Analyzes Python source files for docstring quality issues. Runs all
enabled checks except freshness by default (freshness requires git
context; griffe also excluded when not installed). When *path* is a
directory, only files within that directory tree are checked (not the
entire project). Returns a JSON object with findings, summary
statistics, and optional presence coverage data. Invalid
configuration triggers a structured error response instead of
crashing the server. Input, read, decode, and parse failures report
an incomplete run with unique attempted, analyzed, and failed file counts.
Presence coverage explicitly marks incomplete results, with a null
percentage when no input could be analyzed. Legitimate empty selections
succeed without requiring optional check dependencies or git context.

Args:
    path: Path to a Python file or directory to check.
    checks: Optional list of check names to run. Valid names are
        ``presence``, ``enrichment``, ``freshness``, ``coverage``,
        ``griffe``. Defaults to all except freshness.

Returns:
    JSON string with ``findings``, ``summary``, and optionally
    ``presence_coverage`` keys. Returns an ``error`` key on
    invalid path, unknown check name, or malformed configuration.
    Call docvet_rules() for per-rule fix guidance and format examples.

Returns the complete rule catalog with rule name, associated check
module, human-readable description, category (required or
recommended), prescriptive fix guidance, and a format example.

Returns:
    JSON string with a ``rules`` array containing objects with
    ``name``, ``check``, ``description``, ``category``,
    ``guidance``, and ``fix_example`` keys.