| xj | b04a402 | 2021-11-25 15:01:52 +0800 | [diff] [blame] | 1 | // SPDX-License-Identifier: GPL-2.0 | 
|  | 2 | #include "levenshtein.h" | 
|  | 3 | #include <errno.h> | 
|  | 4 | #include <stdlib.h> | 
|  | 5 | #include <string.h> | 
|  | 6 |  | 
|  | 7 | /* | 
|  | 8 | * This function implements the Damerau-Levenshtein algorithm to | 
|  | 9 | * calculate a distance between strings. | 
|  | 10 | * | 
|  | 11 | * Basically, it says how many letters need to be swapped, substituted, | 
|  | 12 | * deleted from, or added to string1, at least, to get string2. | 
|  | 13 | * | 
|  | 14 | * The idea is to build a distance matrix for the substrings of both | 
|  | 15 | * strings.  To avoid a large space complexity, only the last three rows | 
|  | 16 | * are kept in memory (if swaps had the same or higher cost as one deletion | 
|  | 17 | * plus one insertion, only two rows would be needed). | 
|  | 18 | * | 
|  | 19 | * At any stage, "i + 1" denotes the length of the current substring of | 
|  | 20 | * string1 that the distance is calculated for. | 
|  | 21 | * | 
|  | 22 | * row2 holds the current row, row1 the previous row (i.e. for the substring | 
|  | 23 | * of string1 of length "i"), and row0 the row before that. | 
|  | 24 | * | 
|  | 25 | * In other words, at the start of the big loop, row2[j + 1] contains the | 
|  | 26 | * Damerau-Levenshtein distance between the substring of string1 of length | 
|  | 27 | * "i" and the substring of string2 of length "j + 1". | 
|  | 28 | * | 
|  | 29 | * All the big loop does is determine the partial minimum-cost paths. | 
|  | 30 | * | 
|  | 31 | * It does so by calculating the costs of the path ending in characters | 
|  | 32 | * i (in string1) and j (in string2), respectively, given that the last | 
|  | 33 | * operation is a substition, a swap, a deletion, or an insertion. | 
|  | 34 | * | 
|  | 35 | * This implementation allows the costs to be weighted: | 
|  | 36 | * | 
|  | 37 | * - w (as in "sWap") | 
|  | 38 | * - s (as in "Substitution") | 
|  | 39 | * - a (for insertion, AKA "Add") | 
|  | 40 | * - d (as in "Deletion") | 
|  | 41 | * | 
|  | 42 | * Note that this algorithm calculates a distance _iff_ d == a. | 
|  | 43 | */ | 
|  | 44 | int levenshtein(const char *string1, const char *string2, | 
|  | 45 | int w, int s, int a, int d) | 
|  | 46 | { | 
|  | 47 | int len1 = strlen(string1), len2 = strlen(string2); | 
|  | 48 | int *row0 = malloc(sizeof(int) * (len2 + 1)); | 
|  | 49 | int *row1 = malloc(sizeof(int) * (len2 + 1)); | 
|  | 50 | int *row2 = malloc(sizeof(int) * (len2 + 1)); | 
|  | 51 | int i, j; | 
|  | 52 |  | 
|  | 53 | for (j = 0; j <= len2; j++) | 
|  | 54 | row1[j] = j * a; | 
|  | 55 | for (i = 0; i < len1; i++) { | 
|  | 56 | int *dummy; | 
|  | 57 |  | 
|  | 58 | row2[0] = (i + 1) * d; | 
|  | 59 | for (j = 0; j < len2; j++) { | 
|  | 60 | /* substitution */ | 
|  | 61 | row2[j + 1] = row1[j] + s * (string1[i] != string2[j]); | 
|  | 62 | /* swap */ | 
|  | 63 | if (i > 0 && j > 0 && string1[i - 1] == string2[j] && | 
|  | 64 | string1[i] == string2[j - 1] && | 
|  | 65 | row2[j + 1] > row0[j - 1] + w) | 
|  | 66 | row2[j + 1] = row0[j - 1] + w; | 
|  | 67 | /* deletion */ | 
|  | 68 | if (row2[j + 1] > row1[j + 1] + d) | 
|  | 69 | row2[j + 1] = row1[j + 1] + d; | 
|  | 70 | /* insertion */ | 
|  | 71 | if (row2[j + 1] > row2[j] + a) | 
|  | 72 | row2[j + 1] = row2[j] + a; | 
|  | 73 | } | 
|  | 74 |  | 
|  | 75 | dummy = row0; | 
|  | 76 | row0 = row1; | 
|  | 77 | row1 = row2; | 
|  | 78 | row2 = dummy; | 
|  | 79 | } | 
|  | 80 |  | 
|  | 81 | i = row1[len2]; | 
|  | 82 | free(row0); | 
|  | 83 | free(row1); | 
|  | 84 | free(row2); | 
|  | 85 |  | 
|  | 86 | return i; | 
|  | 87 | } |