action.yml

name: 'DevToolsAi reviewer'
description: 'Inspired by the AI-based PR Reviewer & Summarizer with Chat Capabilities by CodeRabbit'
inputs:
  debug:
    required: false
    description: 'Enable debug mode'
    default: 'false'
  max_files:
    required: false
    description:
      'Max files to summarize and review. Less than or equal to 0 means no
      limit.'
    default: '150'
  review_simple_changes:
    required: false
    description: 'Review even when the changes are simple'
    default: 'false'
  review_comment_lgtm:
    required: false
    description: 'Leave comments even if the patch is LGTM'
    default: 'false'
  path_filters:
    required: false
    description: |
      The path filters, e.g., "src/**.py", "!dist/**", each line will be considered as one pattern.
      See also

      - https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#onpushpull_requestpull_request_targetpathspaths-ignore
      - https://github.com/isaacs/minimatch
    default: |
      !dist/**
      !**/*.3g2
      !**/*.3gp
      !**/*.7z
      !**/*.app
      !**/*.avi
      !**/*.arxml
      !**/*.bin
      !**/*.bmp
      !**/*.bz2
      !**/*.BUILD
      !**/*.cfg
      !**/*.class
      !**/*.csv
      !**/*.dat
      !**/*.db
      !**/*.dll
      !**/*.doc
      !**/*.docx
      !**/*.dot
      !**/*.dylib
      !**/*.egg
      !**/*.eot
      !**/*.exe
      !**/*.flac
      !**/*.flv
      !**/*.gif
      !**/*.glif
      !**/*.gz
      !**/*.ico
      !**/*.ini
      !**/*.iso
      !**/*.jar
      !**/*.jpeg
      !**/*.jpg
      !**/*.json
      !**/*.lo
      !**/*.lock
      !**/*.log
      !**/*.m4a
      !**/*.m4v
      !**/*.md5sum
      !**/*.min.js
      !**/*.min.js.css
      !**/*.min.js.map
      !**/*.mkv
      !**/*.mmd
      !**/*.mod
      !**/*.mov
      !**/*.mp3
      !**/*.mp4
      !**/*.nar
      !**/*.o
      !**/*.ogg
      !**/*.otf
      !**/*.p
      !**/*.parquet
      !**/*.pb.go
      !**/*.pdf
      !**/*.pem
      !**/*.pickle
      !**/*.pkl
      !**/*.png
      !**/*.ppt
      !**/*.pptx
      !**/*.pub
      !**/*.pyc
      !**/*.pyd
      !**/*.pyo
      !**/*.rar
      !**/*.rkt
      !**/*.rm
      !**/*.snap
      !**/*.so
      !**/*.ss
      !**/*.sum
      !**/*.svg
      !**/*.swf
      !**/*.tar
      !**/*.tfstate
      !**/*.tfstate.backup
      !**/*.tiff
      !**/*.toml
      !**/*.tsv
      !**/*.ttf
      !**/*.war
      !**/*.wasm
      !**/*.wav
      !**/*.webm
      !**/*.wma
      !**/*.wmv
      !**/*.woff
      !**/*.woff2
      !**/*.work
      !**/*.xls
      !**/*.xlsx
      !**/*.xz
      !**/*.yaml
      !**/*.yml
      !**/*.zip
      !**/*.zst
      !**/@generated/**
      !**/_gen/**
      !**/gen/**
      !**/generated/**
      !**/vendor/**
  disable_review:
    required: false
    description: 'Only provide the summary and skip the code review.'
    default: 'false'
  disable_release_notes:
    required: false
    description: 'Disable release notes'
    default: 'false'
  api_base_url_azure:
    required: false
    description: 'The url of the ML cluster api interface.'
    default: |
      Mistral-small-azure.swedencentral.models.ai.azure.com
      Mistral-large-azure.swedencentral.models.ai.azure.com
  light_model_name_azure:
    required: false
    description:
      'Model to use for simple tasks like summarizing diff on a file. On severless Azure.'
    default: 'mistral-small-azure'
  heavy_model_name_azure:
    required: false
    description: 'Model to use for complex tasks such as code reviews. On severless Azure.'
    default: 'mistral-large-azure'
  light_model_token_azure:
    required: false
    description: 'Token for the light model on Azure'
    default: ''
  heavy_model_token_azure:
    required: false
    description: 'Token for the heavy model on Azure'
    default: ''
  api_base_url:
    required: false
    description: 'The url of the ML cluster api interface.'
    default: |
      atlas.intra.chrysler.com:41443
      apps.ai-infra.eu-xp.stla-aws.private
  light_model_name:
    required: false
    description:
      'Model to use for simple tasks like summarizing diff on a file.'
    default: 'small'
  light_model_port:
    required: false
    description: 'Port for the light model'
    default: '44901'
  heavy_model_name:
    required: false
    description: 'Model to use for complex tasks such as code reviews.'
    default: 'big'
  heavy_model_port:
    required: false
    description: 'Port for the heavy model'
    default: '44902'
  model_temperature:
    required: false
    description: 'Temperature for language model'
    default: '0.1'
  retries:
    required: false
    description:
      'How many times to retry in case of timeouts or errors?'
    default: '5'
  timeout_ms:
    required: false
    description: 'Timeout for API call in millis'
    default: '360000'
  concurrency_limit:
    required: false
    description: 'How many concurrent API calls to make to ML cluster servers?'
    default: '6'
  github_concurrency_limit:
    required: false
    description: 'How many concurrent API calls to make to GitHub?'
    default: '6'
  system_message:
    required: false
    description: 'System message to be sent to model'
    default: |
      You are `@devtoolsai` (aka `github-actions[bot]`), a language model.
      Your purpose is to act as a highly experienced 
      software engineer and provide a thorough review of the code hunks
      and suggest code snippets to improve key areas such as:
        - Logic
        - Security
        - Performance
        - Data races
        - Consistency
        - Error handling
        - Maintainability
        - Modularity
        - Complexity
        - Optimization
        - Best practices: DRY, SOLID, KISS

      Do not comment on minor code style issues, missing comments,
      or documentation. Focus only on significant, unintentional
      concerns impacting code functionality or performance.
      Deliberate code changes, even if they alter structure or
      variable usage, should be assumed correct if tests are
      included. Only flag changes that are likely to introduce
      actual issues without evident coverage or verification.
  summarize:
    required: false
    description: 'The prompt for final summarization response'
    default: |
      Provide your final response in markdown with the following content:

      - **Walkthrough**: A high-level summary of the overall change instead of 
        specific files within 80 words. Please don't include any code snippets, any comments, any suggestions.
        It must stay a high-level summary of the changes.
      - **Changes**: A markdown table of files and their summaries. Group files 
        with similar changes together into a single row to save space. 
        Files with the same name must be grouped together.
      - **Poem**: Below the changes, include a whimsical, short poem written by 
        a devtoolsai to celebrate the changes. Write only one quatrain.
        Format the poem as a quote using the ">" symbol and feel free to use emojis where relevant.

      Avoid additional commentary as this summary will be added as a comment on the 
      GitHub pull request. Use the titles "Walkthrough" and "Changes" and they must be H2.

  summarize_release_notes:
    required: false
    description:
      'The prompt for generating release notes in the same chat as summarize
      stage'
    default: |
      Craft concise release notes for the pull request. 
      Focus on the purpose and user impact, categorizing changes as "New Feature", "Bug Fix", 
      "Documentation", "Refactor", "Style", "Test", "Chore", or "Revert". Provide a bullet-point list, 
      e.g., "- New Feature: Added search functionality to the UI". Limit your response to 50-100 words 
      and emphasize features visible to the end-user while omitting code-level details.
      DONT propose any code changes or any comments or any recommendations in this stage.
  language:
    required: false
    description: ISO code for the response language
    default: en-US
  bot_icon:
    required: false
    description: 'The icon for the bot'
    default: '<img src="https://avatars.githubusercontent.com/u/124881756" alt="Image description" width="20" height="20">'
  allow_empty_review:
    required: false
    description: 'Allow empty review.'
    default: 'false'
  less_spammy:
    required: false
    description: 'Will remove all bot suggestions, if there is no conversation. Also change request review message.'
    default: 'true'

runs:
  using: 'composite'
  steps:
    - name: Generate hash from requirements.txt
      id: generate-hash
      run: |
        echo "requirements_hash=$(sha256sum ${GITHUB_ACTION_PATH}/requirements.txt | awk '{ print $1 }')" >> $GITHUB_ENV
        ls ${GITHUB_ACTION_PATH}
        cat ${GITHUB_ACTION_PATH}/requirements.txt
      shell: bash

    - name: Cache dependencies
      uses: actions/cache@v3
      id: cache-python
      with:
        path: ~/_work/_tool/
        key: ${{ runner.os }}-pr-reviewer-python-${{ env.requirements_hash }}
        restore-keys: |
          ${{ runner.os }}-pr-reviewer-python-${{ env.requirements_hash }}

    - name: Setup Python
      uses: actions/setup-python@v4
      id: setup-python
      with:
        python-version: 3.11
        update-environment: false

    - name: Cache Requirements
      uses: actions/cache@v3
      id: cache-requirements
      with:
        path: ~/_work/_actions/opensource/ai-pr-reviewer/
        key: ${{ runner.os }}-pr-reviewer-requirements-${{ env.requirements_hash }}
        restore-keys: |
          ${{ runner.os }}-pr-reviewer-requirements-${{ env.requirements_hash }}

    - name: Setup venv
      run: |
        "${{ steps.setup-python.outputs.python-path }}" -m venv "${GITHUB_ACTION_PATH}/venv"
      shell: bash

    - name: Install dependencies
      run: |
        "${GITHUB_ACTION_PATH}/venv/bin/python" -m pip install -r "${GITHUB_ACTION_PATH}/requirements.txt"
      shell: bash

    - name: Run action
      env:
        INPUTS: ${{ toJSON(inputs) }}
      run: |
        source "$GITHUB_ACTION_PATH/venv/bin/activate"
        "${GITHUB_ACTION_PATH}/venv/bin/python" "${GITHUB_ACTION_PATH}/main.py"
      shell: bash

branding:
  icon: 'cpu'
  color: 'green'