quatalog-scraper/.github/workflows/scraper.yml

name: Update data, generate static site, and deploy to Pages
on:
  workflow_dispatch:
  schedule:
    - cron: '15 * * * *'

jobs:
  scrape-data:
    name: Scrape data
    runs-on: ubuntu-latest
    steps:
      - name: Checkout scraping repo
        uses: actions/checkout@v4
        with:
          path: quatalog-scraping
        
      - name: Clone QuACS data
        uses: actions/checkout@v4
        with:
          repository: quacs/quacs-data
          path: quacs-data

      - name: Run scraper
        run: |
          # Usage: CourseOfferingsScraper <data_directory> <terms_offered_file> <prerequisites_file> <list_of_terms_file>
          mkdir new-data
          quatalog-scraping/bin/CourseOfferingsScraper \
            quacs-data/semester_data \
            new-data/terms_offered.json \
            new-data/prerequisites.json \
            new-data/terms_list.json
                      
      - name: Copy catalog.json from QuACS data
        run: |
          CURRENT_TERM="$(grep new-data/terms_list.json -e 'current_term' | cut -f4 -d '"')"
          rsync -avz "quacs-data/semester_data/$CURRENT_TERM/catalog.json" new-data/catalog.json
                    
      - name: Upload data to artifact
        uses: actions/upload-artifact@v4
        with:
          name: new-data
          path: new-data/
            
  push-new-data:
    name: Push new data to data repo
    runs-on: ubuntu-latest
    needs: [scrape-data]
    steps:
      - name: Clone Quatalog data
        uses: actions/checkout@v4
        with:
          repository: quatalog/data
          path: quatalog-data
          token: ${{ secrets.PUSH_TOKEN }}
    
      - name: Download data from artifact
        uses: actions/download-artifact@v4
        with:
          name: new-data
          path: new-data
      
      - name: Copy data to repo directory
        run: |
          rsync -avz new-data/ quatalog-data/
                    
      - name: Push new data
        working-directory: quatalog-data
        run: |
          git config user.name "Quatalog Updater"
          git config user.email "github_actions@quatalog.com"
          git add terms_offered.json prerequisites.json terms_list.json catalog.json
          git commit -m "$(date)" || exit 0
          git push          

  generate-site:
    name: Generate the static site
    runs-on: ubuntu-latest
    needs: [scrape-data]
    steps:
      - name: Checkout scraping repo
        uses: actions/checkout@v4
        with:
          path: quatalog-scraping
      
      - name: Clone Quatalog static site
        uses: actions/checkout@v4
        with:
          repository: quatalog/site
          ref: static-generated
          path: quatalog-site
          token: ${{ secrets.PUSH_TOKEN }}
      
      - name: Download data from artifact
        uses: actions/download-artifact@v4
        with:
          name: new-data
          path: new-data
          
      - name: Run the HTML generator
        run: |
          # Usage: GenerateHtml <terms_offered_file> <prerequisites_file> <list_of_terms_file> <catalog_file> <out_directory> <searchable_catalog_file> <courses_list_file>
          quatalog-scraping/bin/GenerateHtml \
            new-data/terms_offered.json \
            new-data/prerequisites.json \
            new-data/terms_list.json \
            new-data/catalog.json \
            courses/ \
            searchable_catalog.json \
            courses_list.json
                      
      - name: Merge data
        run: |
          rsync -avz courses/ quatalog-site/courses/
          rsync -avz searchable_catalog.json quatalog-site/json/searchable_catalog.json
          rsync -avz courses_list.json quatalog-site/json/courses_list.json          
        
      - name: Push generated HTML
        working-directory: quatalog-site
        run: |
          git config user.name "Quatalog Updater"
          git config user.email "github_actions@quatalog.com"
          git add courses json
          git commit -m "$(date)" || exit 0
          git push          

  push-csv:
    name: Push CSV file to data repo
    runs-on: ubuntu-latest
    needs: [generate-site]
    steps:
      - name: Checkout site repo/static-generated branch
        uses: actions/checkout@v4
        with:
          repository: quatalog/site
          ref: static-generated
          path: static-generated
      
      - name: Checkout data repo
        uses: actions/checkout@v4
        with:
          repository: quatalog/data
          path: quatalog-data
          token: ${{ secrets.PUSH_TOKEN }}
      
      - name: Create CSV file
        run: |
          echo '"Instructor","Course"' > quatalog-data/courses.csv
          cd static-generated/courses
          grep -oe '<li>.*</li>' *.html | 
            sort -u |
            grep -ve "-[0-9]9[4-7][0-9]" -e "-[0-9]9[89]0" -e "-[0-9]00[1-9]" -e "USA[RF]" -e "ADMN" -e "USNA" |
            sed -e 's/^/"/' -e 's/.html:<li>/","/' -e 's#</li>#"#' -e 's/"\([^"]*\)","\([^"]*\)"/"\2","\1"/' | 
            sort >> ../../quatalog-data/courses.csv
                      
      - name: Push CSV file
        working-directory: quatalog-data
        run: |
          git config user.name "Quatalog Updater"
          git config user.email "github_actions@quatalog.com"
          git add courses.csv
          git commit -m "$(date)" || exit 0
          git push          

  prepare-site:
    name: Prepare static site
    runs-on: ubuntu-latest
    needs: [generate-site]
    steps:
      - name: Checkout static-generated branch
        uses: actions/checkout@v4
        with:
          repository: quatalog/site
          ref: static-generated
          
      - name: Setup Pages
        uses: actions/configure-pages@v4
        
      - name: Archive github-pages artifact
        run: |
          tar \
            --dereference --hard-dereference \
            --exclude=.git \
            --exclude=.github \
            --exclude=LICENSE \
            --exclude=README.md \
            -cf "$RUNNER_TEMP/artifact.tar" .          

      - name: Upload github-pages artifact
        uses: actions/upload-artifact@v4
        with:
          name: github-pages
          path: ${{ runner.temp }}/artifact.tar

  deploy-site:
    name: Deploy to GitHub Pages
    runs-on: ubuntu-latest
    needs: [prepare-site]
    permissions:
      pages: write
      id-token: write
    concurrency:
      group: "pages"
      cancel-in-progress: true
    environment:
      name: github-pages
      url: ${{ steps.deployment.outputs.page_url }}
    steps:
      - name: Deploy to GitHub Pages
        id: deployment
        uses: actions/deploy-pages@v4