check-external-doc-links.yml

---
name: Check for External Doc Links

on:
  pull_request:
    branches: [main]
    paths:
      - "**/*.md"
      - "**/*.mdx"

jobs:
  check-external-doc-links:
    name: Block links to old doc sites
    runs-on: ubuntu-latest
    timeout-minutes: 5
    permissions:
      contents: read

    steps:
      - name: Checkout code
        uses: actions/checkout@v6
        with:
          fetch-depth: 0
          persist-credentials: false

      - name: Check for links to js.langchain.com and python.langchain.com
        run: |
          # Get the diff of added lines only (lines starting with +, excluding the +++ header)
          ADDED_LINES=$(git diff origin/${{ github.base_ref }}...HEAD --diff-filter=ACMR -- '*.md' '*.mdx' | grep -E '^\+' | grep -v '^\+\+\+' || true)

          # Check if any added lines contain links to the old doc sites
          if echo "$ADDED_LINES" | grep -qE '(js\.langchain\.com|python\.langchain\.com)'; then
            echo "❌ This PR adds links to deprecated documentation sites."
            echo ""
            echo "Links to js.langchain.com and python.langchain.com are not allowed."
            echo "Please use relative links to reference documentation within this repository."
            echo ""
            echo "Matches found:"
            echo "$ADDED_LINES" | grep -E '(js\.langchain\.com|python\.langchain\.com)' | head -20
            exit 1
          else
            echo "✅ No links to deprecated doc sites found in added content."
          fi