StarShip CodeReviewer
Welcome to the StarShip CodeReviewer
, a next-generation LLM (Large Language Model)-based code reviewing copilot engineered to streamline and enhance the software development workflow. StarShip CodeReviewer
is designed to assist developers in navigating the complexities of code review by providing comprehensive support in analyzing, understanding, and improving merge requests.
Features
StarShip CodeReviewer
empowers developers to efficiently manage their code review process with three functionalities:
Review Merge Requests
The tool meticulously scans merge requests to identify potential issues, offering actionable suggestions for improvement.
For more details: Dive deeper into the review capabilities in the Review Feature Documentation.
Describe Merge Requests
The Describe feature provides developers with a concise, easy-to-understand summary of the merge requests, ensuring you grasp the full context of the changes without digging through lines of code.
For more details: Explore the Description Feature Documentation to learn how we can help you quickly comprehend code modifications.
Evaluate Merge Requests
The Evaluate feature is to ensure your code meets the highest standards. By assessing the code quality from ten dimensions and various angles,
StarShip CodeReviewer
not only gives a comprehensive score based on LLM evaluation but also explains the rationale behind it.For more details: Check out the Evaluation Feature Documentation for an in-depth look into the evaluation process.
Multi-language Code Static Analysis
'StarShip CodeReviewer' integrates linter tools for multiple programming languages to perform in-depth scans of merge request (MR) code. It ensures that the code adheres to standards in terms of style consistency, syntax correctness, and best practices. Additionally, it provides targeted improvement suggestions to help enhance code quality. For more details, please refer to the Multi-language Code Static Analysis documentation.
Why StarShip CodeReviewer?
In the vast sea of code, StarShip CodeReviewer
stands out as your loyal navigational aid, guiding you towards safer, more efficient, and higher quality software development. With StarShip CodeReviewer
, you leverage the power of LLM to:
- Enhance Efficiency: Reduce the time spent on manual code reviews without compromising quality.
- Improve Code Quality: Receive insightful feedback and suggestions to refine and optimize your code.
- Understand Changes Quickly: Get up to speed with changes through succinct, accurate descriptions.
- Make Informed Decisions: Evaluate the quality of code using a comprehensive and explainable scoring system.
Getting Started
Installation
For comprehensive instructions on setting up your system, please refer to the Installation Guide.
Trigger StarShip CodeReviewer
To engage the StarShip CodeReviewer
, there are two primary mechanisms available for activation:
Automatic Trigger: This mechanism springs into action with the creation of a new Merge Request (MR) that encompasses code changes. Automatically detect code changes in new MR and begin its review process without manual intervention.
Manual Trigger: For instances where you require a review at a specific juncture, or for an already opened MR that needs further examination, the manual trigger comes into play. By simply appending a note within the MR that begins with
@codegpt
, you can manually initiate theStarShip CodeReviewer
. For how to manual triggerStarShip CodeReviewer
, please refer the Trigger Guide.
Language Support
Automated Language Detection: If the title or the description of the MR contains Chinese characters,
StarShip CodeReviewer
automatically switches to Chinese for its review outputs. This ensures that the feedback is accessible and easily comprehensible to developers who prefer Chinese for professional communication.Default Language Setting: In the absence of any Chinese characters in the MR title or description,
StarShip CodeReviewer
will default to English for its review comments and suggestions.
Limitations: For locally on-promise LLM, currently only support English since the local LLM may not good for Chinses.