score_cutoff argument not seeming to work for ratio about levenshtein HOT 6 CLOSED

What's the version of Levenshtein you are using?

import Levenshtein
print(Levenshtein.__version__)

from levenshtein.

It looks like Conda has an older version, 0.12.2. And at least with that version,, the version property also seems to be missing.

print(Levenshtein.version)
Traceback (most recent call last):

File "C:\Users\mikem\AppData\Local\Temp\ipykernel_16292\3385038131.py", line 1, in
print(Levenshtein.version)

AttributeError: module 'Levenshtein' has no attribute 'version'

from levenshtein.

The current version is available on conda: https://anaconda.org/conda-forge/levenshtein, so you should be able to upgrade

from levenshtein.

from levenshtein.

When processing large amounts of data you should try to use rapidfuzz directly:

from rapidfuzz.distance import Indel
from rapidfuzz import process

process.cdist(queries, choices, scorer=Indel.normalized_similarity, score_cutoff=score_cutoff, workers=-1)

It saves some overhead by reducing calls into Python, uses SIMD to compare multiple sequences in parallel when possible and allows the use of multiple threads (workers argument).

from levenshtein.

from levenshtein.