.github/workflows/dataset_update.yml

name: WhisperKit Benchmarks Dataset Update Workflow

on:
  schedule:
    - cron: "0 * * * *"
  workflow_dispatch:

jobs:
  update-datasets:
    runs-on: ubuntu-latest
    permissions:
      contents: write
    steps:
      - uses: actions/checkout@v4

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

      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install huggingface_hub requests

      - name: Check dataset updates
        id: check_updates
        env:
          HF_TOKEN: ${{ secrets.HF_TOKEN }}
        run: python .github/scripts/check_dataset_update.py

      - name: Install full requirements
        if: steps.check_updates.outputs.has_updates == 'true'
        run: |
          pip install -r requirements.txt

      - name: Run generate script
        if: steps.check_updates.outputs.has_updates == 'true'
        run: |
          make update-performance-data

      - name: Commit and push if changed
        if: steps.check_updates.outputs.has_updates == 'true'
        env:
          GH_TOKEN: ${{ secrets.BOT_TOKEN }}
          GH_USERNAME: ${{ secrets.BOT_USERNAME }}
          HF_TOKEN: ${{ secrets.HF_TOKEN }}
        run: |
          git config --global user.email "bots@takeargmax.com"
          git config --global user.name $GH_USERNAME

          git remote set-url origin https://x-access-token:${GH_TOKEN}@github.com/${GITHUB_REPOSITORY}.git

          git add .
          git commit -m "update dataset files" || echo "No changes to commit"
          git push
          git push https://HF_USERNAME:$HF_TOKEN@huggingface.co/spaces/argmaxinc/whisperkit-benchmarks