sync-stream-to-lake.yml

name: Sync Stream to Lake

on:
  schedule:
    - cron: '0 */4 * * *'  # Run every 4 hours
  workflow_dispatch:  # Allow manual trigger

permissions:
  contents: write
  pull-requests: write

jobs:
  lake-gate:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout repository
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
          token: ${{ secrets.GITHUB_TOKEN }}

      - name: Configure Git
        run: |
          git config user.name "github-actions[bot]"
          git config user.email "github-actions[bot]@users.noreply.github.com"

      - name: Check for differences between main and stable
        id: check_diff
        run: |
          git fetch origin main stable

          # Get commits in main but not in stable
          COMMITS=$(git log origin/stable..origin/main --oneline)

          if [ -z "$COMMITS" ]; then
            echo "No new commits to sync"
            echo "has_diff=false" >> $GITHUB_OUTPUT
            exit 0
          fi

          echo "has_diff=true" >> $GITHUB_OUTPUT

          # Get latest commit info
          LATEST_COMMIT=$(git rev-parse origin/main)
          LATEST_COMMIT_SHORT=$(git rev-parse --short origin/main)
          LATEST_COMMIT_MSG=$(git log -1 --pretty=%B origin/main)

          echo "latest_commit=${LATEST_COMMIT}" >> $GITHUB_OUTPUT
          echo "latest_commit_short=${LATEST_COMMIT_SHORT}" >> $GITHUB_OUTPUT
          echo "latest_commit_msg<<EOF" >> $GITHUB_OUTPUT
          echo "$LATEST_COMMIT_MSG" >> $GITHUB_OUTPUT
          echo "EOF" >> $GITHUB_OUTPUT

          echo "Commits to sync:"
          echo "$COMMITS"

      - name: Check for existing lake-gate PR
        if: steps.check_diff.outputs.has_diff == 'true'
        id: check_pr
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          # Find open PRs with lake-gate label
          PR_LIST=$(gh pr list --label "lake-gate" --state open --json number,headRefName,title,body)

          if [ "$PR_LIST" = "[]" ]; then
            echo "No existing lake-gate PR found"
            echo "pr_exists=false" >> $GITHUB_OUTPUT
            exit 0
          fi

          # Check if existing PR contains the latest commit
          LATEST_COMMIT="${{ steps.check_diff.outputs.latest_commit }}"
          PR_NUMBER=$(echo "$PR_LIST" | jq -r '.[0].number')
          PR_BRANCH=$(echo "$PR_LIST" | jq -r '.[0].headRefName')

          echo "Found existing PR #${PR_NUMBER} with branch ${PR_BRANCH}"

          # Fetch the PR branch and check if it contains the latest commit
          git fetch origin "$PR_BRANCH"

          if git merge-base --is-ancestor "$LATEST_COMMIT" "origin/$PR_BRANCH"; then
            echo "Existing PR already contains the latest commit"
            echo "pr_exists=true" >> $GITHUB_OUTPUT
            echo "pr_current=true" >> $GITHUB_OUTPUT
            echo "pr_number=${PR_NUMBER}" >> $GITHUB_OUTPUT
          else
            echo "Existing PR does not contain the latest commit - will create new PR"
            echo "pr_exists=true" >> $GITHUB_OUTPUT
            echo "pr_current=false" >> $GITHUB_OUTPUT
            echo "pr_number=${PR_NUMBER}" >> $GITHUB_OUTPUT
            echo "pr_branch=${PR_BRANCH}" >> $GITHUB_OUTPUT
          fi

      - name: Close outdated PR and cleanup branch
        if: steps.check_diff.outputs.has_diff == 'true' && steps.check_pr.outputs.pr_exists == 'true' && steps.check_pr.outputs.pr_current == 'false'
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          PR_NUMBER="${{ steps.check_pr.outputs.pr_number }}"
          PR_BRANCH="${{ steps.check_pr.outputs.pr_branch }}"

          echo "Closing outdated PR #${PR_NUMBER}"
          gh pr close "$PR_NUMBER" --comment "Closing this PR as a newer version with more recent commits is available."

          # Delete the automation branch (only if it matches lake-gate-* pattern for safety)
          if [[ "$PR_BRANCH" =~ ^lake-gate- ]]; then
            echo "Deleting branch ${PR_BRANCH}"
            git push origin --delete "$PR_BRANCH" || echo "Branch already deleted or doesn't exist remotely"
          else
            echo "Skipping branch deletion - branch name doesn't match lake-gate-* pattern"
          fi

      - name: Create new sync PR
        if: steps.check_diff.outputs.has_diff == 'true' && (steps.check_pr.outputs.pr_exists == 'false' || steps.check_pr.outputs.pr_current == 'false')
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          # Create timestamp-based branch name
          TIMESTAMP=$(date +"%Y%m%d-%H%M%S")
          BRANCH_NAME="lake-gate-${TIMESTAMP}"

          echo "Creating branch ${BRANCH_NAME}"

          # Create and push new branch from main
          git checkout -b "$BRANCH_NAME" origin/main
          git push origin "$BRANCH_NAME"

          # Create PR with detailed description
          LATEST_COMMIT_SHORT="${{ steps.check_diff.outputs.latest_commit_short }}"
          LATEST_COMMIT_MSG="${{ steps.check_diff.outputs.latest_commit_msg }}"

          PR_BODY=$(cat <<EOF
          ## Sync Stream (main) to Lake (stable)

          This PR synchronizes changes from the \`main\` branch (Stream) to the \`stable\` branch (Lake).

          **Latest commit:** ${LATEST_COMMIT_SHORT}
          **Commit message:** ${LATEST_COMMIT_MSG}

          ### Automated Merge Process

          This PR will be **automatically fast-forwarded** once all the following conditions are met:

          - [ ] All PR checks have passed
          - [ ] E2E tests are successful
          - [ ] No merge conflicts

          The auto-merge workflow will monitor check status and merge this PR automatically when all checks are green.

          **Note:** After all checks pass, the \`stable\` branch will be fast-forwarded to point to the exact same commit as \`main\`. No merge commit will be created - both branches will share the same commit SHA.

          ---
          *This is an automated PR created by the lake-gate workflow.*
          EOF
          )

          gh pr create \
            --base stable \
            --head "$BRANCH_NAME" \
            --title "[DO NOT MERGE MANUALLY] Sync main to stable (${LATEST_COMMIT_SHORT})" \
            --body "$PR_BODY" \
            --label "lake-gate"

      - name: Summary
        if: steps.check_diff.outputs.has_diff == 'true'
        run: |
          if [ "${{ steps.check_pr.outputs.pr_current }}" = "true" ]; then
            echo "✅ Existing PR #${{ steps.check_pr.outputs.pr_number }} is up to date"
          else
            echo "✅ New sync PR created successfully"
          fi