
          path: ./repo-scripts
          sparse-checkout: |
            ${{ vars.BOOK_TOOLS }}/scripts/publish/modify_dev_announcement.py
            .github/scripts/rewrite-dev-urls.sh
            .github/scripts/flatten-vol-urls.sh
          sparse-checkout-cone-mode: false

      - name: 🔧 Flatten volume URL structure
        # Moves contents/vol{N}/* to the site root so preview URLs are
        # devsite/vol{N}/{chapter}/{chapter}.html (not double-nested).
        # Must run before rewrite-dev-urls so depth calculations are correct.
        run: |
          if [ -d "./preview-site/vol1" ]; then
            bash ./repo-scripts/.github/scripts/flatten-vol-urls.sh ./preview-site/vol1 vol1
          fi
          if [ -d "./preview-site/vol2" ]; then
            bash ./repo-scripts/.github/scripts/flatten-vol-urls.sh ./preview-site/vol2 vol2
          fi

      - name: 🔗 Rewrite URLs for dev site
        run: |
          bash ./repo-scripts/.github/scripts/rewrite-dev-urls.sh vol1 ./preview-site/vol1 --live-external
          bash ./repo-scripts/.github/scripts/rewrite-dev-urls.sh vol2 ./preview-site/vol2 --live-external

      - name: 🔧 Modify announcement for dev preview
        run: |
          echo "🔧 Modifying announcement banner for development preview..."

          # Get commit info for display
          COMMIT_HASH="${{ steps.run_info.outputs.head_sha }}"
          COMMIT_SHORT="${COMMIT_HASH:0:8}"

          echo "📝 Adding dev preview banner with commit info: $COMMIT_SHORT"

          python3 ./repo-scripts/${{ vars.BOOK_TOOLS }}/scripts/publish/modify_dev_announcement.py \
            ./preview-site \
            --verbose \
            --commit-hash "$COMMIT_HASH" \
            --commit-short "$COMMIT_SHORT"

      - name: 🚀 Deploy to Dev Site via SSH
        env:
          SSH_DEPLOY_KEY: ${{ secrets.SSH_DEPLOY_KEY }}
        run: |
          echo "🔐 Starting ssh-agent..."
          eval "$(ssh-agent -s)"
          echo "$SSH_DEPLOY_KEY" | tr -d '\r' | ssh-add - > /dev/null
          # Add github.com to known hosts to avoid prompt. Port 22 intermittently
          # refuses connections from runners, so trust the 443 endpoint too and
          # keep it available as a fallback rather than losing a whole build.
          mkdir -p ~/.ssh
          ssh-keyscan github.com >> ~/.ssh/known_hosts 2>/dev/null || true
          ssh-keyscan -p 443 ssh.github.com >> ~/.ssh/known_hosts

          echo "🔧 Configuring git..."
          git config --global user.email "actions@github.com"
          git config --global user.name "GitHub Actions"

          echo "🔄 Cloning target repository (${{ vars.DEV_REPO }})..."
          clone_target() {
            git clone --depth=1 "${{ vars.DEV_REPO_URL }}" target-repo
          }
          cloned=false
          for attempt in 1 2 3; do
            if clone_target; then
              cloned=true
              break
            fi
            rm -rf target-repo
            echo "⚠️  Clone attempt ${attempt} failed."
            if [ "$attempt" -eq 2 ]; then
              echo "🔁 Falling back to ssh.github.com:443 for the final attempt."
              printf 'Host github.com\n  Hostname ssh.github.com\n  Port 443\n  User git\n' >> ~/.ssh/config
            fi
            sleep $((attempt * 10))
          done
          if [ "$cloned" != true ]; then
            echo "❌ Could not reach the dev site repository over SSH after 3 attempts."
            exit 1
          fi

          echo "🔄 Entering target repository..."
          cd target-repo

          # Per-volume top-level deploy paths. Mirrors book-publish-live.yml so
          # /vol1/ and /vol2/ resolve identically on dev preview and live.
          VOL1_PATH="${{ vars.VOL1_DEPLOY_PATH }}"
          VOL2_PATH="${{ vars.VOL2_DEPLOY_PATH }}"
          if [ -z "$VOL1_PATH" ] || [ -z "$VOL2_PATH" ]; then
            echo "❌ vars.VOL1_DEPLOY_PATH or vars.VOL2_DEPLOY_PATH is unset — refusing to run destructive rm with empty path."
            exit 1
          fi

          echo "🧹 Cleaning /$VOL1_PATH/ and /$VOL2_PATH/ subdirectories (preserving other subsites)..."
          rm -rf "$VOL1_PATH" "$VOL2_PATH"
          # Also clean up the legacy /book/ wrapper from the previous nested layout.
          # Keeping it would leave zombie content at harvard-edge.github.io/cs249r_book_dev/book/.
          rm -rf book
          mkdir -p "$VOL1_PATH" "$VOL2_PATH"

          echo "🚚 Copying Vol1 content to /$VOL1_PATH/..."
          cp -r "${{ github.workspace }}/preview-site/vol1/." "$VOL1_PATH/"
          echo "🚚 Copying Vol2 content to /$VOL2_PATH/..."
          cp -r "${{ github.workspace }}/preview-site/vol2/." "$VOL2_PATH/"
          touch .nojekyll

          echo "🔍 Validating deployment content..."
          if [ ! -f "$VOL1_PATH/index.html" ]; then
            echo "❌ CRITICAL: $VOL1_PATH/index.html is missing after copy. Aborting deployment."
            exit 1
          fi

          if [ ! -f "$VOL2_PATH/index.html" ]; then
            echo "❌ CRITICAL: $VOL2_PATH/index.html is missing after copy. Aborting deployment."
            exit 1
          fi

          echo "📦 Repository structure:"
          ls -la
          echo ""
          echo "📦 /$VOL1_PATH/ contents:"
          ls -la "$VOL1_PATH/" | head -10
          echo ""
          echo "📦 /$VOL2_PATH/ contents:"
          ls -la "$VOL2_PATH/" | head -10

          echo "📦 Committing and pushing changes..."
          git add .
          git commit -m "🚀 Deploy book dev preview to /$VOL1_PATH/ + /$VOL2_PATH/ from ${{ steps.run_info.outputs.head_sha }}" --allow-empty || echo "🟡 Nothing to commit"

          # Retry push with rebase to handle concurrent deploys
          for i in 1 2 3; do
            if git push origin main 2>/dev/null; then
              echo "✅ Push succeeded on attempt $i"
              break
            fi
            echo "⚠️ Push failed (attempt $i/3), pulling with rebase..."
            git pull --rebase origin main || true
          done

          # DEV_REPO contains org/repo format, extract just repo name for GitHub Pages URL
          REPO_NAME="${{ vars.DEV_REPO }}"
          REPO_NAME="${REPO_NAME##*/}"
          echo "✅ Volumes deployed to:"
          echo "   https://harvard-edge.github.io/${REPO_NAME}/$VOL1_PATH/"
          echo "   https://harvard-edge.github.io/${REPO_NAME}/$VOL2_PATH/"