shibing624/TreeSearch

Builds hierarchical tree structures from documents using language-specific parsers (Python AST, regex for Java/Go/JS/C++, heading extraction for Markdown) and indexes them with SQLite FTS5, eliminating the need for embeddings or vector databases. Supports mixed input types (directories, files, globs) with auto-discovery and `.gitignore` respect, plus dual search modes—tree traversal for hierarchical documents and flat FTS5 for code. Includes ripgrep-accelerated GrepFilter for precise symbol matching and optional extractors for PDF/DOCX/HTML via optional dependencies.