todos-extract-from-code.yml

name: Create issues from todos

on:
  push:
    # https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#onpull_requestpull_request_targetbranchesbranches-ignore
    branches:
      - master
    # https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#example-excluding-paths
    paths-ignore:
      - '**.json'
      - '**.jpg'
      - '**.png'
      - '**.ico'
      - '**.md'

# https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#permissions
permissions:
  contents: write # for "git push"
  issues: write # connect-todos-to-issues.sh uses "gh issue create"

defaults:
  # https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#defaultsrun
  run:
    # Enable fail-fast behavior using set -eo pipefail
    # https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#exit-codes-and-error-action-preference
    shell: bash

jobs:
  extract-pdd-puzzles:
    name: Extract todos from code
    # https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#jobsjob_idruns-on
    runs-on: ubuntu-20.04
    steps:
      - name: Clone source code
        uses: actions/checkout@v4.1.0 # https://github.com/actions/checkout
        with:
          # Whether to configure the token or SSH key with the local git config. Default: true
          persist-credentials: true

      - name: Install pdd
        run: sudo gem install pdd:0.23.1 --no-document

      - name: Checkout existing todos to another directory
        run: |
          git fetch --depth=1 origin generated-todos
          git worktree add generated-todos generated-todos

      - name: Extract todos
        run: |
          pdd \
            --skip-gitignore \
            --exclude '**/*.json' \
            --exclude '**/*.jpg' \
            --exclude '**/*.png' \
            --exclude '**/*.ico' \
            --exclude '**/*.md' \
            --exclude '**/*.enc' \
            --exclude 'src/test/wiremock/**/*' \
            --exclude 'generated-todos/**' \
            --verbose \
            --file pdd.xml

      - name: Remove the current date from pdd.xml
        run: sed -i -E '/<time>[-0-9T:Z]+<\/time>/d;/<puzzles/s| date="[-0-9T:Z]+"||' pdd.xml

      - name: Generate todos-in-code.tsv
        run: ./src/main/scripts/ci/pdd-xml-to-json.sh pdd.xml | ./src/main/scripts/ci/pdd-json-to-tsv.sh > todos-in-code.tsv

      # @todo #1610 Close an issue or post a comment when a puzzle got removed from code
      # @todo #1610 Post a comment when issue got closed without removing a puzzle
      - name: Check whether there are new todos
        run: |
          PUZZLES_FILES_MODIFIED=no

          if [ ! -f generated-todos/pdd.xml ]; then
            echo 'pdd.xml has just been created'
            mv pdd.xml generated-todos/pdd.xml
            PUZZLES_FILES_MODIFIED=yes
          elif ! diff --brief generated-todos/pdd.xml pdd.xml >/dev/null; then
            echo 'pdd.xml has been modified'
            mv pdd.xml generated-todos/pdd.xml
            PUZZLES_FILES_MODIFIED=yes
          fi

          if [ ! -f generated-todos/todos-in-code.tsv ]; then
            echo 'todos-in-code.tsv has just been created'
            mv todos-in-code.tsv generated-todos/todos-in-code.tsv
            PUZZLES_FILES_MODIFIED=yes
          elif ! diff --brief generated-todos/todos-in-code.tsv todos-in-code.tsv >/dev/null; then
            echo 'todos-in-code.tsv has been modified'
            mv todos-in-code.tsv generated-todos/todos-in-code.tsv
            PUZZLES_FILES_MODIFIED=yes
          fi

          if [ "$PUZZLES_FILES_MODIFIED" = 'yes' ]; then
            printf 'pdd.xml: %d puzzles\n' "$(grep -c '<puzzle>' generated-todos/pdd.xml)"
            printf 'todos-in-code.tsv: %d todos\n' "$(sed 1d generated-todos/todos-in-code.tsv | wc -l)"
          else
            echo 'neither pdd.xml nor todos-in-code.tsv have been modified'
          fi

          # Make variable available for the next steps
          echo "PUZZLES_FILES_MODIFIED=$PUZZLES_FILES_MODIFIED" | tee -a "$GITHUB_ENV"

      # It's possible that for some puzzles we haven't created issues yet
      # (because of network, rate limit or other kind of errors)
      - name: Check whether there are puzzles without related issues
        if: env.PUZZLES_FILES_MODIFIED == 'no'
        working-directory: generated-todos
        run: |
          NEW_PUZZLES="$(comm -23 <(cut -d$'\t' -f1 todos-in-code.tsv | sort) <(cut -d$'\t' -f1 todos-on-github.tsv | sort))"
          if [ -n "$NEW_PUZZLES" ]; then
            echo "Found puzzles without related issues:"
            echo "$NEW_PUZZLES" | while read PUZZLE_ID; do
              TITLE="$(grep "$PUZZLE_ID" todos-in-code.tsv | cut -d$'\t' -f3 | sed 's|^"||;s|"$||;s|""|"|g')"
              echo "- $PUZZLE_ID: $TITLE"
            done
            HAVE_PUZZLES_WITHOUT_ISSUES=yes
          else
            HAVE_PUZZLES_WITHOUT_ISSUES=no
          fi

          # Make variable available for the next steps
          echo "HAVE_PUZZLES_WITHOUT_ISSUES=$HAVE_PUZZLES_WITHOUT_ISSUES" | tee -a "$GITHUB_ENV"

      - name: Connect todos to the issues
        if: env.PUZZLES_FILES_MODIFIED == 'yes' || env.HAVE_PUZZLES_WITHOUT_ISSUES == 'yes'
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: ./src/main/scripts/ci/connect-todos-to-issues.sh generated-todos/todos-in-code.tsv

      - name: Check whether todos-on-github.tsv has been modified
        if: env.PUZZLES_FILES_MODIFIED == 'yes' || env.HAVE_PUZZLES_WITHOUT_ISSUES == 'yes'
        working-directory: generated-todos
        run: |
          PUZZLES_MAPPING_MODIFIED=no

          if ! git diff-index --quiet HEAD -- todos-on-github.tsv; then
            PUZZLES_MAPPING_MODIFIED=yes
            echo 'todos-on-github.tsv has been modified'
          fi

          # Make variable available for the next steps
          echo "PUZZLES_MAPPING_MODIFIED=$PUZZLES_MAPPING_MODIFIED" | tee -a "$GITHUB_ENV"

      - name: Initialize Git variables for committing the changes
        if: env.PUZZLES_FILES_MODIFIED == 'yes' || env.PUZZLES_MAPPING_MODIFIED == 'yes'
        run: |
          AUTHOR_EMAIL="$(git show "$GITHUB_SHA" --no-patch --format=%aE)"
          AUTHOR_NAME="$(git show "$GITHUB_SHA" --no-patch --format=%aN)"

          # Make variables available for the next steps
          (
            echo "COMMIT_MSG=$(git show "$GITHUB_SHA" --no-patch --format=oneline --abbrev-commit)"
            echo "GIT_AUTHOR_NAME=GitHub Action on behalf of $AUTHOR_NAME"
            echo "GIT_COMMITTER_NAME=GitHub Action on behalf of $AUTHOR_NAME"
            echo "GIT_AUTHOR_EMAIL=$AUTHOR_EMAIL"
            echo "GIT_COMMITTER_EMAIL=$AUTHOR_EMAIL"
          ) | tee -a "$GITHUB_ENV"

      - name: Commit the changes
        if: env.PUZZLES_FILES_MODIFIED == 'yes'
        env:
          NEW_COMMIT_MSG: "chore: processed ${{ env.COMMIT_MSG }}"
        working-directory: generated-todos
        run: |
          git commit pdd.xml todos-in-code.tsv -m "$NEW_COMMIT_MSG"
          git push

      - name: Commit updated mapping
        if: env.PUZZLES_MAPPING_MODIFIED == 'yes'
        env:
          NEW_COMMIT_MSG: "chore: sync issues for ${{ env.COMMIT_MSG }}"
        working-directory: generated-todos
        run: |
          git add todos-on-github.tsv
          git commit todos-on-github.tsv -m "$NEW_COMMIT_MSG"
          git push

      - name: Cleanup
        if: always()
        run: |
          [ ! -f pdd.xml ] || rm -fv pdd.xml
          [ ! -f todos-in-code.tsv ] || rm -fv todos-in-code.tsv
          [ ! -d generated-todos ] || git worktree remove generated-todos