Class PreparedFile

explainshell/extraction/llm/extractor.py:130–161

Result of LLMExtractor.prepare() — everything needed for extraction. Fields ------ synopsis: Synopsis line from lexgrog, or None. aliases: (name, score) tuples for alternative command names. original_lines: 1-indexed line number → original line content (before

@dataclass
class PreparedFile:
    """Result of LLMExtractor.prepare() — everything needed for extraction.

    Fields
    ------
    synopsis:       Synopsis line from lexgrog, or None.
    aliases:        (name, score) tuples for alternative command names.
    original_lines: 1-indexed line number → original line content (before
                    numbering). Used by finalize to map LLM line references
                    back to source text.
    basename:       Manpage file stem without .gz/.section suffixes (e.g. "tar").
    numbered_text:  Full manpage text with "  42| …" line-number prefixes,
                    used for debug dumps.
    plain_text_len: Length of the original plain text before filtering/chunking.
    plain_text:     Original unfiltered manpage text (used for RawManpage storage).
    requests:       Pre-formatted user-content strings, one per chunk, ready to
                    submit to the LLM provider.
    n_chunks:       Derived property — ``len(requests)``.
    """

    synopsis: str | None
    aliases: list[tuple[str, int]]
    original_lines: dict[int, str]
    basename: str
    numbered_text: str
    plain_text_len: int
    plain_text: str
    requests: list[str]

    @property
    def n_chunks(self) -> int:
        return len(self.requests)


@runtime_checkable