Sourcesage

🚀 SourceSage: Efficient Code Memory for Large Language Models (LLMs)

SourceSage is a server based on the Model Context Protocol (MCP), designed to assist Large Language Models (LLMs) in efficiently understanding and memorizing key information in codebases. It supports multiple programming languages and focuses on storing and retrieving code semantics, entities, and their relationships.

🚀 Quick Start

SourceSage is a powerful tool that helps LLMs better handle code-related tasks. By leveraging the understanding ability of LLMs, it can efficiently process and store code information.

✨ Features

1. Leverage LLM's Understanding Ability: SourceSage utilizes the LLM's understanding of code semantics and can handle multiple programming languages.
2. Store Semantic Knowledge: Different from traditional code analysis tools, SourceSage focuses on storing the semantics of code and the relationships between entities.
3. Language Neutral: Supports any programming language that the LLM can understand.
4. Optimize Storage Efficiency: Adopts an efficient storage mechanism to minimize token usage.
5. Evolve with LLM Technology: As the capabilities of LLMs improve, SourceSage can also enhance its code understanding and analysis accordingly.

📦 Installation

Installation Steps

1. Clone the Repository:

   git clone https://github.com/your-repository.git
   cd source-sage
   

2. Run the Server:

   python3 -m sourcesage.server
   

💻 Usage Examples

Basic Usage

1. Analyze Code: Analyze your code files through the LLM.

   "Please analyze this Python file and register the key entities and relationships in it."
   

2. Register Entities: The LLM will use the register_entity tool to store entities in the code.

   "I will register the main class in this file."
   

3. Query Knowledge: You can query the codebase in the following ways later:
   • Get the list of classes defined in the codebase:

     "Please tell me what classes are in my codebase?"
     

   • View detailed information about a specific class:

     "Please show the detailed information of the `User` class."
     

4. Get Coding Patterns: You can query the patterns and conventions used in the code.

   "Please tell me what design patterns are used in my codebase?"
   

Advanced Usage

The advanced usage of SourceSage mainly involves more complex entity registration and relationship recording. For example:

1. Register Relationships:

   "I will record the inheritance relationships between these classes."
   

📚 Documentation

Core Functional Modules

1. Register Entities:
   • register_entity: Used to store key entities (such as classes, functions, etc.) in the code.
   • register_relationship: Used to record the relationships between entities (such as inheritance, dependency, etc.).
2. Query Knowledge:
   • query_entities: Query entities according to language, name pattern, or other conditions.
   • get_entity_details: Get detailed information about a specific entity.
   • query_patterns: Query code patterns by programming language or pattern name.
   • query_style_conventions: Query coding styles by programming language or convention name.
3. Statistics and Cleaning:
   • get_knowledge_statistics: Get statistical data of the knowledge graph.
   • clear_knowledge: Clear all stored knowledge.

🤝 Project Contribution

Community contributions are welcome! If you have any questions or suggestions, please feel free to submit a Pull Request.

📄 License

This project is licensed under the MIT License. Please refer to the LICENSE file for specific details.