links.yaml

---
name: Check URLs 🌐

on:
  push:
    branches:
      - main
  pull_request:
    types:
      - opened
      - synchronize
      - reopened
      - ready_for_review
    branches:
      - main
  workflow_dispatch:
  workflow_call:
    inputs:
      lychee-additional-args:
        description: Additional arguments to pass to lychee
        required: false
        type: string
        default: "--exclude-private"
      lychee-fail:
        description: Fail workflow run on error (i.e. when lychee exit code is not 0)
        required: false
        type: boolean
        default: false
      package-subdirectory:
        description: Subdirectory in the repository, where the R package is located.
        required: false
        type: string
        default: ""
      link-checking-method:
        description: |
          Which link checking method should be used? Supported methods are:
          lychee
          urlchecker
        required: false
        type: string
        default: "urlchecker"

concurrency:
  group: links-${{ github.event.pull_request.number || github.ref }}
  cancel-in-progress: true

jobs:
  links:
    name: Validate Links 🔎
    runs-on: ubuntu-latest
    if: >
      !contains(github.event.commits[0].message, '[skip links]')
        && github.event.pull_request.draft == false
        && inputs.link-checking-method == 'lychee'
    steps:
      - name: Get branch names 🌿
        id: branch-name
        uses: tj-actions/branch-names@v7

      - name: Checkout repo (PR) 🛎
        uses: actions/checkout@v4.1.1
        if: github.event_name == 'pull_request'
        with:
          ref: ${{ steps.branch-name.outputs.head_ref_branch }}
          repository: ${{ github.event.pull_request.head.repo.full_name }}

      - name: Checkout repo 🛎
        uses: actions/checkout@v4.1.1
        if: github.event_name != 'pull_request'
        with:
          ref: ${{ steps.branch-name.outputs.head_ref_branch }}

      - name: Check commit message 💬
        run: |
          git config --global --add safe.directory $(pwd)
          export head_commit_message="$(git show -s --format=%B | tr '\r\n' ' ' | tr '\n' ' ')"
          echo "head_commit_message = $head_commit_message"
          if [[ $head_commit_message == *"$SKIP_INSTRUCTION"* ]]; then
            echo "Skip instruction detected - cancelling the workflow."
            exit 1
          fi
        shell: bash
        env:
          SKIP_INSTRUCTION: "[skip links]"

      - name: Check URLs in docs 🔬
        uses: lycheeverse/lychee-action@v1.10.0
        with:
          args: >-
            --exclude "https://github.com.*.git|https://insightsengineering.github.io.*|lewagon.*|knightdave.*|.*users.noreply.github.com|lycheeverse.*"
            --verbose
            --no-progress
            ${{ inputs.lychee-additional-args }}
            **/*.md
            **/*.html
            **/*.Rmd
            **/*.yaml
            **/*.yml
            *.md
          fail: ${{ inputs.lychee-fail }}
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

  urlchecker:
    name: URL Checker 🔎
    runs-on: ubuntu-latest
    container:
      image: rocker/tidyverse:latest
    if: >
      !contains(github.event.commits[0].message, '[skip links]')
        && github.event.pull_request.draft == false
        && inputs.link-checking-method == 'urlchecker'
    steps:
      - name: Get branch names 🌿
        id: branch-name
        uses: tj-actions/branch-names@v7

      - name: Checkout repo (PR) 🛎
        uses: actions/checkout@v4.1.1
        if: github.event_name == 'pull_request'
        with:
          ref: ${{ steps.branch-name.outputs.head_ref_branch }}
          repository: ${{ github.event.pull_request.head.repo.full_name }}

      - name: Checkout repo 🛎
        uses: actions/checkout@v4.1.1
        if: github.event_name != 'pull_request'
        with:
          ref: ${{ steps.branch-name.outputs.head_ref_branch }}

      - name: Check commit message 💬
        run: |
          git config --global --add safe.directory $(pwd)
          export head_commit_message="$(git show -s --format=%B | tr '\r\n' ' ' | tr '\n' ' ')"
          echo "head_commit_message = $head_commit_message"
          if [[ $head_commit_message == *"$SKIP_INSTRUCTION"* ]]; then
            echo "Skip instruction detected - cancelling the workflow."
            exit 1
          fi
        shell: bash
        env:
          SKIP_INSTRUCTION: "[skip links]"

      - name: Check URLs with urlchecker 🔬
        run: |
          # For unexplained reasons, parallel = FALSE is required to prevent some false positives.
          bad_urls <- nrow(print(urlchecker::url_check(".", parallel = FALSE)))
          if (bad_urls > 0) {
            stop("Looks like a total of ", bad_urls, " URL(s) were found! Please correct them.")
          }
        shell: Rscript {0}
        working-directory: ${{ inputs.package-subdirectory }}