My Project Debugger

A CLI tool for analyzing code and errors using LLM (Language Model) capabilities, with a Godspeed backend for handling LLM integration.

Features

• Analyze code and error logs with AI assistance.
• Interactive chat sessions for real-time debugging.
• Continue previous sessions to maintain conversation history.
• Automated repomap generation for project context.
• Save and load analysis context for future reference.

CLI Options

The CLI supports various options for its commands. For example:

General Options

# Specify a directory to work in and continue an existing session
code-help --chat --directory /path/to/project --continue

# You can also specify files in the command line itself :)
code-help chat --file <filename1> <filename2> ......  

All available commands and their subcommands can be found simply from code-help --help which will help you get started :)

Using the cli tool

Project Structure

• frontend/ - CLI application written in TypeScript.
• backend/ - Godspeed server with LLM integration.

Setup

Backend Setup

1. Navigate to the backend directory:

cd backend
npm install 
godspeed serve

2. Ensure you have tokenjs.yaml configured for your datasource.

3. Configure your LLM settings as per token.js docs:

type: tokenjs
config:
  provider: "your-provider"
  model: "your-model"

For example, using an ollama model:

type: tokenjs
config:
  provider: openai-compatible
  baseURL: http://localhost:11434/v1
  models:
    - name: <model_name>
      config:
        temperature: 0.7
        max_tokens: 1000

For more details, refer to the tokenjs docs: https://docs.tokenjs.ai/providers

Frontend Setup

1. Navigate to the frontend directory:

cd frontend
npm install
npm run build

2. Launch the CLI tool:

code-help

Data Storage

All session history and project context are automatically saved under the .superdebugger folder in your specified directory.

Backend Routes

The backend exposes the following routes:

• POST /api/code/context
  Processes code analysis with LLM.
  Request Body:

  {
    "analysisContext": {
      "files": [
        { "name": "file.ts", "content": "file content" }
      ],
      "errorLog": "Full error log here",
      "projectContext": "Formatted project context",
      "timestamp": "ISO8601 timestamp"
    },
    "prompt": "User question"
  }

  Responses:
  200 - Returns answer, model and usage stats
  400 - Invalid request
  503 - Service unavailable

• POST /repomap
  Generates a repository map based on the provided Git repository URL.
  Request Body:

  {
    "gitUrl": "https://github.com/example/repo.git"
  }

  Response:
  Returns a structured repo map.

Credits

Repomap functionality is based on a pagerank algorithm. Special thanks to Paul Gauthier Aider-AI for his python code, which was ported to TypeScript for this project.