eddie 0.4.2

Fast and well-tested implementations of edit distance/string similarity metrics: Levenshtein, Damerau-Levenshtein, Hamming, Jaro, and Jaro-Winkler.
Documentation 
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
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305

use std::cmp::max;

/// # Hamming distance.
///
/// See [the detailed description][1].
///
/// [1]: https://en.wikipedia.org/wiki/Hamming_distance
///
/// # Usage
///
/// ```rust
/// use eddie::slice::Hamming;
///
/// let hamming = Hamming::new();
/// let dist = hamming.distance(&[1, 2, 3, 4, 5], &[1, 3, 2, 4, 5]);
/// assert_eq!(dist, Some(2));
/// ```
///
/// Returns `None` if slices have different lengths:
/// ```rust
/// # let hamming = eddie::slice::Hamming::new();
/// let dist = hamming.distance(&[1, 2, 3, 4, 5], &[1, 3, 2, 4]);
/// assert_eq!(dist, None);
/// ```
///
/// # Complementary metrics
///
/// Relative distance:
/// ```rust
/// # use std::cmp::max;
/// # fn main() {
/// #     if test().is_none() {
/// #         panic!("Expected test to return Some(f), got None");
/// #     };
/// # }
/// # fn test() -> Option<f64> {
/// #     let hamming = eddie::slice::Hamming::new();
/// #     let s1 = &[1, 2, 3, 4, 5];
/// #     let s2 = &[1, 3, 2, 4, 5];
/// let rel = hamming.rel_dist(s1, s2)?;
/// let dist = hamming.distance(s1, s2).map(|d| d as f64)?;
/// let max_len = max(s1.len(), s2.len()) as f64;
/// assert_eq!(rel, dist / max_len);
/// #     Some(1.0)
/// # }
/// ```
///
/// Similarity:
/// ```rust
/// # use std::cmp::max;
/// # fn main() {
/// #     if test().is_none() {
/// #         panic!("Expected test to return Some(f), got None");
/// #     };
/// # }
/// # fn test() -> Option<f64> {
/// #     let hamming = eddie::slice::Hamming::new();
/// #     let s1 = &[1, 2, 3, 4, 5];
/// #     let s2 = &[1, 3, 2, 4, 5];
/// let rel = hamming.rel_dist(s1, s2)?;
/// let sim = hamming.similarity(s1, s2)?;
/// assert_eq!(sim, 1.0 - rel);
/// #     Some(1.0)
/// # }
/// ```
pub struct Hamming { }


impl Hamming {
    /// Creates a new instance of Hamming struct with an internal state
    /// for the metric methods to reuse.
    ///
    /// # Example
    ///
    /// ```rust
    /// use eddie::slice::Hamming;
    ///
    /// let hamming = Hamming::new();
    /// ```
    pub fn new() -> Self {
        Self { }
    }

    /// Distance metric. Returns a number of positions
    /// at wich slice items are different.
    ///
    /// Returns `None` if slices have different lengths.
    ///
    /// # Example
    ///
    /// ```rust
    /// # let hamming = eddie::slice::Hamming::new();
    /// let dist1 = hamming.distance(&[1, 2, 3, 4, 5], &[1, 3, 2, 4, 5]);
    /// assert_eq!(dist1, Some(2));
    ///
    /// let dist2 = hamming.distance(&[1, 2, 3, 4, 5], &[1, 3, 2, 4]);
    /// assert_eq!(dist2, None);
    /// ```
    pub fn distance<T: PartialEq + Copy>(&self, slice1: &[T], slice2: &[T]) -> Option<usize> {
        if slice1.len() != slice2.len() { return None; }
        let dist = slice1.into_iter()
            .zip(slice2.into_iter())
            .filter(|(x1, x2)| x1 != x2)
            .count();
        Some(dist)
    }

    /// Relative distance metric. Returns a distance relative to the slice length,
    /// ranging from 0.0 (equality) to 1.0 (nothing in common).
    ///
    /// Returns `None` if slices have different lengths.
    ///
    /// # Example
    ///
    /// ```rust
    /// # use std::cmp::max;
    /// # fn main() {
    /// #     if test().is_none() {
    /// #         panic!("Expected test to return Some(f), got None");
    /// #     };
    /// # }
    /// # fn test() -> Option<f64> {
    /// #     let hamming = eddie::slice::Hamming::new();
    /// #     let s1 = &[1, 2, 3, 4, 5];
    /// #     let s2 = &[1, 3, 2, 4, 5];
    /// let dist1 = hamming.rel_dist(&[1, 2, 3, 4, 5], &[1, 3, 2, 4, 5])?;
    /// assert!((dist1 - 0.4).abs() < 0.001);
    ///
    /// let dist2 = hamming.rel_dist(&[1, 2, 3, 4, 5], &[1, 3, 2, 4]);
    /// assert_eq!(dist2, None);
    /// #     Some(1.0)
    /// # }
    /// ```
    pub fn rel_dist<T: PartialEq + Copy>(&self, slice1: &[T], slice2: &[T]) -> Option<f64> {
        self.distance(slice1, slice2)
            .map(|dist| dist as f64 / max(slice1.len(), 1) as f64)
    }

    /// Similarity metric. Inversion of relative distance,
    /// ranging from 1.0 (equality) to 0.0 (nothing in common).
    ///
    /// Returns `None` if slices have different lengths.
    ///
    /// # Example
    ///
    /// ```rust
    /// # use std::cmp::max;
    /// # fn main() {
    /// #     if test().is_none() {
    /// #         panic!("Expected test to return Some(f), got None");
    /// #     };
    /// # }
    /// # fn test() -> Option<f64> {
    /// #     let hamming = eddie::slice::Hamming::new();
    /// #     let s1 = &[1, 2, 3, 4, 5];
    /// #     let s2 = &[1, 3, 2, 4, 5];
    /// let sim1 = hamming.similarity(&[1, 2, 3, 4, 5], &[1, 3, 2, 4, 5])?;
    /// assert!((sim1 - 0.6).abs() < 0.001);
    ///
    /// let sim2 = hamming.similarity(&[1, 2, 3, 4, 5], &[1, 3, 2, 4]);
    /// assert_eq!(sim2, None);
    /// #     Some(1.0)
    /// # }
    /// ```
    pub fn similarity<T: PartialEq + Copy>(&self, slice1: &[T], slice2: &[T]) -> Option<f64> {
        self.rel_dist(slice1, slice2)
            .map(|dist| 1.0 - dist)
    }
}


#[cfg(test)]
mod tests {
    use super::Hamming;

    fn floor3(num: f64) -> f64 {
        let p = 10usize.pow(3) as f64;
        (num * p).floor() / p
    }

    #[test]
    fn equality() {
        let hamming = Hamming::new();
        let sample = [
            vec![],
            vec![1],
            vec![1, 2],
            vec![1, 2, 3],
        ];
        for s in sample.iter() {
            assert_eq!(hamming.distance(s, s), Some(0));
        }
    }

    #[test]
    fn inequality() {
        let hamming = Hamming::new();
        for i in 1..10 {
            let s1: Vec<&usize> = [1].into_iter().cycle().take(i).collect();
            let s2: Vec<&usize> = [2].into_iter().cycle().take(i).collect();
            assert_eq!(hamming.distance(&s1, &s2), Some(i));
        }
    }

    #[test]
    fn length_difference() {
        let hamming = Hamming::new();
        for len1 in 1..10 {
            for len2 in 0 .. len1 - 1 {
                let a1: Vec<&usize> = [1].into_iter().cycle().take(len1).collect();
                let a2: Vec<&usize> = [1].into_iter().cycle().take(len2).collect();
                let b2: Vec<&usize> = [2].into_iter().cycle().take(len2).collect();

                assert_eq!(hamming.distance(&a1, &a2), None);
                assert_eq!(hamming.distance(&a2, &a1), None);

                assert_eq!(hamming.distance(&a1, &b2), None);
                assert_eq!(hamming.distance(&b2, &a1), None);
            }
        }
    }

    #[test]
    fn sub_continuous() {
        let hamming = Hamming::new();
        let sample = [
            (Some(1), vec![1, 2, 3, 4], vec![0, 2, 3, 4]),
            (Some(2), vec![1, 2, 3, 4], vec![0, 0, 3, 4]),
            (Some(3), vec![1, 2, 3, 4], vec![0, 0, 0, 4]),

            (Some(1), vec![1, 2, 3, 4], vec![1, 0, 3, 4]),
            (Some(2), vec![1, 2, 3, 4], vec![1, 0, 0, 4]),

            (Some(1), vec![1, 2, 3, 4], vec![1, 2, 3, 0]),
            (Some(2), vec![1, 2, 3, 4], vec![1, 2, 0, 0]),
            (Some(3), vec![1, 2, 3, 4], vec![1, 0, 0, 0]),
        ];
        for (d, s1, s2) in sample.iter() {
            assert_eq!(hamming.distance(s1, s2), *d);
        }
    }

    #[test]
    fn sub_intermittent() {
        let hamming = Hamming::new();
        let sample = [
            (1, vec![1, 2, 3, 4], vec![0, 2, 3, 4]),
            (2, vec![1, 2, 3, 4], vec![0, 2, 0, 4]),

            (1, vec![1, 2, 3, 4], vec![1, 2, 3, 0]),
            (2, vec![1, 2, 3, 4], vec![1, 0, 3, 0]),
        ];
        for (d, s1, s2) in sample.iter() {
            assert_eq!(hamming.distance(s1, s2), Some(*d));
        }
    }

    #[test]
    fn rel_dist() {
        let hamming = Hamming::new();
        let sample = [
            (Some(0.000), vec![],           vec![]),
            (Some(0.000), vec![1, 2, 3, 4], vec![1, 2, 3, 4]),
            (Some(1.000), vec![1, 2, 3, 4], vec![3, 4, 1, 2]),
            (Some(0.500), vec![1, 2, 3, 4], vec![0, 0, 3, 4]),
            (None,        vec![1, 2, 3, 4], vec![1, 2]),
            (None,        vec![1, 2, 3, 4], vec![]),
        ];
        for (d, s1, s2) in sample.iter() {
            assert_eq!(hamming.rel_dist(s1, s2).map(floor3), *d);
            assert_eq!(hamming.rel_dist(s2, s1).map(floor3), *d);
        }
    }

    #[test]
    fn similarity() {
        let hamming = Hamming::new();
        let sample = [
            (Some(1.000), vec![],           vec![]),
            (Some(1.000), vec![1, 2, 3, 4], vec![1, 2, 3, 4]),
            (Some(0.000), vec![1, 2, 3, 4], vec![3, 4, 1, 2]),
            (Some(0.500), vec![1, 2, 3, 4], vec![0, 0, 3, 4]),
            (None,        vec![1, 2, 3, 4], vec![1, 2]),
            (None,        vec![1, 2, 3, 4], vec![]),
        ];
        for (d, s1, s2) in sample.iter() {
            assert_eq!(hamming.similarity(s1, s2).map(floor3), *d);
            assert_eq!(hamming.similarity(s2, s1).map(floor3), *d);
        }
    }

    #[test]
    fn growth() {
        let hamming = Hamming::new();
        for len in (1..1001).step_by(100) {
            let mut v1 = Vec::with_capacity(len);
            let mut v2 = Vec::with_capacity(len);
            v1.resize(len, 1);
            v2.resize(len, 2);
            assert_eq!(hamming.distance(&v1, &[]), None);
            assert_eq!(hamming.distance(&v1, &v1), Some(0));
            assert_eq!(hamming.distance(&v1, &v2), Some(len));
        }
    }
}