ERRANT v2.3.3

CHANGELOG.md

@@ -2,6 +2,10 @@
 
 This log describes all the changes made to ERRANT since its release.
 
+## v2.3.3 (14-04-22)
+
+1. Missed one case of changing Levenshtein to rapidfuzz... Now fixed.
+
 ## v2.3.2 (14-04-22)
 
 1. Add more details to verbose ERRANT scoring. [#29](https://github.com/chrisjbryant/errant/pull/29)

README.md

@@ -1,4 +1,4 @@
-# ERRANT v2.3.2
+# ERRANT v2.3.3
 
 This repository contains the grammatical ERRor ANnotation Toolkit (ERRANT) described in:
 

errant/__init__.py

@@ -3,7 +3,7 @@
 from errant.annotator import Annotator
 
 # ERRANT version
-__version__ = '2.3.2'
+__version__ = '2.3.3'
 
 # Load an ERRANT Annotator object for a given language
 def load(lang, nlp=None):

errant/en/merger.py

@@ -1,7 +1,7 @@
 from itertools import combinations, groupby
 from re import sub
 from string import punctuation
-import Levenshtein
+from rapidfuzz.distance import Indel
 import spacy.symbols as POS
 from errant.edit import Edit
 
@@ -117,7 +117,7 @@ def is_punct(token):
 
 # Calculate the cost of character alignment; i.e. char similarity
 def char_cost(a, b):
-    return Levenshtein.ratio(a.text, b.text)
+    return 1-Indel.normalized_distance(a.text, b.text)
 
 # Merge the input alignment sequence to a single edit span
 def merge_edits(seq):

setup.py

@@ -10,7 +10,7 @@
 
 setup(
     name = "errant",
-    version = "2.3.2",
+    version = "2.3.3",
     license = "MIT",
     description = "The ERRor ANnotation Toolkit (ERRANT). Automatically extract and classify edits in parallel sentences.",
     long_description = readme,