partial ratio for _token_set is ignored about fuzzywuzzy HOT 6 CLOSED

@ojomio pr please! :)

from fuzzywuzzy.

Closed by #60. Thanks for the pull request!

from fuzzywuzzy.

Today it suddenly hit me, that in current implementation every two strings that have a word in common will produce 100% match, because two of three comparisons here https://github.com/seatgeek/fuzzywuzzy/blob/master/fuzzywuzzy/fuzz.py#L186
are comparisons of intersection with 'intersection + stuff', which, if partial, is always 100. Is that normal or should we modify the behavior, I personally would leave only combined_1to2 to combined_2to1 check

from fuzzywuzzy.

OK, than we have combined_1to2 and combined_2to1, which have intersection at their beginning
Is this possible at all, that partial_ratio will find offset other than zero where these two strings will be more similar?

I mean, now I don't know if there is sense having partial option at all...

from fuzzywuzzy.

@acslater00 thoughts?

from fuzzywuzzy.

@josegonzalez You should have yelled at me in rage :) No matter which offset be found, but at lest at offset zero ratio() will be calculated with the substring of the longer string of length of the shorter string, and not the whole longer string. Thus, it seems there is sense in having such option

from fuzzywuzzy.