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

name: Nubenetes Automated Agentic Curation

on:
  schedule:
    - cron: '0 5 1 * *'  # Monthly: 1st day of the month at 05:00 UTC
  workflow_dispatch:
    inputs:
      start_date:
        description: 'Start date (YYYY-MM-DD). Default: 2026-05-15'
        required: true
        default: '2026-05-15'
      days_back:
        description: 'Last N days (e.g. 7). Overrides start date if filled'
        required: false
        default: ''
      exclude_accounts:
        description: 'Accounts to exclude (comma separated, e.g. nubenetes,OpenAI)'
        required: false
        default: ''
      include_k8s:
        description: 'Include Kubernetes & Cloud Native'
        type: boolean
        default: true
      include_cloud:
        description: 'Include Cloud Providers (AWS/Azure/GCP)'
        type: boolean
        default: true
      include_ai:
        description: 'Include AI & Agentic Systems'
        type: boolean
        default: true
      include_dev:
        description: 'Include Developer Productivity & AI Agents'
        type: boolean
        default: true
      include_data:
        description: 'Include Data & Big Data'
        type: boolean
        default: true
      include_iac:
        description: 'Include Infrastructure as Code & GitOps'
        type: boolean
        default: true
      extraction_strategy:
        description: 'Extraction Strategy'
        required: true
        default: 'search'
        type: choice
        options:
          - search
          - scroll
      historical_mode:
        description: 'Activate Historical Mode'
        required: false
        default: 'true'
        type: boolean
      historical_chunked:
        description: 'Chunk execution (multiple PRs)'
        required: false
        default: 'false'
        type: boolean
      historical_until_date:
        description: 'Upper limit date (chunk)'
        required: false
        default: ''

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

jobs:
  agentic-curation-process:
    runs-on: ubuntu-latest
    if: github.ref == 'refs/heads/develop' || github.event_name == 'schedule'
    steps:
      - name: Repository Synchronization
        uses: actions/checkout@v4
        with:
          ref: develop

      - name: Python 3.11 Environment Provisioning
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'

      - name: Installation of Dependencies
        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 pyyaml
          playwright install chromium --with-deps

      - name: Global Agentic Curation Execution
        id: run_curation
        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 }}
          GEMINI_API_KEY_2: ${{ secrets.GEMINI_API_KEY_2 }}
          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 || '2026-05-15' }}
          CURATION_DAYS_BACK: ${{ github.event.inputs.days_back || '' }}
          EXCLUDE_ACCOUNTS: ${{ github.event.inputs.exclude_accounts || '' }}
          INCLUDE_K8S: ${{ github.event.inputs.include_k8s || 'true' }}
          INCLUDE_CLOUD: ${{ github.event.inputs.include_cloud || 'true' }}
          INCLUDE_AI: ${{ github.event.inputs.include_ai || 'true' }}
          INCLUDE_DEV: ${{ github.event.inputs.include_dev || 'true' }}
          INCLUDE_DATA: ${{ github.event.inputs.include_data || 'true' }}
          INCLUDE_IAC: ${{ github.event.inputs.include_iac || 'true' }}
          HISTORICAL_CHUNK_DAYS: '180'
          PYTHONPATH: .
        run: |
          python -u src/main.py 2>&1 | tee output.log
          
          # Capture PR URL if exists
          PR_URL=$(grep "PULL_REQUEST_URL:" output.log | awk '{print $2}')
          echo "pr_url=$PR_URL" >> $GITHUB_OUTPUT
          
          # Re-trigger logic for Historical Mode
          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 "Triggering next historical chunk until: $NEXT_DATE"
            gh workflow run agentic_cron.yml -f historical_mode=true -f historical_chunked=true -f historical_until_date=$NEXT_DATE
          fi

      - name: Send Email Notification
        if: always()
        uses: dawidd6/action-send-mail@v3
        with:
          server_address: smtp.gmail.com
          server_port: 465
          username: ${{ secrets.MAIL_USERNAME }}
          password: ${{ secrets.MAIL_PASSWORD }}
          subject: "Nubenetes Curation Report - ${{ job.status }}"
          to: "nubenetes@gmail.com"
          from: "Nubenetes Bot <bot@nubenetes.com>"
          body: |
            Nubenetes Automated Curation has finished.
            
            Status: ${{ job.status }}
            Pull Request: ${{ steps.run_curation.outputs.pr_url || 'No PR created (no changes or error)' }}
            
            Check the logs for more details: https://github.com/${{ github.repository }}/actions/runs/${{ github.run_id }}