osu-lazer/.github/workflows/diffcalc.yml

# ## Description
#
# Uses [diffcalc-sheet-generator](https://github.com/smoogipoo/diffcalc-sheet-generator) to run two builds of osu and generate an SR/PP/Score comparison spreadsheet.
#
# ## Requirements
#
# Self-hosted runner with installed:
# - `docker >= 20.10.16`
# - `docker-compose >= 2.5.1`
# - `lbzip2`
# - `jq`
#
# ## Usage
#
# The workflow can be run in two ways:
# 1. Via workflow dispatch.
# 2. By an owner of the repository posting a pull request or issue comment containing `!diffcalc`.  
# For pull requests, the workflow will assume the pull request as the target to compare against (i.e. the `OSU_B` variable).  
# Any lines in the comment of the form `KEY=VALUE` are treated as variables for the generator.
#
# ## Google Service Account
#
# Spreadsheets are uploaded to a Google Service Account, and exposed with read-only permissions to the wider audience.
#
# 1. Create a project at https://console.cloud.google.com
# 2. Enable the `Google Sheets` and `Google Drive` APIs.
# 3. Create a Service Account
# 4. Generate a key in the JSON format.
# 5. Encode the key as base64 and store as an **actions secret** with name **`DIFFCALC_GOOGLE_CREDENTIALS`**
#
# ## Environment variables
#
# The default environment may be configured via **actions variables**.
#
# Refer to [the sample environment](https://github.com/smoogipoo/diffcalc-sheet-generator/blob/master/.env.sample), and prefix each variable with `DIFFCALC_` (e.g. `DIFFCALC_THREADS`, `DIFFCALC_INNODB_BUFFER_SIZE`, etc...).

name: Run difficulty calculation comparison

run-name: "${{ github.event_name == 'workflow_dispatch' && format('Manual run: {0}', inputs.osu-b) || 'Automatic comment trigger' }}"

on:
  issue_comment:
    types: [ created ]
  workflow_dispatch:
    inputs:
      osu-b:
        description: "The target build of ppy/osu"
        type: string
        required: true
      ruleset:
        description: "The ruleset to process"
        type: choice
        required: true
        options:
          - osu
          - taiko
          - catch
          - mania
      converts:
        description: "Include converted beatmaps"
        type: boolean
        required: false
        default: true
      ranked-only:
        description: "Only ranked beatmaps"
        type: boolean
        required: false
        default: true
      generators:
        description: "Comma-separated list of generators (available: [sr, pp, score])"
        type: string
        required: false
        default: 'pp,sr'
      osu-a:
        description: "The source build of ppy/osu"
        type: string
        required: false
        default: 'latest'
      difficulty-calculator-a:
        description: "The source build of ppy/osu-difficulty-calculator"
        type: string
        required: false
        default: 'latest'
      difficulty-calculator-b:
        description: "The target build of ppy/osu-difficulty-calculator"
        type: string
        required: false
        default: 'latest'
      score-processor-a:
        description: "The source build of ppy/osu-queue-score-statistics"
        type: string
        required: false
        default: 'latest'
      score-processor-b:
        description: "The target build of ppy/osu-queue-score-statistics"
        type: string
        required: false
        default: 'latest'

permissions:
  pull-requests: write

env:
  COMMENT_TAG: execution-${{ github.run_id }}-${{ github.run_number }}-${{ github.run_attempt }}

jobs:
  wait-for-queue:
    name: "Wait for previous workflows"
    runs-on: ubuntu-latest
    if: ${{ !cancelled() && (github.event_name == 'workflow_dispatch' || contains(github.event.comment.body, '!diffcalc') && github.event.comment.author_association == 'OWNER') }}
    timeout-minutes: 50400 # 35 days, the maximum for jobs.
    steps:
      - uses: ahmadnassri/action-workflow-queue@v1
        with:
          timeout: 2147483647 # Around 24 days, maximum supported.
          delay: 120000 # Poll every 2 minutes. API seems fairly low on this one.

  create-comment:
    name: Create PR comment
    runs-on: ubuntu-latest
    if: ${{ github.event_name == 'issue_comment' && github.event.issue.pull_request && contains(github.event.comment.body, '!diffcalc') && github.event.comment.author_association == 'OWNER' }}
    steps:
      - name: Create comment
        uses: thollander/actions-comment-pull-request@v2
        with:
          comment_tag: ${{ env.COMMENT_TAG }}
          message: |
            Difficulty calculation queued -- please wait! (${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }})

            *This comment will update on completion*            

  directory:
    name: Prepare directory
    needs: wait-for-queue
    runs-on: self-hosted
    if: ${{ !cancelled() && (github.event_name == 'workflow_dispatch' || contains(github.event.comment.body, '!diffcalc') && github.event.comment.author_association == 'OWNER') }}
    outputs:
      GENERATOR_DIR: ${{ steps.set-outputs.outputs.GENERATOR_DIR }}
      GENERATOR_ENV: ${{ steps.set-outputs.outputs.GENERATOR_ENV }}
      GOOGLE_CREDS_FILE: ${{ steps.set-outputs.outputs.GOOGLE_CREDS_FILE }}
    steps:
      - name: Checkout
        uses: actions/checkout@v3

      - name: Checkout diffcalc-sheet-generator
        uses: actions/checkout@v3
        with:
          path: 'diffcalc-sheet-generator'
          repository: 'smoogipoo/diffcalc-sheet-generator'

      - name: Set outputs
        id: set-outputs
        run: |
          echo "GENERATOR_DIR=${{ github.workspace }}/diffcalc-sheet-generator" >> "${GITHUB_OUTPUT}"
          echo "GENERATOR_ENV=${{ github.workspace }}/diffcalc-sheet-generator/.env" >> "${GITHUB_OUTPUT}"
          echo "GOOGLE_CREDS_FILE=${{ github.workspace }}/diffcalc-sheet-generator/google-credentials.json" >> "${GITHUB_OUTPUT}"          

  environment:
    name: Setup environment
    needs: directory
    runs-on: self-hosted
    if: ${{ !cancelled() && needs.directory.result == 'success' }}
    env:
      VARS_JSON: ${{ toJSON(vars) }}
    steps:
      - name: Add base environment
        run: |
          # Required by diffcalc-sheet-generator
          cp '${{ github.workspace }}/diffcalc-sheet-generator/.env.sample' "${{ needs.directory.outputs.GENERATOR_ENV }}"

          # Add Google credentials
          echo '${{ secrets.DIFFCALC_GOOGLE_CREDENTIALS }}' | base64 -d > "${{ needs.directory.outputs.GOOGLE_CREDS_FILE }}"

          # Add repository variables
          echo "${VARS_JSON}" | jq -c '. | to_entries | .[]' | while read -r line; do
              opt=$(jq -r '.key' <<< ${line})
              val=$(jq -r '.value' <<< ${line})

              if [[ "${opt}" =~ ^DIFFCALC_ ]]; then
                optNoPrefix=$(echo "${opt}" | cut -d '_' -f2-)
                sed -i "s;^${optNoPrefix}=.*$;${optNoPrefix}=${val};" "${{ needs.directory.outputs.GENERATOR_ENV }}"
              fi
          done          

      - name: Add pull-request environment
        if: ${{ github.event_name == 'issue_comment' && github.event.issue.pull_request }}
        run: |
          sed -i "s;^OSU_B=.*$;OSU_B=${{ github.event.issue.pull_request.url }};" "${{ needs.directory.outputs.GENERATOR_ENV }}"          

      - name: Add comment environment
        if: ${{ github.event_name == 'issue_comment' }}
        run: |
          # Add comment environment
          echo '${{ github.event.comment.body }}' | sed -r 's/\r$//' | grep -E '^\w+=' | while read -r line; do
              opt=$(echo ${line} | cut -d '=' -f1)
              sed -i "s;^${opt}=.*$;${line};" "${{ needs.directory.outputs.GENERATOR_ENV }}"
          done          

      - name: Add dispatch environment
        if: ${{ github.event_name == 'workflow_dispatch' }}
        run: |
          sed -i 's;^OSU_B=.*$;OSU_B=${{ inputs.osu-b }};' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          sed -i 's/^RULESET=.*$/RULESET=${{ inputs.ruleset }}/' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          sed -i 's/^GENERATORS=.*$/GENERATORS=${{ inputs.generators }}/' "${{ needs.directory.outputs.GENERATOR_ENV }}"

          if [[ '${{ inputs.osu-a }}' != 'latest' ]]; then
              sed -i 's;^OSU_A=.*$;OSU_A=${{ inputs.osu-a }};' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          fi

          if [[ '${{ inputs.difficulty-calculator-a }}' != 'latest' ]]; then
              sed -i 's;^DIFFICULTY_CALCULATOR_A=.*$;DIFFICULTY_CALCULATOR_A=${{ inputs.difficulty-calculator-a }};' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          fi

          if [[ '${{ inputs.difficulty-calculator-b }}' != 'latest' ]]; then
              sed -i 's;^DIFFICULTY_CALCULATOR_B=.*$;DIFFICULTY_CALCULATOR_B=${{ inputs.difficulty-calculator-b }};' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          fi

          if [[ '${{ inputs.score-processor-a }}' != 'latest' ]]; then
              sed -i 's;^SCORE_PROCESSOR_A=.*$;SCORE_PROCESSOR_A=${{ inputs.score-processor-a }};' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          fi

          if [[ '${{ inputs.score-processor-b }}' != 'latest' ]]; then
              sed -i 's;^SCORE_PROCESSOR_B=.*$;SCORE_PROCESSOR_B=${{ inputs.score-processor-b }};' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          fi

          if [[ '${{ inputs.converts }}' == 'true' ]]; then
              sed -i 's/^NO_CONVERTS=.*$/NO_CONVERTS=0/' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          else
              sed -i 's/^NO_CONVERTS=.*$/NO_CONVERTS=1/' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          fi

          if [[ '${{ inputs.ranked-only }}' == 'true' ]]; then
              sed -i 's/^RANKED_ONLY=.*$/RANKED_ONLY=1/' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          else
              sed -i 's/^RANKED_ONLY=.*$/RANKED_ONLY=0/' "${{ needs.directory.outputs.GENERATOR_ENV }}"
          fi          

  scores:
    name: Setup scores
    needs: [ directory, environment ]
    runs-on: self-hosted
    if: ${{ !cancelled() && needs.environment.result == 'success' }}
    steps:
      - name: Query latest data
        id: query
        run: |
          ruleset=$(cat ${{ needs.directory.outputs.GENERATOR_ENV }} | grep -E '^RULESET=' | cut -d '=' -f2-)
          performance_data_name=$(curl -s "https://data.ppy.sh/" | grep "performance_${ruleset}_top_1000\b" | tail -1 | awk -F "'" '{print $2}' | sed 's/\.tar\.bz2//g')

          echo "TARGET_DIR=${{ needs.directory.outputs.GENERATOR_DIR }}/sql/${ruleset}" >> "${GITHUB_OUTPUT}"
          echo "DATA_NAME=${performance_data_name}" >> "${GITHUB_OUTPUT}"          

      - name: Restore cache
        id: restore-cache
        uses: maxnowack/local-cache@v1
        with:
          path: ${{ steps.query.outputs.DATA_NAME }}.tar.bz2
          key: ${{ steps.query.outputs.DATA_NAME }}

      - name: Download
        if: steps.restore-cache.outputs.cache-hit != 'true'
        run: |
          wget -q -nc "https://data.ppy.sh/${{ steps.query.outputs.DATA_NAME }}.tar.bz2"          

      - name: Extract
        run: |
          tar -I lbzip2 -xf "${{ steps.query.outputs.DATA_NAME }}.tar.bz2"
          rm -r "${{ steps.query.outputs.TARGET_DIR }}"
          mv "${{ steps.query.outputs.DATA_NAME }}" "${{ steps.query.outputs.TARGET_DIR }}"          

  beatmaps:
    name: Setup beatmaps
    needs: directory
    runs-on: self-hosted
    if: ${{ !cancelled() && needs.directory.result == 'success' }}
    steps:
      - name: Query latest data
        id: query
        run: |
          beatmaps_data_name=$(curl -s "https://data.ppy.sh/" | grep "osu_files" | tail -1 | awk -F "'" '{print $2}' | sed 's/\.tar\.bz2//g')

          echo "TARGET_DIR=${{ needs.directory.outputs.GENERATOR_DIR }}/beatmaps" >> "${GITHUB_OUTPUT}"
          echo "DATA_NAME=${beatmaps_data_name}" >> "${GITHUB_OUTPUT}"          

      - name: Restore cache
        id: restore-cache
        uses: maxnowack/local-cache@v1
        with:
          path: ${{ steps.query.outputs.DATA_NAME }}.tar.bz2
          key: ${{ steps.query.outputs.DATA_NAME }}

      - name: Download
        if: steps.restore-cache.outputs.cache-hit != 'true'
        run: |
          wget -q -nc "https://data.ppy.sh/${{ steps.query.outputs.DATA_NAME }}.tar.bz2"          

      - name: Extract
        run: |
          tar -I lbzip2 -xf "${{ steps.query.outputs.DATA_NAME }}.tar.bz2"
          rm -r "${{ steps.query.outputs.TARGET_DIR }}"
          mv "${{ steps.query.outputs.DATA_NAME }}" "${{ steps.query.outputs.TARGET_DIR }}"          

  generator:
    name: Run generator
    needs: [ directory, environment, scores, beatmaps ]
    runs-on: self-hosted
    timeout-minutes: 720
    if: ${{ !cancelled() && needs.scores.result == 'success' && needs.beatmaps.result == 'success' }}
    outputs:
      TARGET: ${{ steps.run.outputs.TARGET }}
      SPREADSHEET_LINK: ${{ steps.run.outputs.SPREADSHEET_LINK }}
    steps:
      - name: Run
        id: run
        run: |
          # Add the GitHub token. This needs to be done here because it's unique per-job.
          sed -i 's/^GH_TOKEN=.*$/GH_TOKEN=${{ github.token }}/' "${{ needs.directory.outputs.GENERATOR_ENV }}"

          cd "${{ needs.directory.outputs.GENERATOR_DIR }}"
          docker-compose up --build generator

          link=$(docker-compose logs generator -n 10 | grep 'http' | sed -E 's/^.*(http.*)$/\1/')
          target=$(cat "${{ needs.directory.outputs.GENERATOR_ENV }}" | grep -E '^OSU_B=' | cut -d '=' -f2-)

          echo "TARGET=${target}" >> "${GITHUB_OUTPUT}"
          echo "SPREADSHEET_LINK=${link}" >> "${GITHUB_OUTPUT}"          

      - name: Shutdown
        if: ${{ always() }}
        run: |
          cd "${{ needs.directory.outputs.GENERATOR_DIR }}"
          docker-compose down          

      - name: Output info
        if: ${{ success() }}
        run: |
          echo "Target: ${{ steps.run.outputs.TARGET }}"
          echo "Spreadsheet: ${{ steps.run.outputs.SPREADSHEET_LINK }}"          

  update-comment:
    name: Update PR comment
    needs: [ create-comment, generator ]
    runs-on: ubuntu-latest
    if: ${{ github.event_name == 'issue_comment' && github.event.issue.pull_request && contains(github.event.comment.body, '!diffcalc') && github.event.comment.author_association == 'OWNER' }}
    steps:
      - name: Update comment on success
        if: ${{ needs.generator.result == 'success' }}
        uses: thollander/actions-comment-pull-request@v2
        with:
          comment_tag: ${{ env.COMMENT_TAG }}
          mode: upsert
          create_if_not_exists: false
          message: |
            Target: ${{ needs.generator.outputs.TARGET }}
            Spreadsheet: ${{ needs.generator.outputs.SPREADSHEET_LINK }}            

      - name: Update comment on failure
        if: ${{ needs.generator.result == 'failure' }}
        uses: thollander/actions-comment-pull-request@v2
        with:
          comment_tag: ${{ env.COMMENT_TAG }}
          mode: upsert
          create_if_not_exists: false
          message: |
            Difficulty calculation failed: ${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}