v/vlib/strings/similarity.v

71 lines
1.8 KiB
V
Raw Normal View History

module strings
2019-12-19 22:29:37 +01:00
// #-js
// use levenshtein distance algorithm to calculate
// the distance between between two strings (lower is closer)
pub fn levenshtein_distance(a, b string) int {
2019-12-18 19:56:30 +01:00
mut f := [0].repeat(b.len + 1)
for j in 0..f.len {
f[j] = j
}
for ca in a {
mut j := 1
mut fj1 := f[0]
f[0]++
for cb in b {
2019-12-18 19:56:30 +01:00
mut mn := if f[j] + 1 <= f[j - 1] + 1 { f[j] + 1 } else { f[j - 1] + 1 }
if cb != ca {
2019-12-18 19:56:30 +01:00
mn = if mn <= fj1 + 1 { mn } else { fj1 + 1 }
}
else {
mn = if mn <= fj1 { mn } else { fj1 }
}
fj1 = f[j]
f[j] = mn
j++
}
}
2019-12-18 19:56:30 +01:00
return f[f.len - 1]
}
// use levenshtein distance algorithm to calculate
// how similar two strings are as a percentage (higher is closer)
2019-09-13 15:15:30 +02:00
pub fn levenshtein_distance_percentage(a, b string) f32 {
d := levenshtein_distance(a, b)
l := if a.len >= b.len { a.len } else { b.len }
2019-12-18 19:56:30 +01:00
return (1.00 - f32(d) / f32(l)) * 100.00
}
// implementation of SørensenDice coefficient.
// find the similarity between two strings.
2019-09-21 13:59:06 +02:00
// returns coefficient between 0.0 (not similar) and 1.0 (exact match).
2019-09-13 15:15:30 +02:00
pub fn dice_coefficient(s1, s2 string) f32 {
2019-12-18 19:56:30 +01:00
if s1.len == 0 || s2.len == 0 {
return 0.0
}
if s1 == s2 {
return 1.0
}
if s1.len < 2 || s2.len < 2 {
return 0.0
}
2019-09-15 19:07:12 +02:00
a := if s1.len > s2.len { s1 } else { s2 }
b := if a == s1 { s2 } else { s1 }
mut first_bigrams := map[string]int
for i in 0..a.len - 1 {
2019-12-18 19:56:30 +01:00
bigram := a[i..i + 2]
q := if bigram in first_bigrams { first_bigrams[bigram] + 1 } else { 1 }
2019-10-07 00:31:01 +02:00
first_bigrams[bigram] = q
}
mut intersection_size := 0
for i in 0..b.len - 1 {
2019-12-18 19:56:30 +01:00
bigram := b[i..i + 2]
2019-09-19 04:22:24 +02:00
count := if bigram in first_bigrams { first_bigrams[bigram] } else { 0 }
if count > 0 {
2019-09-15 19:07:12 +02:00
first_bigrams[bigram] = count - 1
intersection_size++
}
}
return (2.0 * f32(intersection_size)) / (f32(a.len) + f32(b.len) - 2)
}
2019-12-19 22:29:37 +01:00