action.yml

# GitHub Actions definition for Waystation.
#
# Copyright 2022-2024 California Institute of Technology.
# License: Modified BSD 3-clause – see file "LICENSE" in the project website.
# Website: https://github.com/caltechlibrary/waystation

name: Waystation
description: Archive a repository's GitHub Pages in the Wayback Machine
author: Michael Hucka – https://github.com/mhucka
branding:
  icon: camera
  color: orange

inputs:
  dry_run:
    description: Print what would be done without actually doing it.
    default: false
    required: false
  save_errors:
    # Corresponds to "saveErrors" in the Wayback Machine GitHub Action.
    description: Whether the Wayback Machine should save pages that return HTTP status 4xx-5xx.
    default: false
    required: false
  save_outlinks:
    # Corresponds to "saveOutlinks" in the Wayback Machine GitHub Action.
    description: Tell the Wayback Machine to save links to external web pages.
    default: true
    required: false
  save_screenshot:
    # Corresponds to "saveScreenshot" in the Wayback Machine GitHub Action.
    description: Tell the Wayback Machine to save a screenshot of the front page.
    default: true
    required: false
  debug:
    description: Print context and variables in the GitHub Actions log.
    default: false
    required: false

runs:
  using: composite
  steps:
    - name: Get GitHub Pages URL.
      # This stores the URL in env.pages_url. The value will be "null" if
      # no pages URL is returned for any reason.
      shell: bash
      run: |
        echo pages_url=$(curl -s -H 'Accept: application/vnd.github+json' -H "Authorization: Bearer ${{github.token}}" https://api.github.com/repos/${{github.repository}}/pages | jq '.html_url' | tr -d '"') >> $GITHUB_ENV

    - name: Dump GitHub Actions context and variables if debugging is on.
      if: inputs.debug == true || inputs.debug == 'true'
      env:
        # Setting this variable and then printing it below is not strictly
        # necessary; you could put the expression directly into the echo
        # command below. But when I do that, I get a doubling of the output.
        GITHUB_CONTEXT: ${{toJSON(github)}}
      shell: bash
      run: |
        echo "dry_run = ${{inputs.dry_run}}"
        echo "pages_url = ${{env.pages_url}}"
        echo "save_outlinks = ${{inputs.save_outlinks}}"
        echo "save_screenshots = ${{inputs.save_screenshots}}"
        echo "debug = ${{inputs.debug}}"
        echo "====================="
        echo "$GITHUB_CONTEXT"

    - name: Quit if GitHub Pages URL is null.
      if: startsWith(env.pages_url, 'null') == true
      uses: actions/github-script@v7
      with:
        script: |
          core.setFailed("Could not retrieve the GitHub Pages URL for repository ${{github.repository}}.")

    - name: Proceed if this is not a dry run.
      if: startsWith(env.pages_url, 'null') != true && inputs.dry_run != true && inputs.dry_run != 'true'
      uses: JamieMagee/wayback@v1.3.51
      with:
        url: ${{env.pages_url}}
        saveOutlinks: ${{inputs.save_outlinks}}
        saveScreenshot: ${{inputs.save_screenshot}}

    - name: Summarize the results of the run.
      shell: bash
      env:
        wm_url: https://web.archive.org/web/${{env.pages_url}}
      run: |
        echo \#\# Results of sending the GitHub Pages to the Wayback Machine >> $GITHUB_STEP_SUMMARY
        echo Dry run: ${{inputs.dry_run}}                                    >> $GITHUB_STEP_SUMMARY
        echo GitHub repository for this action: \`${{github.repository}}\`   >> $GITHUB_STEP_SUMMARY
        echo GitHub Pages URL for the repository: ${{env.pages_url}}         >> $GITHUB_STEP_SUMMARY
        echo **URL in the Wayback Machine**: ${{env.wm_url}}                 >> $GITHUB_STEP_SUMMARY
        echo Date and time of this run: $(date "+%F %T (%Z)")                >> $GITHUB_STEP_SUMMARY
        echo ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━                               >> $GITHUB_STEP_SUMMARY
        echo Note: sometimes the Wayback Machine takes extra long to start   >> $GITHUB_STEP_SUMMARY
        echo archiving a site. If your GitHub Pages URL is not yet found at  >> $GITHUB_STEP_SUMMARY
        echo Wayback Machine URL above, wait for a time and then try to      >> $GITHUB_STEP_SUMMARY
        echo look up your GitHub Pages URL again.                            >> $GITHUB_STEP_SUMMARY
        echo ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━                               >> $GITHUB_STEP_SUMMARY