CodeSpector

CodeSpector is a Python package designed to review code changes for quality and security issues using AI chat agents. It supports different chat agents like Codestral and ChatGPT.

Features

• Automated code review using AI chat agents.
• Supports multiple chat agents and models.
• Generates detailed review reports in markdown format.
• Configurable via environment variables and command-line options.

Installation

To install the package, use the following command:

pip install codespector

uv add codespector

Usage

Command-Line Interface

You can use the codespector command to start a code review. Below are the available options:

Usage: codespector [OPTIONS]

Options:
  --chat-token TEXT          Chat agent token  [env var: CODESPECTOR_CHAT_TOKEN]
  --chat-model TEXT          Choose the chat model to use  [env var: CODESPECTOR_CHAT_MODEL]
  --chat-agent TEXT          Choose the chat agent to use (codestral, chatgpt, deepseek)
                            or set your own  [env var: CODESPECTOR_CHAT_AGENT]
  -b, --compare-branch TEXT  Select the branch to compare the current one with
  --output-dir TEXT         Select the output directory  [default: codespector]
                            [env var: CODESPECTOR_OUTPUT_DIR]
  --system-content TEXT     Content which used in system field for agent
                            [env var: CODESPECTOR_SYSTEM_CONTENT]
  --prompt-content TEXT     Prompt content which included to review prompt
                            [env var: CODESPECTOR_PROMPT_CONTENT]
  --result-file TEXT        Set file for saving the result
                            [env var: CODESPECTOR_RESULT_FILE]
  --exclude-file-ext LIST   Exclude file extensions from the review
                            [env var: CODESPECTOR_EXCLUDE_FILE_EXT]
  --version                 Show the version and exit.
  --help                    Show this message and exit.

Example

To run a code review, use the following command:

codespector --chat-token YOUR_CHAT_TOKEN --chat-agent codestral --compare-branch develop --result-file result.md --system-content "system content" --prompt-content "prompt content"

Configuration

You can also configure CodeSpector using environment variables. Create a .env file in the root directory of your project with the following content:

CODESPECTOR_CHAT_TOKEN=your_token
CODESPECTOR_OUTPUT_DIR=codespector
CODESPECTOR_SYSTEM_CONTENT="Ты код ревьювер. Отвечай на русском"
CODESPECTOR_PROMPT_CONTENT="Оцени код на безопасноть, соблюдение лучших техник"
CODESPECTOR_RESULT_FILE="result.md"
CODESPECTOR_CHAT_AGENT=codestral
CODESPECTOR_CHAT_MODEL=your_model
CODESPECTOR_EXCLUDE_FILE_EXT=.pyc,.pyo

Makefile Commands

• lint: Run linting and formatting checks.
• format: Format the code.
• fix: Fix linting issues and format the code.
• test: Run the tests.

License

This project is licensed under the MIT License. See the LICENSE file for more details.