Unexpected results from token_set_ratio()

[timdagit]

I've been playing with the library today and am a bit confused by the behaviour of token_set_ratio(). Regardless of the token manipulation I would only expect a result of 100 if both strings were identical, but I also get 100 from the example below:

from fuzzywuzzy import fuzz

result = fuzz.token_set_ratio("word1 word2 word3", "word1 word2")

I would have expected that from partial_token_set_ratio() but not here, unless I've missed something.

[maxbachmann]

It is 100 as well when all words of one of the two strings appear in the other string

[timdagit]

Ah, thanks, how does that differ from partial_token_set_ratio()?

[maxbachmann]

Yes partial_token_set_ratio is based on partial_ratio instead of ratio and is already 100 when one word is similar.

fuzz.token_set_ratio("word1 word2 word3", "word1 word4")
# 71
fuzz.partial_token_set_ratio("word1 word2 word3", "word1 word4")
# 100

[timdagit]

Thanks for clarifying