platisd/openai-pr-description GitHub Action

Autofill the description of your pull requests with the power of OpenAI!

What does it do?

chalermporn17/openai-pr-description is forked from

`platisd/openai-pr-description` which is a GitHub Action that looks at the title as well as the contents
of your pull request and uses the [OpenAI API](https://openai.com/blog/openai-api) to automatically
fill up the description of your pull request. Just like ChatGPT would! 🎉<br>
The Action tries to focus on **why** the changes are needed rather on **what** they are,
like any proper pull request description should.

The GitHub Action will only run when a PR description is not already provided.
In other words it will not accidentally overwrite your existing description.
The idea is this Action will save you the time and trouble of writing **meaningful** pull request descriptions.<br>
You can customize it in different ways. One of them allows the Action to only run on pull requests started
by specific users, e.g. the main maintainers of the repository.
Keep in mind the OpenAI API is not free to use. That being said, so far it's been rather cheap,
i.e. around ~$0.10 for 15-20 pull requests so far.

New Feature

• model selection base on change which will choose bigger model if pull request if big.
• terminated if there aren't any avialable model for pull request if pull request if too big.
• specific file type. so some trash file doesn't included.

How can you use it?

1. Create an account on OpenAI, set up a payment method and get your OpenAI API key.
2. Add the OpenAI API key as a secret in your repository's settings.
3. Create a workflow YAML file, e.g. .github/workflows/openai-pr-description.yml with the following contents:

Example workflow

name: Autofill PR description

on:
  pull_request:
    branches:
      - dev
      - main
      - master
      
jobs:
  openai-pr-description:
    runs-on: ubuntu-22.04
    timeout-minutes: 10
    steps:
      - uses: chalermporn17/openai-pr-description@master
        with:
          github_token: ${{ secrets.GITHUB_TOKEN }}
          openai_api_key: ${{ secrets.OPENAI_API_KEY }}

Input	Description	Required	Default
github_token	The GitHub token used for the Action	Yes
openai_api_key	The OpenAI API key to use, ensure it remains hidden	Yes
pull_request_id	The ID of the pull request to be used	No	Extracted from metadata
openai_models	The OpenAI model to use, contains a JSON key indicating the model name and value as the max tokens for the model	No	{ "gpt-3.5-turbo": 4096, "gpt-3.5-turbo-16k": 16384}
max_response_tokens	The maximum number of response tokens to be used	No	2048
temperature	Higher values increase model creativity (range: 0-2)	No	0.6
header_sample_prompt	The prompt used to provide context to the model	No	See SAMPLE_PROMPT
sample_response	A sample response used to provide context to the model	No	See GOOD_SAMPLE_RESPONSE
file_types	A comma separated list of file types to include in the prompt.	No	See File types that will include in prompt
saver_mode	If true, will ignore chat guide before the sample prompt and only use the sample prompt as the input to the model	No	false

example pull request chalermporn17#2

File types that will include in prompt

.java,.py,.js,.ts,.cpp,.c,.cs,.rb,.go,.php,.swift,.html,.htm,.css,.scss,.sass,.less,.xml,.json,.yaml,.yml,.sh,.bat,.ps1,.cfg,.conf,.ini,.md,.txt,.rtf,.pdf,.sql,.db,.sqlite,.xaml,.plist

too include or exclude file type, you can edit file_types in action argument.

Default sample prompt

• SAMPLE_PROMPT
  • input prompt will generate from header_sample_prompt concat with pull request file change
  • default header_sample_prompt

    Write a pull request description , describe the summary of change.
    Go straight to the point.
    
    answer in below format
    
    ## Overview
    tell overview here
    ## Changes Made
    - **Header**: Description
    - **Header**: Description
    ## Impact
    - **Header**: Description
    - **Header**: Description
    
    content of pull request is below
    

  • example of generated pull request file change

    The title of the pull request is "Enable valgrind on CI" and the following changes took place: 
    
    Changes in file .github/workflows/build-ut-coverage.yml: 
    @@ -24,6 +24,7 @@ jobs:
            run: |
              sudo apt-get update
              sudo apt-get install -y lcov
    +          sudo apt-get install -y valgrind
              sudo apt-get install -y ${{ matrix.compiler.cc }}
              sudo apt-get install -y ${{ matrix.compiler.cxx }}
          - name: Checkout repository
    @@ -48,3 +49,7 @@ jobs:
            with:
              files: coverage.info
              fail_ci_if_error: true
    +      - name: Run valgrind
    +        run: |
    +          valgrind --tool=memcheck --leak-check=full --leak-resolution=med \
    +            --track-origins=yes --vgdb=no --error-exitcode=1 ${build_dir}/test/command_parser_test
    Changes in file test/CommandParserTest.cpp: @@ -566,7 +566,7 @@ TEST(CommandParserTest, ParsedCommandImpl_WhenArgumentIsSupportedNumericTypeWill
        unsigned long long expectedUnsignedLongLong { std::numeric_limits<unsigned long long>::max() };
        float expectedFloat { -164223.123f }; // std::to_string does not play well with floating point min()
        double expectedDouble { std::numeric_limits<double>::max() };
    -    long double expectedLongDouble { std::numeric_limits<long double>::max() };
    +    long double expectedLongDouble { 123455678912349.1245678912349L };
    

• GOOD_SAMPLE_RESPONSE

  This pull request aims to integrate Valgrind into our Continuous Integration (CI) process to enhance memory leak detection and code reliability.
  
  ## Changes Made
      - **CI Workflow Update**: Included Valgrind installation and configured it to run with our test suite.
      - **Test Suite Adjustments**: Made minor adjustments to the test suite for better compatibility with Valgrind and improved test coverage.
  
  ## Impact
      - **Enhanced Code Quality**: Integrating Valgrind allows for automatic detection of memory leaks and access errors, leading to a more robust codebase.
      - **Improved Testing**: The adjustments in the test suite ensure comprehensive testing and accuracy.
  

Model Selection Process

1. Calculate the total number of tokens used in the model prompt (prompt_tokens), which includes the tokens from the example prompt, example response, and the prompt generated from the pull request.

2. Select an appropriate model based on the following criteria:

   • Ensure that the model's maximum token capacity (model_max_token) is greater than the sum of prompt_tokens and max_response_tokens.
   • If multiple models satisfy this condition, select the one with the smallest maximum token capacity.

403 error when updating the PR description

If you get a 403 error when trying to update the PR description, it's most likely because the GitHub Action is not allowed to do so. The easiest way forward is to grant the necessary permissions to the GITHUB_TOKEN secret at <your_repo_url>/settings/actions under Workflow permissions.