Closed v1nam closed 3 years ago
Don't forget to resolve the conflicts in bot/constants.py
:)
Hey @Akarys42, could you have a second look at this and see if your requested changes are addressed?
Has anyone tested this and can confirm that it not only works, but that the rates that appear make sense?
I have tested it, i will do it one more time rn, and about the rates, it depends on how you want them to make sense, but to me its better than using fuzzywuzzy which will show 90% on the letter k
, which is not the case here and we are using regex, so any name that doesnt match gurkan|urkan|gurk
will have 0, i added gurkan in the regex too because if i just use gurk and urkan, so if someone has gurkan in there names the rate will be lower
It seems like this would make the steps more incremental. As long as it makes sense on most gurkan names (and non-gurkan ones), I’d be fine with doing it like this.
Could you try it with 5 random names (from our member list?) and post the results here, so I get get an overview of how it would perform.
sure, i will do it in a second
@gustavwilliam so basically, it checks if you have gurkan|urkan|gurk
in your name, (in that order) and then the rate will be the span of of the match, (i.e if your match is gurkan span will be 6) divided by the length of your name
Hmmm okay. It looks slightly weird when UncomfyMelonGurkan only got 33.33% (also, round it to 33%).
@gustavwilliam that is because i am wearing a disguise. i am actually a melon.
Is this ready? We have got 3 approvals on this.
@gustavwilliam is it necessary to round it? IMO decimals seem better :thinking:
I mean, that precision just looks weird. I don’t like the decimals
Alright then, tbh I don't mind either way, and the decimals can be quite annoying sometimes
Removed use of fuzzywuzzy completely, now function uses regex to check the span of string match and rate is decided based upon that