collect-purls_template.yml

name: Collect base PURLs for an ecosystem from AboutCode federated package metadata
on:
  workflow_call:
    inputs:
      ecosystem:
        description: "Ecosystem for which to collect base PackageURLs (eg. npm)"
        required: true
        type: string
      path:
        description: "Path to store collected PURLs (eg. 'data/npm.txt')"
        required: true
        type: string

permissions:
  contents: write

jobs:
  clone-repos:
    runs-on: ubuntu-latest

    steps:
      - name: Install GitHub CLI
        run: |
          sudo apt update && sudo apt install -y gh

      - name: Fetch repo list
        env:
          GH_TOKEN: ${{ github.token }}
        run: |
          mkdir ../repos && cd ../repos
          gh repo list aboutcode-data --limit 20000 --json name \
            --jq '.[] | select(.name | startswith("purls-${{ inputs.ecosystem }}")) | .name' \
            > repos.txt

      - name: Shallow clone repos
        run: |
          cd ../repos
          cat repos.txt | xargs -n 1 -P 8 -I {} bash -c '
            echo "Cloning {}"
            git clone --depth 1 "https://github.com/aboutcode-data/{}.git"
          '
      - name: Extract base purl from purls.yml
        run: |
          cd ../repos
          find . -type f -name "purls.yml" -exec sed -n '1{ s/^- *//; s/@.*//; p }' {} \; | sort \
          >  "${{ inputs.ecosystem }}.txt"

      - name: Checkout repository
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd  # v6.0.2

      - name: Commit and push if it changed
        run: |-
          dest="${{ inputs.path }}"
          git config user.name "AboutCode Automation"
          git config user.email "automation@aboutcode.org"
          git pull
          mkdir -p "$(dirname "$dest")" && mv ../repos/${{ inputs.ecosystem }}.txt "$dest"
          git add -A
          git commit -m "$(echo -e "Sync latest ${{ inputs.ecosystem }} PURLs from FederatedCode\n\nSigned-off-by: AboutCode Automation <automation@aboutcode.org>")" || exit 0
          git push