dvc-diff.yml

# Report changes in test images
#
# This workflow checks for image diffs in a pull request and adds a GitHub
# comment showing the diff.
#
# It is triggered in a PR when any *.dvc files have been added, modified,
# or deleted. A GitHub comment will be published that contains a summary table
# of the images that have changed along with a visual report.
#
name: DVC image diff

on:
  pull_request:
    paths:
      - '**/*.dvc'

permissions:
  contents: read
  pull-requests: write

jobs:
  dvc-diff:
    name: DVC image diff
    runs-on: ubuntu-latest
    defaults:
      run:
        shell: bash -l {0}

    steps:
    - name: Checkout
      uses: actions/checkout@v6.0.2

    - name: Setup data version control (DVC)
      uses: iterative/setup-dvc@v2

    - name: Pull baseline image data from dvc remote
      id: dvc-pull
      continue-on-error: true
      env:
        DAGSHUB_TOKEN: ${{ secrets.DAGSHUB_TOKEN }}
      run: |
        dvc remote modify origin url https://${DAGSHUB_TOKEN}@dagshub.com/GenericMappingTools/gmt.dvc --local
        dvc pull --remote origin --no-run-cache
        dvc pull --remote origin --no-run-cache

    - name: Download DVC cache from GitHub Artifacts and restore baseline image data
      if: steps.dvc-pull.outcome == 'failure'
      run: |
        gh run download -n dvc-cache -D .dvc/cache/
        dvc checkout --force
      env:
        GH_TOKEN: ${{ github.token }}

    - name: Setup continuous machine learning (CML)
      uses: iterative/setup-cml@v3

    - name: Setup Micromamba
      uses: mamba-org/setup-micromamba@v3.0.0
      with:
        environment-name: gmt
        condarc: |
          channels:
            - conda-forge
            - nodefaults
        cache-downloads: true
        cache-environment: true
        create-args: >-
          gmt

    # Produce the markdown diff report, which should look like:
    # ## Summary of changed images
    #
    # This is an auto-generated report of images that have changed on the DVC remote
    #
    # | Status   | Path                                |
    # |----------|-------------------------------------|
    # | added    | test/baseline/test_image.png |
    # | deleted  | test/baseline/test_image2.png |
    # | modified | test/baseline/test_image3.png |
    #
    # ## Image diff(s)
    #
    # <details>
    # ...
    # </details>
    #
    # Report last updated at commit abcdef
    - name: Generate the image diff report
      env:
        REPO_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        DAGSHUB_TOKEN: ${{ secrets.DAGSHUB_TOKEN }}
      run: |
        echo -e "## Summary of changed images\n" > report.md
        echo -e "This is an auto-generated report of images that have changed on the DVC remote\n" >> report.md

        # Fetch the master branch
        git fetch origin master

        dvc diff --md master HEAD >> report.md

        # Get just the filename of the added and modified image from the report
        awk 'NF==5 && NR>=7 && $2=="added" && $4 ~ /.ps$/ {print $4}' report.md > added_files.txt
        awk 'NF==5 && NR>=7 && $2=="modified" && $4 ~ /.ps$/ {print $4}' report.md > modified_files.txt

        # Backup new images in a different directory
        mkdir -p newbaseline/test && cp -r test/baseline newbaseline/test/
        mkdir -p newbaseline/doc/examples && cp -r doc/examples/images newbaseline/doc/examples/
        mkdir -p newbaseline/doc/scripts && cp -r doc/scripts/images newbaseline/doc/scripts/
        rm -r test/baseline/**/*.ps doc/examples/images/*.ps doc/scripts/images/*.ps
        # Restore images for the master branch, using the DVC remote when available.
        git checkout master
        if ! dvc pull --remote origin --no-run-cache; then
          dvc checkout --force
        fi
        if ! dvc pull --remote origin --no-run-cache; then
          dvc checkout --force
        fi

        # Append each image to the markdown report
        echo -e "## Image diff(s)\n" >> report.md
        echo -e "<details>\n" >> report.md

        # Added images
        echo -e "### Added images\n" >> report.md
        while IFS= read -r line; do
          gmt psconvert -A -P -Tg $line "newbaseline/$line"
          echo -e "- ${line%.*}.png \n" >> report.md
          echo -e "![](newbaseline/${line%.*}.png)" >> report.md
        done < added_files.txt

        # Modified images
        echo -e "### Modified images\n" >> report.md
        echo -e "| Path | Old | New |" >> report.md
        echo -e "|---|---|---|" >> report.md
        while IFS= read -r line; do
          gmt psconvert -A -P -Tg $line "newbaseline/$line"
          echo -e "| $line | ![](${line%.*}.png) | ![](newbaseline/${line%.*}.png) |" >> report.md
        done < modified_files.txt

        echo -e "</details>\n" >> report.md

        # Mention git commit SHA in the report
        echo -e "Report last updated at commit ${{ github.event.pull_request.head.sha }}" >> report.md

        # create/update PR comment
        cml comment update report.md