1//! This library implements string similarity metrics.
2
3#![forbid(unsafe_code)]
4
5use std::char;
6use std::cmp::{max, min};
7use std::collections::HashMap;
8use std::error::Error;
9use std::fmt::{self, Display, Formatter};
10use std::hash::Hash;
11use std::str::Chars;
12
13#[derive(Debug, PartialEq)]
14pub enum StrSimError {
15 DifferentLengthArgs,
16}
17
18impl Display for StrSimError {
19 fn fmt(&self, fmt: &mut Formatter) -> Result<(), fmt::Error> {
20 let text: &str = match self {
21 StrSimError::DifferentLengthArgs => "Differing length arguments provided",
22 };
23
24 write!(fmt, "{}", text)
25 }
26}
27
28impl Error for StrSimError {}
29
30pub type HammingResult = Result<usize, StrSimError>;
31
32/// Calculates the number of positions in the two sequences where the elements
33/// differ. Returns an error if the sequences have different lengths.
34pub fn generic_hamming<Iter1, Iter2, Elem1, Elem2>(a: Iter1, b: Iter2) -> HammingResult
35 where Iter1: IntoIterator<Item=Elem1>,
36 Iter2: IntoIterator<Item=Elem2>,
37 Elem1: PartialEq<Elem2> {
38 let (mut ita: Iter1, mut itb: Iter2) = (a.into_iter(), b.into_iter());
39 let mut count: usize = 0;
40 loop {
41 match (ita.next(), itb.next()){
42 (Some(x), Some(y)) => if x != y { count += 1 },
43 (None, None) => return Ok(count),
44 _ => return Err(StrSimError::DifferentLengthArgs),
45 }
46 }
47}
48
49/// Calculates the number of positions in the two strings where the characters
50/// differ. Returns an error if the strings have different lengths.
51///
52/// ```
53/// use strsim::{hamming, StrSimError::DifferentLengthArgs};
54///
55/// assert_eq!(Ok(3), hamming("hamming", "hammers"));
56///
57/// assert_eq!(Err(DifferentLengthArgs), hamming("hamming", "ham"));
58/// ```
59pub fn hamming(a: &str, b: &str) -> HammingResult {
60 generic_hamming(a:a.chars(), b:b.chars())
61}
62
63/// Calculates the Jaro similarity between two sequences. The returned value
64/// is between 0.0 and 1.0 (higher value means more similar).
65pub fn generic_jaro<'a, 'b, Iter1, Iter2, Elem1, Elem2>(a: &'a Iter1, b: &'b Iter2) -> f64
66 where &'a Iter1: IntoIterator<Item=Elem1>,
67 &'b Iter2: IntoIterator<Item=Elem2>,
68 Elem1: PartialEq<Elem2> {
69 let a_len = a.into_iter().count();
70 let b_len = b.into_iter().count();
71
72 // The check for lengths of one here is to prevent integer overflow when
73 // calculating the search range.
74 if a_len == 0 && b_len == 0 {
75 return 1.0;
76 } else if a_len == 0 || b_len == 0 {
77 return 0.0;
78 } else if a_len == 1 && b_len == 1 {
79 return if a.into_iter().eq(b.into_iter()) { 1.0} else { 0.0 };
80 }
81
82 let search_range = (max(a_len, b_len) / 2) - 1;
83
84 let mut b_consumed = Vec::with_capacity(b_len);
85 for _ in 0..b_len {
86 b_consumed.push(false);
87 }
88 let mut matches = 0.0;
89
90 let mut transpositions = 0.0;
91 let mut b_match_index = 0;
92
93 for (i, a_elem) in a.into_iter().enumerate() {
94 let min_bound =
95 // prevent integer wrapping
96 if i > search_range {
97 max(0, i - search_range)
98 } else {
99 0
100 };
101
102 let max_bound = min(b_len - 1, i + search_range);
103
104 if min_bound > max_bound {
105 continue;
106 }
107
108 for (j, b_elem) in b.into_iter().enumerate() {
109 if min_bound <= j && j <= max_bound && a_elem == b_elem &&
110 !b_consumed[j] {
111 b_consumed[j] = true;
112 matches += 1.0;
113
114 if j < b_match_index {
115 transpositions += 1.0;
116 }
117 b_match_index = j;
118
119 break;
120 }
121 }
122 }
123
124 if matches == 0.0 {
125 0.0
126 } else {
127 (1.0 / 3.0) * ((matches / a_len as f64) +
128 (matches / b_len as f64) +
129 ((matches - transpositions) / matches))
130 }
131}
132
133struct StringWrapper<'a>(&'a str);
134
135impl<'a, 'b> IntoIterator for &'a StringWrapper<'b> {
136 type Item = char;
137 type IntoIter = Chars<'b>;
138
139 fn into_iter(self) -> Self::IntoIter {
140 self.0.chars()
141 }
142}
143
144/// Calculates the Jaro similarity between two strings. The returned value
145/// is between 0.0 and 1.0 (higher value means more similar).
146///
147/// ```
148/// use strsim::jaro;
149///
150/// assert!((0.392 - jaro("Friedrich Nietzsche", "Jean-Paul Sartre")).abs() <
151/// 0.001);
152/// ```
153pub fn jaro(a: &str, b: &str) -> f64 {
154 generic_jaro(&StringWrapper(a), &StringWrapper(b))
155}
156
157/// Like Jaro but gives a boost to sequences that have a common prefix.
158pub fn generic_jaro_winkler<'a, 'b, Iter1, Iter2, Elem1, Elem2>(a: &'a Iter1, b: &'b Iter2) -> f64
159 where &'a Iter1: IntoIterator<Item=Elem1>,
160 &'b Iter2: IntoIterator<Item=Elem2>,
161 Elem1: PartialEq<Elem2> {
162 let jaro_distance: f64 = generic_jaro(a, b);
163
164 // Don't limit the length of the common prefix
165 let prefix_length: usize = aTakeWhile, …>.into_iter()
166 .zip(b.into_iter())
167 .take_while(|&(ref a_elem: &{unknown}, ref b_elem: &{unknown})| a_elem == b_elem)
168 .count();
169
170 let jaro_winkler_distance: f64 =
171 jaro_distance + (0.1 * prefix_length as f64 * (1.0 - jaro_distance));
172
173 if jaro_winkler_distance <= 1.0 {
174 jaro_winkler_distance
175 } else {
176 1.0
177 }
178}
179
180/// Like Jaro but gives a boost to strings that have a common prefix.
181///
182/// ```
183/// use strsim::jaro_winkler;
184///
185/// assert!((0.911 - jaro_winkler("cheeseburger", "cheese fries")).abs() <
186/// 0.001);
187/// ```
188pub fn jaro_winkler(a: &str, b: &str) -> f64 {
189 generic_jaro_winkler(&StringWrapper(a), &StringWrapper(b))
190}
191
192/// Calculates the minimum number of insertions, deletions, and substitutions
193/// required to change one sequence into the other.
194///
195/// ```
196/// use strsim::generic_levenshtein;
197///
198/// assert_eq!(3, generic_levenshtein(&[1,2,3], &[1,2,3,4,5,6]));
199/// ```
200pub fn generic_levenshtein<'a, 'b, Iter1, Iter2, Elem1, Elem2>(a: &'a Iter1, b: &'b Iter2) -> usize
201 where &'a Iter1: IntoIterator<Item=Elem1>,
202 &'b Iter2: IntoIterator<Item=Elem2>,
203 Elem1: PartialEq<Elem2> {
204 let b_len: usize = b.into_iter().count();
205
206 if a.into_iter().next().is_none() { return b_len; }
207
208 let mut cache: Vec<usize> = (1..b_len+1).collect();
209
210 let mut result: usize = 0;
211
212 for (i, a_elem) in a.into_iter().enumerate() {
213 result = i + 1;
214 let mut distance_b = i;
215
216 for (j, b_elem) in b.into_iter().enumerate() {
217 let cost: usize = if a_elem == b_elem { 0usize } else { 1usize };
218 let distance_a: usize = distance_b + cost;
219 distance_b = cache[j];
220 result = min(v1:result + 1, v2:min(v1:distance_a, v2:distance_b + 1));
221 cache[j] = result;
222 }
223 }
224
225 result
226}
227
228/// Calculates the minimum number of insertions, deletions, and substitutions
229/// required to change one string into the other.
230///
231/// ```
232/// use strsim::levenshtein;
233///
234/// assert_eq!(3, levenshtein("kitten", "sitting"));
235/// ```
236pub fn levenshtein(a: &str, b: &str) -> usize {
237 generic_levenshtein(&StringWrapper(a), &StringWrapper(b))
238}
239
240/// Calculates a normalized score of the Levenshtein algorithm between 0.0 and
241/// 1.0 (inclusive), where 1.0 means the strings are the same.
242///
243/// ```
244/// use strsim::normalized_levenshtein;
245///
246/// assert!((normalized_levenshtein("kitten", "sitting") - 0.57142).abs() < 0.00001);
247/// assert!((normalized_levenshtein("", "") - 1.0).abs() < 0.00001);
248/// assert!(normalized_levenshtein("", "second").abs() < 0.00001);
249/// assert!(normalized_levenshtein("first", "").abs() < 0.00001);
250/// assert!((normalized_levenshtein("string", "string") - 1.0).abs() < 0.00001);
251/// ```
252pub fn normalized_levenshtein(a: &str, b: &str) -> f64 {
253 if a.is_empty() && b.is_empty() {
254 return 1.0;
255 }
256 1.0 - (levenshtein(a, b) as f64) / (a.chars().count().max(b.chars().count()) as f64)
257}
258
259/// Like Levenshtein but allows for adjacent transpositions. Each substring can
260/// only be edited once.
261///
262/// ```
263/// use strsim::osa_distance;
264///
265/// assert_eq!(3, osa_distance("ab", "bca"));
266/// ```
267pub fn osa_distance(a: &str, b: &str) -> usize {
268 let a_len = a.chars().count();
269 let b_len = b.chars().count();
270 if a == b { return 0; }
271 else if a_len == 0 { return b_len; }
272 else if b_len == 0 { return a_len; }
273
274 let mut prev_two_distances: Vec<usize> = Vec::with_capacity(b_len + 1);
275 let mut prev_distances: Vec<usize> = Vec::with_capacity(b_len + 1);
276 let mut curr_distances: Vec<usize> = Vec::with_capacity(b_len + 1);
277
278 let mut prev_a_char = char::MAX;
279 let mut prev_b_char = char::MAX;
280
281 for i in 0..(b_len + 1) {
282 prev_two_distances.push(i);
283 prev_distances.push(i);
284 curr_distances.push(0);
285 }
286
287 for (i, a_char) in a.chars().enumerate() {
288 curr_distances[0] = i + 1;
289
290 for (j, b_char) in b.chars().enumerate() {
291 let cost = if a_char == b_char { 0 } else { 1 };
292 curr_distances[j + 1] = min(curr_distances[j] + 1,
293 min(prev_distances[j + 1] + 1,
294 prev_distances[j] + cost));
295 if i > 0 && j > 0 && a_char != b_char &&
296 a_char == prev_b_char && b_char == prev_a_char {
297 curr_distances[j + 1] = min(curr_distances[j + 1],
298 prev_two_distances[j - 1] + 1);
299 }
300
301 prev_b_char = b_char;
302 }
303
304 prev_two_distances.clone_from(&prev_distances);
305 prev_distances.clone_from(&curr_distances);
306 prev_a_char = a_char;
307 }
308
309 curr_distances[b_len]
310
311}
312
313/* Returns the final index for a value in a single vector that represents a fixed
314 2d grid */
315fn flat_index(i: usize, j: usize, width: usize) -> usize {
316 j * width + i
317}
318
319/// Like optimal string alignment, but substrings can be edited an unlimited
320/// number of times, and the triangle inequality holds.
321///
322/// ```
323/// use strsim::generic_damerau_levenshtein;
324///
325/// assert_eq!(2, generic_damerau_levenshtein(&[1,2], &[2,3,1]));
326/// ```
327pub fn generic_damerau_levenshtein<Elem>(a_elems: &[Elem], b_elems: &[Elem]) -> usize
328 where Elem: Eq + Hash + Clone {
329 let a_len = a_elems.len();
330 let b_len = b_elems.len();
331
332 if a_len == 0 { return b_len; }
333 if b_len == 0 { return a_len; }
334
335 let width = a_len + 2;
336 let mut distances = vec![0; (a_len + 2) * (b_len + 2)];
337 let max_distance = a_len + b_len;
338 distances[0] = max_distance;
339
340 for i in 0..(a_len + 1) {
341 distances[flat_index(i + 1, 0, width)] = max_distance;
342 distances[flat_index(i + 1, 1, width)] = i;
343 }
344
345 for j in 0..(b_len + 1) {
346 distances[flat_index(0, j + 1, width)] = max_distance;
347 distances[flat_index(1, j + 1, width)] = j;
348 }
349
350 let mut elems: HashMap<Elem, usize> = HashMap::with_capacity(64);
351
352 for i in 1..(a_len + 1) {
353 let mut db = 0;
354
355 for j in 1..(b_len + 1) {
356 let k = match elems.get(&b_elems[j - 1]) {
357 Some(&value) => value,
358 None => 0
359 };
360
361 let insertion_cost = distances[flat_index(i, j + 1, width)] + 1;
362 let deletion_cost = distances[flat_index(i + 1, j, width)] + 1;
363 let transposition_cost = distances[flat_index(k, db, width)] +
364 (i - k - 1) + 1 + (j - db - 1);
365
366 let mut substitution_cost = distances[flat_index(i, j, width)] + 1;
367 if a_elems[i - 1] == b_elems[j - 1] {
368 db = j;
369 substitution_cost -= 1;
370 }
371
372 distances[flat_index(i + 1, j + 1, width)] = min(substitution_cost,
373 min(insertion_cost, min(deletion_cost, transposition_cost)));
374 }
375
376 elems.insert(a_elems[i - 1].clone(), i);
377 }
378
379 distances[flat_index(a_len + 1, b_len + 1, width)]
380}
381
382/// Like optimal string alignment, but substrings can be edited an unlimited
383/// number of times, and the triangle inequality holds.
384///
385/// ```
386/// use strsim::damerau_levenshtein;
387///
388/// assert_eq!(2, damerau_levenshtein("ab", "bca"));
389/// ```
390pub fn damerau_levenshtein(a: &str, b: &str) -> usize {
391 let (x: Vec, y: Vec): (Vec<_>, Vec<_>) = (a.chars().collect(), b.chars().collect());
392 generic_damerau_levenshtein(a_elems:x.as_slice(), b_elems:y.as_slice())
393}
394
395/// Calculates a normalized score of the Damerau–Levenshtein algorithm between
396/// 0.0 and 1.0 (inclusive), where 1.0 means the strings are the same.
397///
398/// ```
399/// use strsim::normalized_damerau_levenshtein;
400///
401/// assert!((normalized_damerau_levenshtein("levenshtein", "löwenbräu") - 0.27272).abs() < 0.00001);
402/// assert!((normalized_damerau_levenshtein("", "") - 1.0).abs() < 0.00001);
403/// assert!(normalized_damerau_levenshtein("", "flower").abs() < 0.00001);
404/// assert!(normalized_damerau_levenshtein("tree", "").abs() < 0.00001);
405/// assert!((normalized_damerau_levenshtein("sunglasses", "sunglasses") - 1.0).abs() < 0.00001);
406/// ```
407pub fn normalized_damerau_levenshtein(a: &str, b: &str) -> f64 {
408 if a.is_empty() && b.is_empty() {
409 return 1.0;
410 }
411 1.0 - (damerau_levenshtein(a, b) as f64) / (a.chars().count().max(b.chars().count()) as f64)
412}
413
414/// Returns an Iterator of char tuples.
415fn bigrams(s: &str) -> impl Iterator<Item=(char, char)> + '_ {
416 s.chars().zip(s.chars().skip(1))
417}
418
419
420/// Calculates a Sørensen-Dice similarity distance using bigrams.
421/// See http://en.wikipedia.org/wiki/S%C3%B8rensen%E2%80%93Dice_coefficient.
422///
423/// ```
424/// use strsim::sorensen_dice;
425///
426/// assert_eq!(1.0, sorensen_dice("", ""));
427/// assert_eq!(0.0, sorensen_dice("", "a"));
428/// assert_eq!(0.0, sorensen_dice("french", "quebec"));
429/// assert_eq!(1.0, sorensen_dice("ferris", "ferris"));
430/// assert_eq!(1.0, sorensen_dice("ferris", "ferris"));
431/// assert_eq!(0.8888888888888888, sorensen_dice("feris", "ferris"));
432/// ```
433pub fn sorensen_dice(a: &str, b: &str) -> f64 {
434 // implementation guided by
435 // https://github.com/aceakash/string-similarity/blob/f83ba3cd7bae874c20c429774e911ae8cff8bced/src/index.js#L6
436
437 let a: String = a.chars().filter(|&x| !char::is_whitespace(x)).collect();
438 let b: String = b.chars().filter(|&x| !char::is_whitespace(x)).collect();
439
440 if a.len() == 0 && b.len() == 0 {
441 return 1.0;
442 }
443
444 if a.len() == 0 || b.len() == 0 {
445 return 0.0;
446 }
447
448 if a == b {
449 return 1.0;
450 }
451
452 if a.len() == 1 && b.len() == 1 {
453 return 0.0;
454 }
455
456 if a.len() < 2 || b.len() < 2 {
457 return 0.0;
458 }
459
460 let mut a_bigrams: HashMap<(char, char), usize> = HashMap::new();
461
462 for bigram in bigrams(&a) {
463 *a_bigrams.entry(bigram).or_insert(0) += 1;
464 }
465
466 let mut intersection_size = 0;
467
468 for bigram in bigrams(&b) {
469 a_bigrams.entry(bigram).and_modify(|bi| {
470 if *bi > 0 {
471 *bi -= 1;
472 intersection_size += 1;
473 }
474 });
475 }
476
477 (2 * intersection_size) as f64 / (a.len() + b.len() - 2) as f64
478}
479
480
481#[cfg(test)]
482mod tests {
483 use super::*;
484
485 #[test]
486 fn bigrams_iterator() {
487 let mut bi = bigrams("abcde");
488
489 assert_eq!(Some(('a', 'b')), bi.next());
490 assert_eq!(Some(('b', 'c')), bi.next());
491 assert_eq!(Some(('c', 'd')), bi.next());
492 assert_eq!(Some(('d', 'e')), bi.next());
493 assert_eq!(None, bi.next());
494 }
495
496 fn assert_hamming_dist(dist: usize, str1: &str, str2: &str) {
497 assert_eq!(Ok(dist), hamming(str1, str2));
498 }
499
500 #[test]
501 fn hamming_empty() {
502 assert_hamming_dist(0, "", "")
503 }
504
505 #[test]
506 fn hamming_same() {
507 assert_hamming_dist(0, "hamming", "hamming")
508 }
509
510 #[test]
511 fn hamming_numbers() {
512 assert_eq!(Ok(1), generic_hamming(&[1, 2, 4], &[1, 2, 3]));
513 }
514
515 #[test]
516 fn hamming_diff() {
517 assert_hamming_dist(3, "hamming", "hammers")
518 }
519
520 #[test]
521 fn hamming_diff_multibyte() {
522 assert_hamming_dist(2, "hamming", "h香mmüng");
523 }
524
525 #[test]
526 fn hamming_unequal_length() {
527 assert_eq!(
528 Err(StrSimError::DifferentLengthArgs),
529 generic_hamming("ham".chars(), "hamming".chars())
530 );
531 }
532
533 #[test]
534 fn hamming_names() {
535 assert_hamming_dist(14, "Friedrich Nietzs", "Jean-Paul Sartre")
536 }
537
538 #[test]
539 fn jaro_both_empty() {
540 assert_eq!(1.0, jaro("", ""));
541 }
542
543 #[test]
544 fn jaro_first_empty() {
545 assert_eq!(0.0, jaro("", "jaro"));
546 }
547
548 #[test]
549 fn jaro_second_empty() {
550 assert_eq!(0.0, jaro("distance", ""));
551 }
552
553 #[test]
554 fn jaro_same() {
555 assert_eq!(1.0, jaro("jaro", "jaro"));
556 }
557
558 #[test]
559 fn jaro_multibyte() {
560 assert!((0.818 - jaro("testabctest", "testöঙ香test")) < 0.001);
561 assert!((0.818 - jaro("testöঙ香test", "testabctest")) < 0.001);
562 }
563
564 #[test]
565 fn jaro_diff_short() {
566 assert!((0.767 - jaro("dixon", "dicksonx")).abs() < 0.001);
567 }
568
569 #[test]
570 fn jaro_diff_one_character() {
571 assert_eq!(0.0, jaro("a", "b"));
572 }
573
574 #[test]
575 fn jaro_same_one_character() {
576 assert_eq!(1.0, jaro("a", "a"));
577 }
578
579 #[test]
580 fn generic_jaro_diff() {
581 assert_eq!(0.0, generic_jaro(&[1, 2], &[3, 4]));
582 }
583
584 #[test]
585 fn jaro_diff_one_and_two() {
586 assert!((0.83 - jaro("a", "ab")).abs() < 0.01);
587 }
588
589 #[test]
590 fn jaro_diff_two_and_one() {
591 assert!((0.83 - jaro("ab", "a")).abs() < 0.01);
592 }
593
594 #[test]
595 fn jaro_diff_no_transposition() {
596 assert!((0.822 - jaro("dwayne", "duane")).abs() < 0.001);
597 }
598
599 #[test]
600 fn jaro_diff_with_transposition() {
601 assert!((0.944 - jaro("martha", "marhta")).abs() < 0.001);
602 }
603
604 #[test]
605 fn jaro_names() {
606 assert!((0.392 - jaro("Friedrich Nietzsche",
607 "Jean-Paul Sartre")).abs() < 0.001);
608 }
609
610 #[test]
611 fn jaro_winkler_both_empty() {
612 assert_eq!(1.0, jaro_winkler("", ""));
613 }
614
615 #[test]
616 fn jaro_winkler_first_empty() {
617 assert_eq!(0.0, jaro_winkler("", "jaro-winkler"));
618 }
619
620 #[test]
621 fn jaro_winkler_second_empty() {
622 assert_eq!(0.0, jaro_winkler("distance", ""));
623 }
624
625 #[test]
626 fn jaro_winkler_same() {
627 assert_eq!(1.0, jaro_winkler("Jaro-Winkler", "Jaro-Winkler"));
628 }
629
630 #[test]
631 fn jaro_winkler_multibyte() {
632 assert!((0.89 - jaro_winkler("testabctest", "testöঙ香test")).abs() <
633 0.001);
634 assert!((0.89 - jaro_winkler("testöঙ香test", "testabctest")).abs() <
635 0.001);
636 }
637
638 #[test]
639 fn jaro_winkler_diff_short() {
640 assert!((0.813 - jaro_winkler("dixon", "dicksonx")).abs() < 0.001);
641 assert!((0.813 - jaro_winkler("dicksonx", "dixon")).abs() < 0.001);
642 }
643
644 #[test]
645 fn jaro_winkler_diff_one_character() {
646 assert_eq!(0.0, jaro_winkler("a", "b"));
647 }
648
649 #[test]
650 fn jaro_winkler_same_one_character() {
651 assert_eq!(1.0, jaro_winkler("a", "a"));
652 }
653
654 #[test]
655 fn jaro_winkler_diff_no_transposition() {
656 assert!((0.840 - jaro_winkler("dwayne", "duane")).abs() < 0.001);
657 }
658
659 #[test]
660 fn jaro_winkler_diff_with_transposition() {
661 assert!((0.961 - jaro_winkler("martha", "marhta")).abs() < 0.001);
662 }
663
664 #[test]
665 fn jaro_winkler_names() {
666 assert!((0.562 - jaro_winkler("Friedrich Nietzsche",
667 "Fran-Paul Sartre")).abs() < 0.001);
668 }
669
670 #[test]
671 fn jaro_winkler_long_prefix() {
672 assert!((0.911 - jaro_winkler("cheeseburger", "cheese fries")).abs() <
673 0.001);
674 }
675
676 #[test]
677 fn jaro_winkler_more_names() {
678 assert!((0.868 - jaro_winkler("Thorkel", "Thorgier")).abs() < 0.001);
679 }
680
681 #[test]
682 fn jaro_winkler_length_of_one() {
683 assert!((0.738 - jaro_winkler("Dinsdale", "D")).abs() < 0.001);
684 }
685
686 #[test]
687 fn jaro_winkler_very_long_prefix() {
688 assert!((1.0 - jaro_winkler("thequickbrownfoxjumpedoverx",
689 "thequickbrownfoxjumpedovery")).abs() <
690 0.001);
691 }
692
693 #[test]
694 fn levenshtein_empty() {
695 assert_eq!(0, levenshtein("", ""));
696 }
697
698 #[test]
699 fn levenshtein_same() {
700 assert_eq!(0, levenshtein("levenshtein", "levenshtein"));
701 }
702
703 #[test]
704 fn levenshtein_diff_short() {
705 assert_eq!(3, levenshtein("kitten", "sitting"));
706 }
707
708 #[test]
709 fn levenshtein_diff_with_space() {
710 assert_eq!(5, levenshtein("hello, world", "bye, world"));
711 }
712
713 #[test]
714 fn levenshtein_diff_multibyte() {
715 assert_eq!(3, levenshtein("öঙ香", "abc"));
716 assert_eq!(3, levenshtein("abc", "öঙ香"));
717 }
718
719 #[test]
720 fn levenshtein_diff_longer() {
721 let a = "The quick brown fox jumped over the angry dog.";
722 let b = "Lorem ipsum dolor sit amet, dicta latine an eam.";
723 assert_eq!(37, levenshtein(a, b));
724 }
725
726 #[test]
727 fn levenshtein_first_empty() {
728 assert_eq!(7, levenshtein("", "sitting"));
729 }
730
731 #[test]
732 fn levenshtein_second_empty() {
733 assert_eq!(6, levenshtein("kitten", ""));
734 }
735
736 #[test]
737 fn normalized_levenshtein_diff_short() {
738 assert!((normalized_levenshtein("kitten", "sitting") - 0.57142).abs() < 0.00001);
739 }
740
741 #[test]
742 fn normalized_levenshtein_for_empty_strings() {
743 assert!((normalized_levenshtein("", "") - 1.0).abs() < 0.00001);
744 }
745
746 #[test]
747 fn normalized_levenshtein_first_empty() {
748 assert!(normalized_levenshtein("", "second").abs() < 0.00001);
749 }
750
751 #[test]
752 fn normalized_levenshtein_second_empty() {
753 assert!(normalized_levenshtein("first", "").abs() < 0.00001);
754 }
755
756 #[test]
757 fn normalized_levenshtein_identical_strings() {
758 assert!((normalized_levenshtein("identical", "identical") - 1.0).abs() < 0.00001);
759 }
760
761 #[test]
762 fn osa_distance_empty() {
763 assert_eq!(0, osa_distance("", ""));
764 }
765
766 #[test]
767 fn osa_distance_same() {
768 assert_eq!(0, osa_distance("damerau", "damerau"));
769 }
770
771 #[test]
772 fn osa_distance_first_empty() {
773 assert_eq!(7, osa_distance("", "damerau"));
774 }
775
776 #[test]
777 fn osa_distance_second_empty() {
778 assert_eq!(7, osa_distance("damerau", ""));
779 }
780
781 #[test]
782 fn osa_distance_diff() {
783 assert_eq!(3, osa_distance("ca", "abc"));
784 }
785
786 #[test]
787 fn osa_distance_diff_short() {
788 assert_eq!(3, osa_distance("damerau", "aderua"));
789 }
790
791 #[test]
792 fn osa_distance_diff_reversed() {
793 assert_eq!(3, osa_distance("aderua", "damerau"));
794 }
795
796 #[test]
797 fn osa_distance_diff_multibyte() {
798 assert_eq!(3, osa_distance("öঙ香", "abc"));
799 assert_eq!(3, osa_distance("abc", "öঙ香"));
800 }
801
802 #[test]
803 fn osa_distance_diff_unequal_length() {
804 assert_eq!(6, osa_distance("damerau", "aderuaxyz"));
805 }
806
807 #[test]
808 fn osa_distance_diff_unequal_length_reversed() {
809 assert_eq!(6, osa_distance("aderuaxyz", "damerau"));
810 }
811
812 #[test]
813 fn osa_distance_diff_comedians() {
814 assert_eq!(5, osa_distance("Stewart", "Colbert"));
815 }
816
817 #[test]
818 fn osa_distance_many_transpositions() {
819 assert_eq!(4, osa_distance("abcdefghijkl", "bacedfgihjlk"));
820 }
821
822 #[test]
823 fn osa_distance_diff_longer() {
824 let a = "The quick brown fox jumped over the angry dog.";
825 let b = "Lehem ipsum dolor sit amet, dicta latine an eam.";
826 assert_eq!(36, osa_distance(a, b));
827 }
828
829 #[test]
830 fn osa_distance_beginning_transposition() {
831 assert_eq!(1, osa_distance("foobar", "ofobar"));
832 }
833
834 #[test]
835 fn osa_distance_end_transposition() {
836 assert_eq!(1, osa_distance("specter", "spectre"));
837 }
838
839 #[test]
840 fn osa_distance_restricted_edit() {
841 assert_eq!(4, osa_distance("a cat", "an abct"));
842 }
843
844 #[test]
845 fn damerau_levenshtein_empty() {
846 assert_eq!(0, damerau_levenshtein("", ""));
847 }
848
849 #[test]
850 fn damerau_levenshtein_same() {
851 assert_eq!(0, damerau_levenshtein("damerau", "damerau"));
852 }
853
854 #[test]
855 fn damerau_levenshtein_first_empty() {
856 assert_eq!(7, damerau_levenshtein("", "damerau"));
857 }
858
859 #[test]
860 fn damerau_levenshtein_second_empty() {
861 assert_eq!(7, damerau_levenshtein("damerau", ""));
862 }
863
864 #[test]
865 fn damerau_levenshtein_diff() {
866 assert_eq!(2, damerau_levenshtein("ca", "abc"));
867 }
868
869 #[test]
870 fn damerau_levenshtein_diff_short() {
871 assert_eq!(3, damerau_levenshtein("damerau", "aderua"));
872 }
873
874 #[test]
875 fn damerau_levenshtein_diff_reversed() {
876 assert_eq!(3, damerau_levenshtein("aderua", "damerau"));
877 }
878
879 #[test]
880 fn damerau_levenshtein_diff_multibyte() {
881 assert_eq!(3, damerau_levenshtein("öঙ香", "abc"));
882 assert_eq!(3, damerau_levenshtein("abc", "öঙ香"));
883 }
884
885 #[test]
886 fn damerau_levenshtein_diff_unequal_length() {
887 assert_eq!(6, damerau_levenshtein("damerau", "aderuaxyz"));
888 }
889
890 #[test]
891 fn damerau_levenshtein_diff_unequal_length_reversed() {
892 assert_eq!(6, damerau_levenshtein("aderuaxyz", "damerau"));
893 }
894
895 #[test]
896 fn damerau_levenshtein_diff_comedians() {
897 assert_eq!(5, damerau_levenshtein("Stewart", "Colbert"));
898 }
899
900 #[test]
901 fn damerau_levenshtein_many_transpositions() {
902 assert_eq!(4, damerau_levenshtein("abcdefghijkl", "bacedfgihjlk"));
903 }
904
905 #[test]
906 fn damerau_levenshtein_diff_longer() {
907 let a = "The quick brown fox jumped over the angry dog.";
908 let b = "Lehem ipsum dolor sit amet, dicta latine an eam.";
909 assert_eq!(36, damerau_levenshtein(a, b));
910 }
911
912 #[test]
913 fn damerau_levenshtein_beginning_transposition() {
914 assert_eq!(1, damerau_levenshtein("foobar", "ofobar"));
915 }
916
917 #[test]
918 fn damerau_levenshtein_end_transposition() {
919 assert_eq!(1, damerau_levenshtein("specter", "spectre"));
920 }
921
922 #[test]
923 fn damerau_levenshtein_unrestricted_edit() {
924 assert_eq!(3, damerau_levenshtein("a cat", "an abct"));
925 }
926
927 #[test]
928 fn normalized_damerau_levenshtein_diff_short() {
929 assert!((normalized_damerau_levenshtein("levenshtein", "löwenbräu") - 0.27272).abs() < 0.00001);
930 }
931
932 #[test]
933 fn normalized_damerau_levenshtein_for_empty_strings() {
934 assert!((normalized_damerau_levenshtein("", "") - 1.0).abs() < 0.00001);
935 }
936
937 #[test]
938 fn normalized_damerau_levenshtein_first_empty() {
939 assert!(normalized_damerau_levenshtein("", "flower").abs() < 0.00001);
940 }
941
942 #[test]
943 fn normalized_damerau_levenshtein_second_empty() {
944 assert!(normalized_damerau_levenshtein("tree", "").abs() < 0.00001);
945 }
946
947 #[test]
948 fn normalized_damerau_levenshtein_identical_strings() {
949 assert!((normalized_damerau_levenshtein("sunglasses", "sunglasses") - 1.0).abs() < 0.00001);
950 }
951
952 #[test]
953 fn sorensen_dice_all() {
954 // test cases taken from
955 // https://github.com/aceakash/string-similarity/blob/f83ba3cd7bae874c20c429774e911ae8cff8bced/src/spec/index.spec.js#L11
956
957 assert_eq!(1.0, sorensen_dice("a", "a"));
958 assert_eq!(0.0, sorensen_dice("a", "b"));
959 assert_eq!(1.0, sorensen_dice("", ""));
960 assert_eq!(0.0, sorensen_dice("a", ""));
961 assert_eq!(0.0, sorensen_dice("", "a"));
962 assert_eq!(1.0, sorensen_dice("apple event", "apple event"));
963 assert_eq!(0.9090909090909091, sorensen_dice("iphone", "iphone x"));
964 assert_eq!(0.0, sorensen_dice("french", "quebec"));
965 assert_eq!(1.0, sorensen_dice("france", "france"));
966 assert_eq!(0.2, sorensen_dice("fRaNce", "france"));
967 assert_eq!(0.8, sorensen_dice("healed", "sealed"));
968 assert_eq!(
969 0.7878787878787878,
970 sorensen_dice("web applications", "applications of the web")
971 );
972 assert_eq!(
973 0.92,
974 sorensen_dice(
975 "this will have a typo somewhere",
976 "this will huve a typo somewhere"
977 )
978 );
979 assert_eq!(
980 0.6060606060606061,
981 sorensen_dice(
982 "Olive-green table for sale, in extremely good condition.",
983 "For sale: table in very good condition, olive green in colour."
984 )
985 );
986 assert_eq!(
987 0.2558139534883721,
988 sorensen_dice(
989 "Olive-green table for sale, in extremely good condition.",
990 "For sale: green Subaru Impreza, 210,000 miles"
991 )
992 );
993 assert_eq!(
994 0.1411764705882353,
995 sorensen_dice(
996 "Olive-green table for sale, in extremely good condition.",
997 "Wanted: mountain bike with at least 21 gears."
998 )
999 );
1000 assert_eq!(
1001 0.7741935483870968,
1002 sorensen_dice("this has one extra word", "this has one word")
1003 );
1004 }
1005}
1006