Consider heuristics for imperative detection about opinionated-commit-message HOT 8 CLOSED

Hi @pschanely ,
Actually, many of the exceptions on your list are already in the main whitelist: https://github.com/mristin/opinionated-commit-message/blob/master/src/mostFrequentEnglishVerbs.ts

(I spotted at least "disable", "redo", "standardize" and "realize".)

Are you sure you had to add these verbs to the list? I'm actually suspecting a bug on the opinionated-commit-message side.

So far, I observed very little additional verbs in the other projects, but that can be due to the fact that most or all contributors were non-native speakers and thus had a limited vocabulary.

from opinionated-commit-message.

Let me at least please first add the verbs from your list so that other users can benefit in the meanwhile.

from opinionated-commit-message.

@pschanely I've just released 2.3.2 with your exceptions included: https://github.com/mristin/opinionated-commit-message/releases/tag/v2.3.2.

Could you please try to test it on your side (or let me know if I should have a look)? Perhaps you kept adding verbs to the whitelist though they were already whitelisted?

from opinionated-commit-message.

Yeah, I think I just got into the habit of adding everything because I kept hitting new ones. It could also be that I'm just extra creative with phrasing. I'll update and see how it goes - thanks.

from opinionated-commit-message.

Quick follow up: sadly, I seem to keep hitting new ones. (example - though at least one of these is a mis-spelling!)

Looks like pydocstyle uses a real english stemmer, which might be another idea to consider someday.

from opinionated-commit-message.

Hi @pschanely ,
Thanks for the suggestion! If I understood PyCQA/pydocstyle#235 correctly, they also just used a whitelist, didn't they? I have some experience in NLP (though limited, and things definitely changed since I worked in it) -- context-free stemmers for English are usually indeed whitelists as the language is too irregular for anything else.

Are these the only false negatives you encountered or are there more? If there were more, would you mind creating an issue and listing them so that I could add them to the whitelist? Otherwise, I'll just insert the ones from https://github.com/pschanely/CrossHair/actions/runs/1168073763 and https://github.com/pschanely/CrossHair/blob/master/additional-verbs-in-imperative-mood.txt.

from opinionated-commit-message.

(I'm reopening the issue to give it more visibility.)

from opinionated-commit-message.

Closing the issue for no activity in a while.

from opinionated-commit-message.