1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162

use std::cmp;
use unicode_segmentation::UnicodeSegmentation;

enum JaroVersion {
    Pure,
    Winkler,
    WinklerLongTolerance,
}

fn vec_jaro_or_winkler<T: PartialEq>(s1: &Vec<T>, s2: &Vec<T>, version: JaroVersion) -> f64 {
    let s1_len = s1.len();
    let s2_len = s2.len();

    if s1_len == 0 || s2_len == 0 {
        return 0.0;
    }

    let min_len = cmp::max(s1_len, s2_len);
    let search_range = if min_len > 1 { (min_len / 2) - 1 } else { 0 };

    let mut s1_flags = vec![false; s1_len];
    let mut s2_flags = vec![false; s2_len];
    let mut common_chars = 0;

    // looking only within search range, count & flag matched pairs
    for (i, s1_ch) in s1.iter().enumerate() {
        // avoid underflow on i - search_range
        let low = if search_range >= i {
            0
        } else {
            i - search_range
        };
        let hi = cmp::min(i + search_range, s2_len - 1);
        for j in low..hi + 1 {
            if !s2_flags[j] && s2[j] == *s1_ch {
                s1_flags[i] = true;
                s2_flags[j] = true;
                common_chars += 1;
                break;
            }
        }
    }

    // no characters match
    if common_chars == 0 {
        return 0.0;
    }

    // count transpositions
    let mut k = 0;
    let mut trans_count = 0;
    for (i, s1_f) in s1_flags.iter().enumerate() {
        if *s1_f {
            let mut j = k;
            while j < s2_len {
                if s2_flags[j] {
                    k = j + 1;
                    break;
                }
                j += 1;
            }
            if s1[i] != s2[j] {
                trans_count += 1
            }
        }
    }
    // need to do floor division then cast to float
    let trans_count = (trans_count / 2) as f64;
    let common_charsf = common_chars as f64;
    let s1_lenf = s1_len as f64;
    let s2_lenf = s2_len as f64;

    // adjust for similarities in nonmatched characters
    let mut weight = (common_charsf / s1_lenf
        + common_charsf / s2_lenf
        + (common_charsf - trans_count) / common_charsf)
        / 3.0;

    // check which version to run
    let (winklerize, long_tolerance) = match version {
        JaroVersion::Pure => (false, false),
        JaroVersion::Winkler => (true, false),
        JaroVersion::WinklerLongTolerance => (true, true),
    };

    // winkler moddification: continue to boost similar strings
    if winklerize && weight > 0.7 && s1_len > 3 && s2_len > 3 {
        let mut i = 0;
        let j = cmp::min(min_len, 4);
        while i < j && s1[i] == s2[i] {
            // TODO: also had s1[i] in Python, necessary?
            i += 1;
        }
        if i > 0 {
            weight += (i as f64) * 0.1 * (1.0 - weight);
        }

        // optional adjustment for long strings
        // after agreeing beginning items, at least two or more must agree
        // and agreed items must be more than half of remaining items
        if long_tolerance && min_len > 4 && common_chars > i + 1 && 2 * common_chars >= min_len + i
        {
            weight += (1.0 - weight) * (common_charsf - i as f64 - 1.0)
                / (s1_lenf + s2_lenf - i as f64 * 2.0 + 2.0);
        }
    }

    return weight;
}

pub fn vec_jaro_similarity<T: PartialEq>(s1: &Vec<T>, s2: &Vec<T>) -> f64 {
    vec_jaro_or_winkler(s1, s2, JaroVersion::Pure)
}

pub fn vec_jaro_winkler_similarity<T: PartialEq>(s1: &Vec<T>, s2: &Vec<T>) -> f64 {
    vec_jaro_or_winkler(s1, s2, JaroVersion::Winkler)
}

pub fn vec_jaro_winkler_similarity_longtol<T: PartialEq>(s1: &Vec<T>, s2: &Vec<T>) -> f64 {
    vec_jaro_or_winkler(s1, s2, JaroVersion::WinklerLongTolerance)
}

pub fn jaro_similarity(s1: &str, s2: &str) -> f64 {
    let us1 = UnicodeSegmentation::graphemes(s1, true).collect::<Vec<&str>>();
    let us2 = UnicodeSegmentation::graphemes(s2, true).collect::<Vec<&str>>();
    vec_jaro_similarity(&us1, &us2)
}

pub fn jaro_winkler_similarity(s1: &str, s2: &str) -> f64 {
    let us1 = UnicodeSegmentation::graphemes(s1, true).collect::<Vec<&str>>();
    let us2 = UnicodeSegmentation::graphemes(s2, true).collect::<Vec<&str>>();
    vec_jaro_winkler_similarity(&us1, &us2)
}

pub fn jaro_winkler_similarity_longtol(s1: &str, s2: &str) -> f64 {
    let us1 = UnicodeSegmentation::graphemes(s1, true).collect::<Vec<&str>>();
    let us2 = UnicodeSegmentation::graphemes(s2, true).collect::<Vec<&str>>();
    vec_jaro_winkler_similarity_longtol(&us1, &us2)
}

#[cfg(test)]
mod test {
    use super::*;
    use crate::testutils::testutils;
    #[test]
    fn test_jaro() {
        testutils::test_similarity_func("testdata/jaro_distance.csv", jaro_similarity);
    }

    #[test]
    fn test_jaro_winkler() {
        testutils::test_similarity_func("testdata/jaro_winkler.csv", jaro_winkler_similarity);
    }

    #[test]
    fn test_jaro_winkler_longtol() {
        testutils::test_similarity_func(
            "testdata/jaro_winkler_longtol.csv",
            jaro_winkler_similarity_longtol,
        );
    }
}