code-semantic-labeler

Automate code semantic labeling by mapping code nodes to urban business semantics using rule-based and LLM-based approaches. This process applies AST analysis to extract core variables and function calls, providing semantic understanding for large urban brain codebases.

AST analysis uses the code-to-ast-new library to parse code and extract core variables and function calls. When rules do not match, it summarizes the code to provide an overall semantic understanding of the codebase structure and logic.

Yes, you need Python 3.7+ along with the abc, ast, json, os, sys, and code-to-ast-new libraries to perform code analysis and semantic labeling. These dependencies are required for the AST analysis and semantic mapping processes.

Yes, code summarization is automatically applied when rule-based semantic labeling fails to match. It provides an overall semantic understanding of the code by summarizing the extracted AST nodes and function calls.

The best way to map code nodes to urban business semantics is by combining rule-based matching with LLM-based code summarization. This dual approach ensures accurate semantic labeling even when predefined rules do not cover specific code patterns.

Rule-based code semantic labeling is limited by its predefined rules, which may not cover all code patterns. When rules fail to match, the system falls back to code summarization to provide an overall semantic understanding of the code.