Show stats in summary line (files / rules / elapsed)

[ggueret]

Motivation

After rude check succeeds, the only output is:

All checks passed

It is correct but unsatisfying — no signal that real work happened. On a fresh integration, a maintainer cannot tell at a glance whether rude:

• looked at the expected number of files,
• ran the expected number of rules,
• completed in a reasonable time.

Silence-on-success conflates "everything is fine" with "nothing was checked" (e.g. paths typo, select accidentally narrowed to zero rules via a broad ignore, glob mismatch). A one-line stat dispels that doubt cheaply.

Proposal

Append (files, rules, elapsed) to both the success and failure summary so the two paths are symmetric:

Success:

All checks passed (62 files, 87 rules, 0.42s)

Failure: (currently Found 3 error(s) and 2 warning(s))

Found 3 error(s) and 2 warning(s) (62 files, 87 rules, 0.42s)

With --fix:

Fixed 4 issue(s)
All checks passed (62 files, 87 rules, 0.42s)

Details

• files: count of .py files actually analyzed (post-glob, post-skip).
• rules: count of rule instances loaded after select/ignore/plugins/local-rules filtering — answers "what did rude really enforce?", not "what is built-in".
• elapsed: wall time from after argument parsing to before summary print. Format Xms under 1 s, X.XXs above, to keep the visual scale obvious (850ms vs 12.50s rather than 0.85s vs 12.50s).

Optional, follow-on

• A Skipped: N files (parse error) line when non-empty — silent skips are easy to miss.
• A Auto-fixable: N issue(s) (rerun with --fix) line in the failure summary when any diagnostic has a [fix]. The per-diag [fix] marker is already there; a final count makes the call to action visible.
• Respect --quiet: suppress the stats line in --quiet mode (machine-consumer territory).

Reference

Tried out while integrating rude into a downstream project — having the stats inline would have removed a "did this actually run?" verification step.