refresh_cache.yml

name: Refresh Model Cache

on:
  schedule:
    - cron: '0 0 * * *'  # Every night at 00:00 UTC
  workflow_dispatch:

permissions:
  contents: write

jobs:
  refresh:
    name: Nightly Refresh
    runs-on: ubuntu-latest
    timeout-minutes: 30
    steps:
      - name: Checkout
        uses: actions/checkout@v4

      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'

      - name: Install dependencies
        run: pip install requests urllib3

      - name: Run cache refresh
        run: python scripts/refresh_cache.py

      - name: Validate cache.json
        run: |
          python -c "
          import json, sys
          with open('data/cache.json') as f:
              d = json.load(f)
          sources = d.get('sources', {})
          models = d.get('models', [])
          gpus = d.get('gpus', [])
          leaderboard_ok = sources.get('huggingface_leaderboard', {}).get('status') == 'success'
          models_ok = len(models) > 0
          print(f'Models: {len(models)}, GPUs: {len(gpus)}')
          print(f'Leaderboard: {sources.get(\"huggingface_leaderboard\", {}).get(\"status\")} ({sources.get(\"huggingface_leaderboard\", {}).get(\"row_count\", 0)} rows)')
          for k, v in sources.items():
              print(f'  {k}: {v.get(\"status\")} ({v.get(\"row_count\", 0)} rows)')
          if not models_ok:
              print('ERROR: No models in cache!')
              sys.exit(1)
          if not leaderboard_ok:
              print('WARNING: Leaderboard fetch failed - benchmarks will use fallback estimates')
          else:
              measured = sum(1 for m in models if m.get('benchmarks', {}).get('status') in ('measured', 'derived'))
              print(f'Benchmark coverage: {measured}/{len(models)} models have measured/derived scores')
          "

      - name: Commit and push updated cache
        run: |
          git config --global user.name "github-actions[bot]"
          git config --global user.email "github-actions[bot]@users.noreply.github.com"
          git add data/cache.json
          git diff --staged --quiet || git commit -m "Nightly refresh: update model & GPU cache database"
          git pull --rebase origin main && git push

  # NOTE: No explicit deploy job here. This repo's Pages is configured as
  # "Deploy from a branch" (legacy mode), so GitHub's built-in
  # "pages build and deployment" workflow builds and deploys automatically on
  # every push to main — including the cache.json push above. Adding our own
  # actions/deploy-pages step here (as was done before) created a second,
  # concurrent deployment that raced with the built-in one and intermittently
  # failed with "Deployment request failed ... due to in progress deployment"
  # (see run 31551404326, 2026-08-12).