Class ExtractorConfig

explainshell/extraction/types.py:111–132

Shared configuration for all extractors. ``run_dir``: single directory for all run artifacts (logs, debug files, failed responses). When set, debug artifacts land in ``markdown/``, ``prompts/``, ``responses/`` subdirs keyed by an encoded form of the input path (see ``repo_root``).

@dataclass(frozen=True)
class ExtractorConfig:
    """Shared configuration for all extractors.

    ``run_dir``: single directory for all run artifacts (logs, debug
    files, failed responses).  When set, debug artifacts land in
    ``markdown/``, ``prompts/``, ``responses/`` subdirs keyed by an
    encoded form of the input path (see ``repo_root``).

    ``repo_root``: when set, per-page artifact filenames encode the
    .gz file's repo-relative path (slashes replaced with ``__``) so
    same-basename pages from different distros/sections don't collide.
    When unset, falls back to the bare basename.

    ``debug``: when True, ``_finalize()`` writes full prompt/response
    artifacts (.md, .prompt.json, .response.txt) into *run_dir*.
    Failed responses are always written when *run_dir* is set.
    """

    model: str | None = None
    run_dir: str | None = None
    repo_root: str | None = None
    debug: bool = False


@runtime_checkable