Finding best matches in a list gives wrong results.

[ardoi]

The process.extract function doesn't seem to handle capitalised queries well with some scorers:

>>>fuzzywuzzy.fuzz.partial_ratio('Santa Ana','Santa Ana')
100
>>>fuzzywuzzy.process.extract('Santa Ana', ['Santa Ana', 'Manta'], scorer=fuzzywuzzy.fuzz.partial_ratio)
[('Manta', 80), ('Santa Ana', 77)]

This is because here the choice string is processed but the query string is not:

>>>fuzzywuzzy.process.extract('santa ana', ['Santa Ana', 'Manta'], scorer=fuzzywuzzy.fuzz.partial_ratio)
[('Santa Ana', 100), ('Manta', 80)]

With some other scorers (e.g., WRatio) things work fine because those processes both strings internally anyway.

A workaround I use right now is:

>>>fuzzywuzzy.process.extract('Santa Ana', ['Santa Ana', 'Manta'], scorer=fuzzywuzzy.fuzz.partial_ratio, processor=lambda x:x)
[('Santa Ana', 100), ('Manta', 80)]

[josegonzalez]

@acslater00 Thoughts?

[acslater00]

Yeah this definitely looks broken.

Unfortunately I don't think it's safe to run processor() on the query because it isn't and hasn't been part of the spec. It makes sense when processing is just doing an encoding conversion and running .lower(), but it wouldn't make sense in a situation like this

choices = [
    "[2014-01-01] abcd",
    "[2014-01-01] cdef"
    "[2014-01-01] ghij"
]

def strip_date(s):
    return s[12:]

process.extract("abcde", choices, processor=strip_date, scorer=fuzz.ratio)

This example is contrived but not too unrealistic.

In retrospect there probably shouldn't be a default processor at all, since the default scorer processes its own input. However, changing that would not be BC.

The problem here, I think, is that there is a mismatch between the expected input to the custom scorer (in this case, ratio.partial_ratio) and the input to extract. Another reasonable solution is the following

In [28]: process.extract(
    utils.full_process("Santa Ana"), 
    ['Santa Ana', 'Manta'], 
   scorer=fuzz.partial_ratio)
Out[28]: [('Santa Ana', 100), ('Manta', 80)]