Merge pull request #23 from matthijsz/master

Added SIFT4

README.md

@@ -393,8 +393,6 @@ Distance is computed as 1 - similarity.
 ### SIFT4
 SIFT4 is a general purpose string distance algorithm inspired by JaroWinkler and Longest Common Subsequence. It was developed to produce a distance measure that matches as close as possible to the human perception of string distance. Hence it takes into account elements like character substitution, character distance, longest common subsequence etc. It was developed using experimental testing, and without theoretical background.
 
-**Not implemented yet**
-
 
 
 ## Users

setup.py

@@ -5,7 +5,7 @@ with open("README.md", "r") as fh:
 
 setuptools.setup(
     name="strsimpy",
-    version="0.1.8",
+    version="0.1.9",
     description="A library implementing different string similarity and distance measures",
     long_description=long_description,
     long_description_content_type="text/markdown",

strsimpy/__init__.py

@@ -34,6 +34,7 @@ from .sorensen_dice import SorensenDice
 from .string_distance import StringDistance
 from .string_similarity import StringSimilarity
 from .weighted_levenshtein import WeightedLevenshtein
+from .sift4 import SIFT4
 
 __name__ = 'strsimpy'
-__version__ = '0.1.8'
+__version__ = '0.1.9'

strsimpy/sift4.py

@@ -0,0 +1,188 @@
+# Permission is hereby granted, free of charge, to any person obtaining a copy
+# of this software and associated documentation files (the "Software"), to deal
+# in the Software without restriction, including without limitation the rights
+# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+# copies of the Software, and to permit persons to whom the Software is
+# furnished to do so, subject to the following conditions:
+#
+# The above copyright notice and this permission notice shall be included in all
+# copies or substantial portions of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+# SOFTWARE.
+from .string_distance import MetricStringDistance
+
+
+class SIFT4Options(MetricStringDistance):
+    def __init__(self, options=None):
+        self.options = {
+            'maxdistance': 0,
+            'tokenizer': lambda x: [i for i in x],
+            'tokenmatcher': lambda t1, t2: t1 == t2,
+            'matchingevaluator': lambda t1, t2: 1,
+            'locallengthevaluator': lambda x: x,
+            'transpositioncostevaluator': lambda c1, c2: 1,
+            'transpositionsevaluator': lambda lcss, trans: lcss - trans
+        }
+        otheroptions = {
+            'tokenizer': {
+                'ngram': self.ngramtokenizer,
+                'wordsplit': self.wordsplittokenizer,
+                'characterfrequency': self.characterfrequencytokenizer
+            },
+            'tokematcher': {'sift4tokenmatcher': self.sift4tokenmatcher},
+            'matchingevaluator': {'sift4matchingevaluator': self.sift4matchingevaluator},
+            'locallengthevaluator': {
+                'rewardlengthevaluator': self.rewardlengthevaluator,
+                'rewardlengthevaluator2': self.rewardlengthevaluator2
+            },
+            'transpositioncostevaluator': {'longertranspositionsaremorecostly':self.longertranspositionsaremorecostly},
+            'transpositionsevaluator': {}
+        }
+        if isinstance(options, dict):
+            for k, v in options.items():
+                if k in self.options.keys():
+                    if k == 'maxdistance':
+                        if isinstance(v, int):
+                            self.options[k] = v
+                        else:
+                            raise ValueError("Option maxdistance should be int")
+                    else:
+                        if callable(v):
+                            self.options[k] = v
+                        else:
+                            if v in otheroptions[k].keys():
+                                self.options[k] = otheroptions[k][v]
+                            else:
+                                msg = "Option {} should be callable or one of [{}]".format(k, ', '.join(otheroptions[k].keys()))
+                                raise ValueError(msg)
+                else:
+                    raise ValueError("Option {} not recognized.".format(k))
+        elif options is not None:
+            raise ValueError("options should be a dictionary")
+        self.maxdistance = self.options['maxdistance']
+        self.tokenizer = self.options['tokenizer']
+        self.tokenmatcher = self.options['tokenmatcher']
+        self.matchingevaluator = self.options['matchingevaluator']
+        self.locallengthevaluator = self.options['locallengthevaluator']
+        self.transpositioncostevaluator = self.options['transpositioncostevaluator']
+        self.transpositionsevaluator = self.options['transpositionsevaluator']
+
+    # tokenizers:
+    @staticmethod
+    def ngramtokenizer(s, n):
+        result = []
+        if not s:
+            return result
+        for i in range(len(s) - n - 1):
+            result.append(s[i:(i + n)])
+        return result
+
+    @staticmethod
+    def wordsplittokenizer(s):
+        if not s:
+            return []
+        return s.split()
+
+    @staticmethod
+    def characterfrequencytokenizer(s):
+        letters = [i for i in 'abcdefghijklmnopqrstuvwxyz']
+        return [s.lower().count(x) for x in letters]
+
+    # tokenMatchers:
+    @staticmethod
+    def sift4tokenmatcher(t1, t2):
+        similarity = 1 - SIFT4().distance(t1, t2, 5) / max(len(t1), len(t2))
+        return similarity > 0.7
+
+    # matchingEvaluators:
+    @staticmethod
+    def sift4matchingevaluator(t1, t2):
+        similarity = 1 - SIFT4().distance(t1, t2, 5) / max(len(t1), len(t2))
+        return similarity
+
+    # localLengthEvaluators:
+    @staticmethod
+    def rewardlengthevaluator(l):
+        if l < 1:
+            return l
+        return l - 1 / (l + 1)
+
+    @staticmethod
+    def rewardlengthevaluator2(l):
+        return pow(l, 1.5)
+
+    # transpositionCostEvaluators:
+    @staticmethod
+    def longertranspositionsaremorecostly(c1, c2):
+        return abs(c2 - c1) / 9 + 1
+
+
+class SIFT4:
+    # As described in https://siderite.dev/blog/super-fast-and-accurate-string-distance.html/
+    def distance(self, s1, s2, maxoffset=5, options=None):
+        options = SIFT4Options(options)
+        t1, t2 = options.tokenizer(s1), options.tokenizer(s2)
+        l1, l2 = len(t1), len(t2)
+        if l1 == 0:
+            return l2
+        if l2 == 0:
+            return l1
+
+        c1, c2, lcss, local_cs, trans, offset_arr = 0, 0, 0, 0, 0, []
+        while (c1 < l1) and (c2 < l2):
+            if options.tokenmatcher(t1[c1], t2[c2]):
+                local_cs += options.matchingevaluator(t1[c1], t2[c2])
+                isTrans = False
+                i = 0
+                while i < len(offset_arr):
+                    ofs = offset_arr[i]
+                    if (c1 <= ofs['c1']) or (c2 <= ofs['c2']):
+                        isTrans = abs(c2 - c1) >= abs(ofs['c2'] - ofs['c1'])
+                        if isTrans:
+                            trans += options.transpositioncostevaluator(c1, c2)
+                        else:
+                            if not ofs['trans']:
+                                ofs['trans'] = True
+                                trans += options.transpositioncostevaluator(ofs['c1'], ofs['c2'])
+                        break
+                    else:
+                        if (c1 > ofs['c2']) and (c2 > ofs['c1']):
+                            offset_arr.pop(i)
+                        else:
+                            i += 1
+                offset_arr.append({'c1': c1, 'c2': c2, 'trans': isTrans})
+            else:
+                lcss += options.locallengthevaluator(local_cs)
+                local_cs = 0
+                if c1 != c2:
+                    c1 = c2 = min(c1, c2)
+                for i in range(maxoffset):
+                    if (c1 + i < l1) or (c2 + i < l2):
+                        if (c1 + i < l1) and options.tokenmatcher(t1[c1 + i], t2[c2]):
+                            c1 += i - 1
+                            c2 -= 1
+                            break
+                    if (c2 + i < l2) and options.tokenmatcher(t1[c1], t2[c2 + i]):
+                        c1 -= 1
+                        c2 += i - 1
+                        break
+            c1 += 1
+            c2 += 1
+            if options.maxdistance:
+                temporarydistance = options.locallengthevaluator(max(c1, c2)) - options.transpositionsevaluator(lcss, trans)
+                if temporarydistance >= options.maxdistance:
+                    return round(temporarydistance)
+            if (c1 >= l1) or (c2 >= l2):
+                lcss += options.locallengthevaluator(local_cs)
+                local_cs = 0
+                c1 = c2 = min(c1, c2)
+        lcss += options.locallengthevaluator(local_cs)
+        return round(options.locallengthevaluator(max(l1, l2)) - options.transpositionsevaluator(lcss, trans))
+
+