Wrong (non symmetric) results for some strings if Levenshtein module is not present

[spotamianos]

If the python-Levenshtein library is not installed, then fuzz.ratio('abc', 'cadb') returns 57 and fuzz.ratio('cadb', 'abc') returns 29.
If the python-Levenshtein library is used, then both calls return 57.

Similar errors happen for other strings too, e.g:
'abcd', 'cbda'
'ONYZBOHON', 'ZRKFULFORD'

[spotamianos]

Another example is strings "9038119840000001" and "AS19970718106866036". We get 3 different answers:
levenshtein library: 34
no levensthein library: 17
no levenshtein library and order reversed: 23

[rnegron]

This is probably because without the Levenshtein library, this package uses a different algorithm from the standard library'sdifflib. See this issue from the previous name of this project: seatgeek/fuzzywuzzy#128

[saeed2402]

This is all very confusing, could someone please explain what "If the python-Levenshtein library is not installed" means? Does it mean if both "thefuzz" and "python-Levenshtein" are installed on the machine then we should get the same results for "fuzz.ratio('abc', 'cadb') " and "fuzz.ratio('cadb', 'abc') "?
I have both of them installed, and still get different results..

[maxbachmann]

Yes if both are installed you should get the same results:

>>> from thefuzz import fuzz
>>> fuzz.ratio('abc', 'cadb')
57
>>> fuzz.ratio('cadb', 'abc')
57

[saeed2402]

@maxbachmann I am using this library in Amazon Redshift in a Python UDF. Both libraries are installed on my Amazon Redshift cluster, running "select * from pg_library;" gives me "python_levenshtein, setuptools and thefuzz"

and still get 29 for "SELECT f_fuzzy_string_match( 'cadb', 'abc');"

[maxbachmann]

Then you do not have a valid installation of python-Levenshtein. You can check whether

from thefuzz import fuzz
import difflib
print(fuzz.SequenceMatcher == difflib.SequenceMatcher)

[saeed2402]

@maxbachmann Something is not right here mate. In "StringMatcher.py", line 11 is "from Levenshtein import *", which looks like the project you're maintaining on: https://github.com/maxbachmann/Levenshtein
Whereas yours and all other references are for https://pypi.org/project/python-Levenshtein/ , which apparently is not maintained anymore.
Can you please clarify which one is the correct one to have installed?
Also, your project "Levenshtein" errors out when used with Redshift Python UDF:

ERROR: File "/rdsdbdata/user_lib/1/0/105733.zip/Levenshtein/init.py", line 17 author: str = "Max Bachmann" ^ SyntaxError: invalid syntax. Please look at svl_udf_log for more information Detail: ----------------------------------------------- error: File "/rdsdbdata/user_lib/1/0/105733.zip/Levenshtein/init.py", line 17 author: str = "Max Bachmann" ^ SyntaxError: invalid syntax. Please look at svl_udf_log for more information code: 10000 context: UDF query: 0 location: udf_client.cpp:364 process: padbmaster [pid=3125] ----------------------------------------------- [ErrorId: 1-63081de5-55324c9c41e8b72f3fdf7abb]

[maxbachmann]

both should work