Extractors API

Extractors module.

contextbench.extractors.extract_defs(file_path: str) → List[Tuple[str, int, int]][source]: Extract definition nodes from file.

contextbench.extractors.extract_def_set_in_spans(spans_by_file: Dict[str, List[Tuple[int, int]]], repo_dir: str) → Set[Tuple[str, str, int, int]][source]: Extract definitions that overlap with given byte spans. Returns {(file, kind, start_byte, end_byte)}.

contextbench.extractors.extract_def_set_from_symbol_names(pred_symbols_by_file: Dict[str, List[str]], repo_dir: str) → Set[Tuple[str, str, int, int]][source]

Map predicted symbol names to tree-sitter def byte ranges.

Returns {(file, kind, start_byte, end_byte)}.

contextbench.extractors.available() → bool[source]: Return True only if tree-sitter + parsers are usable.