Search papers, labs, and topics across Lattice.
The paper introduces abcoder-ts-parser, a new TypeScript parser for the ABCoder framework that leverages the TypeScript Compiler API to build function-level code indexes (UniAST) for improved LLM code agent retrieval. By directly accessing the compiler's AST and semantic information, the parser avoids the performance bottleneck of JSON-RPC calls inherent in language server-based approaches. Evaluation on large TypeScript repositories shows that abcoder-ts-parser generates reliable indexes much more efficiently than existing methods.
Parsing TypeScript codebases for LLM code agents just got an order of magnitude faster by ditching language servers for direct compiler access.
Graph-based code indexing can improve context retrieval for LLM-based code agents by preserving call chains and dependency relationships that keyword search and similarity retrieval often miss. ABCoder is an open-source framework that parses codebases into a function-level code index called UniAST, but its existing parsers combine lightweight AST parsers for syntactic analysis with language servers for semantic resolution, but because LSP-based resolution requires a JSON-RPC call for each symbol lookup, these per-symbol calls become a bottleneck on large TypeScript repositories. We present abcoder-ts-parser, a TypeScript parser built on the TypeScript Compiler API that works directly with the compiler's AST, semantic information, and module resolution logic. We evaluate the parser on three open-source TypeScript projects with up to 1.2 million lines of code and find that it produces reliable indexes significantly more efficiently than the existing architecture. For a live demonstration, watch: https://youtu.be/ryssr7ouvdE
