From 6e64a5cd0d49437592bc15c3a7272a1494d18af6 Mon Sep 17 00:00:00 2001
From: Kevin Humphreys <kevin.humphreys@dialpad.com>
Date: Mon, 31 Oct 2022 15:52:49 -0700
Subject: [PATCH] move fuzzy_match

to its own Python module to allow patching
---
 spacy/matcher/fuzzy.py    | 19 +++++++++++++++++++
 spacy/matcher/matcher.pyx | 20 ++++----------------
 2 files changed, 23 insertions(+), 16 deletions(-)
 create mode 100644 spacy/matcher/fuzzy.py

diff --git a/spacy/matcher/fuzzy.py b/spacy/matcher/fuzzy.py
new file mode 100644
index 000000000..3b29303be
--- /dev/null
+++ b/spacy/matcher/fuzzy.py
@@ -0,0 +1,19 @@
+from .levenshtein import levenshtein
+
+def fuzzy_match(input_string: str, rule_string: str, distance: int=0) -> bool:
+    """Define in pure Python outside Matcher to allow patching.
+
+    Patch with e.g.:
+        import wrapt
+        from spacy.matcher import fuzzy
+        @wrapt.patch_function_wrapper('spacy.matcher.fuzzy', 'fuzzy_match')
+    *before* import spacy
+    """
+    min_length = min(len(input_string), len(rule_string))
+    if distance: # FUZZYn operators with explicit distance
+        threshold = min(distance, min_length - 1)
+    else: # FUZZY operator with default distance
+        threshold = min(5, min_length - 2)
+    if threshold > 0:
+        return levenshtein(input_string, rule_string) <= threshold
+    return False
diff --git a/spacy/matcher/matcher.pyx b/spacy/matcher/matcher.pyx
index 443947cf1..257fd3ea4 100644
--- a/spacy/matcher/matcher.pyx
+++ b/spacy/matcher/matcher.pyx
@@ -25,8 +25,7 @@ from ..errors import Errors, MatchPatternError, Warnings
 from ..strings import get_string_id
 from ..attrs import IDS
 
-from .levenshtein import levenshtein
-
+import fuzzy
 
 DEF PADDING = 5
 
@@ -205,17 +204,6 @@ cdef class Matcher:
                 else:
                     yield doc
 
-    @staticmethod
-    def fuzzy_match(s1: str, s2: str, distance: int) -> bool:
-        min_length = min(len(s1), len(s2))
-        if distance: # FUZZYn operators with explicit distance
-            threshold = min(distance, min_length - 1)
-        else: # FUZZY operator with default distance
-            threshold = min(5, min_length - 2)
-        if threshold > 0:
-            return levenshtein(s1, s2) <= threshold
-        return False
-
     def __call__(self, object doclike, *, as_spans=False, allow_missing=False, with_alignments=False):
         """Find all token sequences matching the supplied pattern.
 
@@ -862,7 +850,7 @@ class _FuzzyPredicate:
             value = token.vocab.strings[get_token_attr_for_matcher(token.c, self.attr)]
         if self.value == value:
             return True
-        return Matcher.fuzzy_match(value, self.value, self.fuzzy)
+        return fuzzy.fuzzy_match(value, self.value, self.fuzzy)
 
 
 class _RegexPredicate:
@@ -945,7 +933,7 @@ class _SetPredicate:
                 return True
             elif self.fuzzy is not None:
                 value = self.vocab.strings[value]
-                return any(Matcher.fuzzy_match(value, self.vocab.strings[v], self.fuzzy)
+                return any(fuzzy.fuzzy_match(value, self.vocab.strings[v], self.fuzzy)
                            for v in self.value)
             else:
                 return False
@@ -957,7 +945,7 @@ class _SetPredicate:
                 return False
             elif self.fuzzy is not None:
                 value = self.vocab.strings[value]
-                return not any(Matcher.fuzzy_match(value, self.vocab.strings[v], self.fuzzy)
+                return not any(fuzzy.fuzzy_match(value, self.vocab.strings[v], self.fuzzy)
                                for v in self.value)
             else:
                 return True