Skip to content

Improve raw-source inspector: file_tree_content should explain directory content shape #111

Description

@abegong

Context

file_tree_content currently reports directory content shape through the same generic classes / outliers structure used elsewhere. That makes the evidence hard to read: the user cannot quickly tell which directories contain parseable markdown, which directories have frontmatter, or what metadata/key-set conventions Katalyst found.

For katalyst inspect ., this inspector should contribute to the story of how the directory works, not just list summarizer artifacts.

Goal

Make file_tree_content answer, at a glance:

  • Which directories contain markdown content?
  • Which directories parse cleanly, partially, or not at all?
  • Where does frontmatter appear, and where is it absent?
  • What frontmatter key sets are common by directory?
  • Which directories look like likely content collections based on content shape?
  • Which directories have content that does not fit their surrounding pattern?

Possible shape

Render a directory-oriented content table or grouped summary with concrete paths. Prefer labels like all parse, some parse failures, frontmatter keys: title, status, and no frontmatter over raw feature tokens such as fmkey:title and parsed.

The output should distinguish structural observations from judgment. It can say what was observed and where; it should not recommend a schema or mutation.

Acceptance criteria

  • katalyst inspect . --inspector file_tree_content describes content shape by directory in plain language.
  • Parse/frontmatter status is visible without requiring the reader to decode feature tokens.
  • Common frontmatter key sets are summarized with representative directories or paths.
  • Exceptions are grounded in concrete directories/files and are phrased as observations.
  • Existing JSON output remains complete and parseable; any schema changes are intentional and covered by tests.
  • Snapshot tests cover directories with frontmatter, without frontmatter, and with at least one parse failure or non-markdown region.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions