refresh-repo-catalog.yml

name: Refresh Tiered Repo Catalog

on:
  workflow_dispatch:
    inputs:
      force_all_time:
        description: 'Force refresh all-time tier'
        required: false
        type: boolean
        default: false
  schedule:
    # Weekly refresh (weekly, monthly, 6-month, yearly tiers)
    - cron: "0 3 * * 1" # Every Monday at 3 AM UTC
    # All-time refresh (every 5 years - manual trigger recommended)
    # Note: For practical purposes, all-time tier can be triggered manually
    # or updated once per year if needed

permissions:
  contents: write

jobs:
  refresh:
    name: Refresh Catalog with Trending Algorithm
    runs-on: ubuntu-latest

    steps:
      - name: Checkout
        uses: actions/checkout@v4

      - name: Setup Node.js
        uses: actions/setup-node@v4
        with:
          node-version: 20
          cache: npm

      - name: Install dependencies
        run: npm ci --ignore-scripts

      - name: Check if all-time tier needs refresh
        id: check_alltime
        run: |
          # Check last all-time refresh date from git log
          LAST_ALLTIME=$(git log --all --grep="all-time tier" --date=format:'%Y-%m-%d' --pretty=format:'%ad' -1 2>/dev/null || echo "1970-01-01")
          DAYS_SINCE=$(( ($(date +%s) - $(date -d "$LAST_ALLTIME" +%s)) / 86400 ))

          # Refresh all-time tier every 1825 days (5 years) or if forced
          if [ "${{ github.event.inputs.force_all_time }}" == "true" ] || [ $DAYS_SINCE -gt 1825 ]; then
            echo "refresh_alltime=true" >> $GITHUB_OUTPUT
            echo "All-time tier will be refreshed (last refresh: $DAYS_SINCE days ago)"
          else
            echo "refresh_alltime=false" >> $GITHUB_OUTPUT
            echo "All-time tier skipped (last refresh: $DAYS_SINCE days ago, next in $((1825 - DAYS_SINCE)) days)"
          fi

      - name: Refresh trending tiers (weekly, monthly, 6-month, yearly)
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          SKIP_ALL_TIME: ${{ steps.check_alltime.outputs.refresh_alltime == 'false' }}
        run: |
          if [ "$SKIP_ALL_TIME" == "true" ]; then
            echo "⏭️  Skipping all-time tier (not due for refresh)"
            # Run script with all-time tier disabled
            node scripts/fetch-trending-repos.mjs --skip-all-time
          else
            echo "🔄 Refreshing all tiers including all-time"
            npm run data:refresh:repos
          fi

      - name: Validate data integrity
        run: |
          REPO_COUNT=$(grep -c '"owner"' public/data/top-repos.json)
          echo "Found $REPO_COUNT repositories."
          if [ "$REPO_COUNT" -lt 3000 ]; then
            echo "Error: Catalog size ($REPO_COUNT) is below the minimum threshold (3000)."
            exit 1
          fi

      - name: Commit refreshed dataset
        run: |
          if git diff --quiet -- public/data/top-repos.json; then
            echo "No dataset changes detected."
            exit 0
          fi

          git config user.name "github-actions[bot]"
          git config user.email "41898282+github-actions[bot]@users.noreply.github.com"
          git add public/data/top-repos.json

          # Create commit message based on what was refreshed
          if [ "${{ steps.check_alltime.outputs.refresh_alltime }}" == "true" ]; then
            git commit -m "chore(data): refresh all tiers including all-time tier (trending algorithm)"
          else
            git commit -m "chore(data): refresh trending tiers (weekly, monthly, 6-month, yearly)"
          fi

          git push
        # NOTE: Sitemap is refreshed automatically on the next Vercel deployment after this commit.