awesome-kubernetes/.github/workflows/agentic_cron.yml

name: Nubenetes Automated Agentic Curation

on:
  schedule:
    - cron: '0 5 * * 0'
  workflow_dispatch:
    inputs:
      start_date:
        description: 'Fecha inicial para la búsqueda (YYYY-MM-DD)'
        required: true
        default: '2024-10-01'
      extraction_strategy:
        description: 'Estrategia de Extracción'
        required: true
        default: 'search'
        type: choice
        options:
          - search
          - scroll
      historical_mode:
        description: 'Activar Modo Histórico'
        required: false
        default: 'true'
        type: boolean
      historical_chunked:
        description: 'Trocear ejecución (múltiples PRs)'
        required: false
        default: 'false'
        type: boolean
      historical_until_date:
        description: 'Fecha límite superior (tramo)'
        required: false
        default: ''

permissions:
  contents: write
  pull-requests: write
  actions: write

jobs:
  agentic-curation-process:
    runs-on: ubuntu-latest
    # Solo ejecutar en develop
    if: github.ref == 'refs/heads/develop'
    steps:
      - name: Sincronización del repositorio
        uses: actions/checkout@v4
        with:
          ref: develop

      - name: Provisión del Entorno Python 3.11
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'

      - name: Instalación de dependencias (LIGERO Y ROBUSTO)
        run: |
          python -m pip install --upgrade pip
          pip install --no-cache-dir pydantic PyGithub aiohttp beautifulsoup4 httpx fake-useragent pytz python-dotenv twikit>=2.1.2 playwright playwright-stealth
          playwright install chromium --with-deps

      - name: Ejecución de la Canalización Agéntica Integral
        env:
          TWITTER_USERNAME: ${{ secrets.TWITTER_USERNAME }}
          TWITTER_EMAIL: ${{ secrets.TWITTER_EMAIL }}
          TWITTER_PASSWORD: ${{ secrets.TWITTER_PASSWORD }}
          TWITTER_COOKIES: ${{ secrets.TWITTER_COOKIES }}
          GEMINI_API_KEY_1: ${{ secrets.GEMINI_API_KEY_1 }}
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          EXTRACTION_STRATEGY: ${{ github.event.inputs.extraction_strategy || 'search' }}
          HISTORICAL_MODE: ${{ github.event.inputs.historical_mode || 'false' }}
          HISTORICAL_CHUNKED: ${{ github.event.inputs.historical_chunked || 'false' }}
          HISTORICAL_UNTIL_DATE: ${{ github.event.inputs.historical_until_date || '' }}
          CURATION_START_DATE: ${{ github.event.inputs.start_date || '' }}
          HISTORICAL_CHUNK_DAYS: '180'
          PYTHONPATH: .
        run: |
          python -u src/main.py 2>&1 | tee output.log
          
          # Lógica de Re-disparo para Modo Histórico (SOLO SI SE SOLICITA TROCEADO)
          if [ "${{ github.event.inputs.historical_chunked }}" == "true" ] && grep -q "NEXT_CHUNK_START:" output.log; then
            NEXT_DATE=$(grep "NEXT_CHUNK_START:" output.log | awk '{print $2}')
            echo "Disparando siguiente tramo histórico hasta: $NEXT_DATE"
            gh workflow run agentic_cron.yml -f historical_mode=true -f historical_chunked=true -f historical_until_date=$NEXT_DATE
          fi