Add cli flag to scan command to show detector-level scores

[omkar-foss]

Currently with #12 (PR #65), we show score and confidence for each finding in the output. By default, this is useful as it gives very granular scoring.

But in case the user wants to see per-detector aggregate scores for each commit, the user would've to manually go through the list and aggregate it at detector-level. This is especially true for folks who'd be interested in a particular type of detector (e.g. commit trailers).

So it'd be good if allow the user to pass a flag to see per-detector scores for each commit in output, something like:

disclosure scan --format=json --show-per-detector-scoring

Output could look like:

...

commit abc (score=125.0, confidence=high)
  per-detector scores:
  detector          score
  trailer           85.0
  committer         40.0

  findings:
  ...(what we have in current output)

commit def (score=85.0, confidence=high)
  per-detector scores:
  detector          score
  trailer           85.0

  findings:
  ...(what we have in current output)

...other commits similarly

[MoralCode]

long as

1. we actually output JSON when the user requests JSON (your example doesnt)
2. the scores we expose are always the scores from the end of the scoring process (i dont want to expose anything "internal" that could become depended on somewhere thats hard to change

[omkar-foss]

we actually output JSON when the user requests JSON (your example doesnt)

Yeah the structure is same for text and json, so what you see in example above for text will come in json too

the scores we expose are always the scores from the end of the scoring process (i dont want to expose anything "internal" that could become depended on somewhere thats hard to change

It isn't internal, but now that you mention I share your concern that it shouldn't be hard to change. May be we could look at this in later releases when we've more clarity