The improve
tool scans the PR code changes, and automatically generates committable suggestions for improving the PR code.
It can be invoked manually by commenting on any PR:
/improve
For example:
The improve
tool can also be triggered automatically every time a new PR is opened. See examples for automatic triggers for GitHub App and GitHub Action
An extended mode, which does not involve PR Compression and provides more comprehensive suggestions, can be invoked by commenting on any PR:
/improve --extended
Note that the extended mode divides the PR code changes into chunks, up to the token limits, where each chunk is handled separately (multiple calls to GPT-4). Hence, the total number of suggestions is proportional to the number of chunks, i.e., the size of the PR.
Under the section 'pr_code_suggestions', the configuration file contains options to customize the 'improve' tool:
num_code_suggestions
: number of code suggestions provided by the 'improve' tool. Default is 4.extra_instructions
: Optional extra instructions to the tool. For example: "focus on the changes in the file X. Ignore change in ...".rank_suggestions
: if set to true, the tool will rank the suggestions, based on importance. Default is false.include_improved_code
: if set to true, the tool will include an improved code implementation in the suggestion. Default is true.
num_code_suggestions_per_chunk
: number of code suggestions provided by the 'improve' tool, per chunk. Default is 8.rank_extended_suggestions
: if set to true, the tool will rank the suggestions, based on importance. Default is true.max_number_of_calls
: maximum number of chunks. Default is 5.final_clip_factor
: factor to remove suggestions with low confidence. Default is 0.9.
summarize
: if set to true, the tool will present the code suggestions in a compact way. Default is false.
In this mode, instead of presenting committable suggestions, the different suggestions will be combined into a single compact comment, with significantly smaller PR footprint.
For example:
/improve --pr_code_suggestions.summarize=true
-
While the current AI for code is getting better and better (GPT-4), it's not flawless. Not all the suggestions will be perfect, and a user should not accept all of them automatically.
-
Suggestions are not meant to be simplistic. Instead, they aim to give deep feedback and raise questions, ideas and thoughts to the user, who can then use his judgment, experience, and understanding of the code base.
-
Recommended to use the 'extra_instructions' field to guide the model to suggestions that are more relevant to the specific needs of the project.
-
Best quality will be obtained by using 'improve --extended' mode.