mini-levenshtein 0.1.1 → 0.1.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/Gemfile.lock +1 -1
- data/README.md +2 -2
- data/ext/mini_levenshtein/extconf.rb +1 -0
- data/ext/mini_levenshtein/mini_levenshtein.c +91 -14
- data/lib/mini-levenshtein/version.rb +1 -1
- data/lib/mini-levenshtein.rb +0 -54
- metadata +2 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 87d8dfd831b540e20d080df316abaf2195f371822d68e819e5c717fd2cbb637d
|
4
|
+
data.tar.gz: acd26d48b08feeaecea2ff7b9e4afe7f751e6590dafd54c87971fec244d79e28
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 4a380bcd7f95a4bc0ae68a871a40c0ec18beab58ef5af6bcc3313de38bd836f1e0e40538e491c9c0be46ba7882f25607b747cce8b579654d58cbac55303e67f1
|
7
|
+
data.tar.gz: f46268d96b7dfb651bb42db3d2f509bb5e2bdb908375832ee0a12a21ce388461cf52434156e042c8ba201c24041451dc03dd74d17b43cf3d2efec913089472e1
|
data/Gemfile.lock
CHANGED
data/README.md
CHANGED
@@ -5,9 +5,9 @@ Simple, fast, levenshtein distance
|
|
5
5
|
# Synopsis
|
6
6
|
|
7
7
|
```ruby
|
8
|
-
MiniLevenshtein.
|
8
|
+
MiniLevenshtein.edit_distance("levenshtein", "levenstien") # => 3
|
9
9
|
|
10
|
-
MiniLevenshtein.
|
10
|
+
MiniLevenshtein.similarity("levenshtein", "levenstien") # 0.857142...
|
11
11
|
```
|
12
12
|
|
13
13
|
# Credits
|
@@ -1,23 +1,20 @@
|
|
1
|
-
#include <
|
2
|
-
#include
|
3
|
-
#include <wchar.h>
|
4
|
-
#include <stdlib.h>
|
1
|
+
#include <ruby.h>
|
2
|
+
#include "extconf.h"
|
5
3
|
|
6
|
-
#include "ruby.h"
|
7
4
|
#include "levenshtein.h"
|
8
5
|
|
9
|
-
VALUE
|
6
|
+
VALUE MiniLevenshtein = Qnil;
|
10
7
|
|
11
|
-
VALUE
|
8
|
+
VALUE rb_lev_edit_distance(VALUE self, VALUE s1, VALUE s2, VALUE xcost);
|
9
|
+
VALUE rb_edit_distance(VALUE self, VALUE s1, VALUE s2);
|
10
|
+
VALUE rb_similarity(VALUE self, VALUE s1, VALUE s2);
|
12
11
|
|
13
|
-
|
14
|
-
{
|
15
|
-
MiniLevenshteinInternal = rb_define_module("MiniLevenshteinInternal");
|
16
|
-
rb_define_method(MiniLevenshteinInternal, "internal_distance", method_internal_distance, 3);
|
17
|
-
}
|
18
|
-
|
19
|
-
VALUE method_internal_distance(VALUE self, VALUE s1, VALUE s2, VALUE xcost)
|
12
|
+
VALUE rb_lev_edit_distance(VALUE self, VALUE s1, VALUE s2, VALUE xcost)
|
20
13
|
{
|
14
|
+
Check_Type(s1, T_STRING);
|
15
|
+
Check_Type(s2, T_STRING);
|
16
|
+
Check_Type(xcost, T_FIXNUM);
|
17
|
+
|
21
18
|
size_t len1 = RSTRING_LEN(s1);
|
22
19
|
size_t len2 = RSTRING_LEN(s2);
|
23
20
|
|
@@ -30,3 +27,83 @@ VALUE method_internal_distance(VALUE self, VALUE s1, VALUE s2, VALUE xcost)
|
|
30
27
|
|
31
28
|
return INT2NUM(distance);
|
32
29
|
}
|
30
|
+
|
31
|
+
/*
|
32
|
+
* Compute absolute Levenshtein distance of two strings.
|
33
|
+
*
|
34
|
+
* @example it's hard to spell Levenshtein correctly
|
35
|
+
* edit_distance('Levenshtein', 'Lenvinsten') # => 4
|
36
|
+
* edit_distance('Levenshtein', 'Levensthein') # => 2
|
37
|
+
* edit_distance('Levenshtein', 'Levenshten') # => 1
|
38
|
+
* edit_distance('Levenshtein', 'Levenshtein') # => 0
|
39
|
+
* "Yeah, we've managed it at last."
|
40
|
+
*
|
41
|
+
* @param string1 [String]
|
42
|
+
* @param string2 [String]
|
43
|
+
*
|
44
|
+
* @return [Integer]
|
45
|
+
*/
|
46
|
+
VALUE rb_edit_distance(VALUE self, VALUE string1, VALUE string2)
|
47
|
+
{
|
48
|
+
Check_Type(string1, T_STRING);
|
49
|
+
Check_Type(string2, T_STRING);
|
50
|
+
|
51
|
+
size_t len1 = RSTRING_LEN(string1);
|
52
|
+
size_t len2 = RSTRING_LEN(string2);
|
53
|
+
|
54
|
+
const lev_byte *str1 = StringValuePtr(string1);
|
55
|
+
const lev_byte *str2 = StringValuePtr(string2);
|
56
|
+
|
57
|
+
long distance = lev_edit_distance(len1, str1, len2, str2, 0);
|
58
|
+
|
59
|
+
return INT2NUM(distance);
|
60
|
+
}
|
61
|
+
|
62
|
+
/*
|
63
|
+
* Compute similarity of two strings.
|
64
|
+
*
|
65
|
+
* The similarity is a number between 0 and 1.
|
66
|
+
*
|
67
|
+
* @example
|
68
|
+
* similarity('Hello World!', 'Holly grail!') # => 0.5833...
|
69
|
+
*
|
70
|
+
* @example
|
71
|
+
* similarity('Holmes', 'Jack') # => 0.0
|
72
|
+
*
|
73
|
+
* @param string1 [String]
|
74
|
+
* @param string2 [String]
|
75
|
+
*
|
76
|
+
* @return [Float]
|
77
|
+
*/
|
78
|
+
VALUE rb_similarity(VALUE self, VALUE string1, VALUE string2)
|
79
|
+
{
|
80
|
+
Check_Type(string1, T_STRING);
|
81
|
+
Check_Type(string2, T_STRING);
|
82
|
+
|
83
|
+
size_t len1 = RSTRING_LEN(string1);
|
84
|
+
size_t len2 = RSTRING_LEN(string2);
|
85
|
+
|
86
|
+
const lev_byte *str1 = StringValuePtr(string1);
|
87
|
+
const lev_byte *str2 = StringValuePtr(string2);
|
88
|
+
|
89
|
+
size_t lensum = len1 + len2;
|
90
|
+
if (lensum == 0) {
|
91
|
+
return DBL2NUM(1.0);
|
92
|
+
}
|
93
|
+
|
94
|
+
long distance = lev_edit_distance(len1, str1, len2, str2, 0);
|
95
|
+
if (distance == 0) {
|
96
|
+
return DBL2NUM(1.0);
|
97
|
+
}
|
98
|
+
|
99
|
+
return DBL2NUM((lensum - distance) / (double)lensum);
|
100
|
+
}
|
101
|
+
|
102
|
+
void Init_mini_levenshtein()
|
103
|
+
{
|
104
|
+
MiniLevenshtein = rb_define_module("MiniLevenshtein");
|
105
|
+
rb_define_singleton_method(MiniLevenshtein, "lev_edit_distance", rb_lev_edit_distance, 3);
|
106
|
+
rb_define_singleton_method(MiniLevenshtein, "edit_distance", rb_edit_distance, 2);
|
107
|
+
rb_define_singleton_method(MiniLevenshtein, "similarity", rb_similarity, 2);
|
108
|
+
}
|
109
|
+
|
data/lib/mini-levenshtein.rb
CHANGED
@@ -1,62 +1,8 @@
|
|
1
1
|
require 'mini-levenshtein/mini_levenshtein'
|
2
2
|
|
3
3
|
module MiniLevenshtein
|
4
|
-
extend MiniLevenshteinInternal
|
5
|
-
|
6
4
|
class << self
|
7
|
-
# Compute absolute Levenshtein distance of two strings.
|
8
|
-
#
|
9
|
-
# @example (it's hard to spell Levenshtein correctly):
|
10
|
-
# distance('Levenshtein', 'Lenvinsten') # => 4
|
11
|
-
# distance('Levenshtein', 'Levensthein') # => 2
|
12
|
-
# distance('Levenshtein', 'Levenshten') # => 1
|
13
|
-
# distance('Levenshtein', 'Levenshtein') # => 0
|
14
|
-
# "Yeah, we've managed it at last."
|
15
|
-
#
|
16
|
-
# @param string1 [String]
|
17
|
-
# @param string2 [String]
|
18
|
-
#
|
19
|
-
# @return [Integer]
|
20
|
-
def edit_distance(string1, string2)
|
21
|
-
validate_string!(string1)
|
22
|
-
validate_string!(string2)
|
23
|
-
|
24
|
-
internal_distance(string1, string2, 0)
|
25
|
-
end
|
26
5
|
alias distance edit_distance
|
27
|
-
|
28
|
-
# Compute similarity of two strings.
|
29
|
-
#
|
30
|
-
# The similarity is a number between 0 and 1.
|
31
|
-
#
|
32
|
-
# @example
|
33
|
-
# similarity('Hello World!', 'Holly grail!') # => 0.5833...
|
34
|
-
#
|
35
|
-
# @example
|
36
|
-
# similarity('Holmes', 'Jack') # => 0.0
|
37
|
-
#
|
38
|
-
# @param string1 [String]
|
39
|
-
# @param string2 [String]
|
40
|
-
#
|
41
|
-
# @return [Float]
|
42
|
-
def similarity(string1, string2)
|
43
|
-
validate_string!(string1)
|
44
|
-
validate_string!(string2)
|
45
|
-
|
46
|
-
lensum = string1.length + string2.length
|
47
|
-
return 1.0 if lensum.zero?
|
48
|
-
|
49
|
-
distance = internal_distance(string1, string2, 1)
|
50
|
-
return 1.0 if distance.zero?
|
51
|
-
|
52
|
-
(lensum - distance) / lensum.to_f
|
53
|
-
end
|
54
6
|
alias ratio similarity
|
55
|
-
|
56
|
-
private
|
57
|
-
|
58
|
-
def validate_string!(string)
|
59
|
-
raise TypeError, "no implicit conversion of #{string.class} to String" unless string.is_a?(String)
|
60
|
-
end
|
61
7
|
end
|
62
8
|
end
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: mini-levenshtein
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.1.
|
4
|
+
version: 0.1.2
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Delon Newman
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2022-05-
|
11
|
+
date: 2022-05-05 00:00:00.000000000 Z
|
12
12
|
dependencies: []
|
13
13
|
description: Simple, fast, levenshtein distance
|
14
14
|
email: contact@delonnewman.name
|