From abb1d5cb1bd545e7ac0cdcf44701e2f586924aab Mon Sep 17 00:00:00 2001
From: Unbewohnte <me@unbewohnte.su>
Date: Fri, 20 Jan 2023 20:26:40 +0300
Subject: [PATCH] Levenshtein string distance

---
 src/strings/levenshtein.c | 54 +++++++++++++++++++++++++++++++++++++++
 src/strings/levenshtein.h |  4 +++
 testing/test.c            | 32 +++++++++++++++++++++++
 3 files changed, 90 insertions(+)
 create mode 100644 src/strings/levenshtein.c
 create mode 100644 src/strings/levenshtein.h

diff --git a/src/strings/levenshtein.c b/src/strings/levenshtein.c
new file mode 100644
index 0000000..c37edae
--- /dev/null
+++ b/src/strings/levenshtein.c
@@ -0,0 +1,54 @@
+#include "levenshtein.h"
+
+int _min(int a, int b) {
+    if (a < b) {
+        return a;
+    }
+
+    return b;
+}
+
+// Calculates Levenshtein distance for str1 and str2 of length len1 and len2 respectively
+int levenshtein_distance(const char* str1, const char* str2) {
+    unsigned int len1 = strlen(str1) + 1;
+    unsigned int len2 = strlen(str2) + 1;
+
+    if (len1 == 0) {
+        return len2;
+    }
+
+    if (len2 == 0) {
+        return len1;
+    }
+
+    if (strcmp(str1, str2) == 0) {
+        return 0;
+    }
+
+    char s1[len2];
+    char s2[len2];
+    char row[len1];
+
+    strcpy(s1, str1);
+    strcpy(s2, str2);
+
+    for (unsigned int i = 0; i < len1; i++) {
+        row[i] = i;
+    }
+
+    for (unsigned int i = 1; i < len2; i++) {
+        int previous = i;
+
+        for (unsigned int j = 1; j < len1; j++) {
+            int current = row[j-1];
+            if (s2[i-1] != s1[j-1]) {
+                current = _min(_min(row[j-1]+1, previous+1), row[j]+1);
+            }
+            row[j-1] = previous;
+            previous = current;
+        }
+        row[len1] = previous;
+    }
+
+    return row[len1];
+}
\ No newline at end of file
diff --git a/src/strings/levenshtein.h b/src/strings/levenshtein.h
new file mode 100644
index 0000000..9f8d942
--- /dev/null
+++ b/src/strings/levenshtein.h
@@ -0,0 +1,4 @@
+#include <string.h>
+
+// Calculates Levenshtein distance for str1 and str2
+int levenshtein_distance(const char* str1, const char* str2);
\ No newline at end of file
diff --git a/testing/test.c b/testing/test.c
index bb31709..6771bc7 100644
--- a/testing/test.c
+++ b/testing/test.c
@@ -26,6 +26,7 @@ THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR I
 #include "../src/math/vector.h"
 #include "../src/math/misc.h"
 #include "../src/datastruct/cvec.h"
+#include "../src/strings/levenshtein.h"
 
 int test_rng() {
     lcg(76);
@@ -302,6 +303,29 @@ int test_datastruct() {
     return EXIT_SUCCESS;
 }
 
+int test_levenshtein() {
+    const char* str1 = "Fellow";
+    const char* str2 = "Fella";
+
+    int distance = levenshtein_distance(str1, str2);
+
+    if (distance != 2) {
+        printf("[ERROR] Invalid string distance between \"%s\" and \"%s\": expected to be %d; got %d\n", str1, str2, 2, distance);
+        return EXIT_FAILURE;
+    }
+
+    return EXIT_SUCCESS;
+}
+
+int test_strings() {
+    if (test_levenshtein() == EXIT_FAILURE) {
+        printf("[ERROR] Levenshtein test failed\n");
+        return EXIT_FAILURE;
+    }
+
+    return EXIT_SUCCESS;
+}
+
 int main() {
     // rng
     printf("[INFO] Testing rng...\n");
@@ -359,5 +383,13 @@ int main() {
         printf("[INFO] Datastruct test passed\n\n");
     }
 
+    // strings
+        printf("[INFO] Testing strings...\n");
+    if (test_strings() == EXIT_FAILURE) {
+        printf("[INFO] Strings test failed\n\n");
+    } else {
+        printf("[INFO] Strings test passed\n\n");
+    }
+
     return EXIT_SUCCESS;
 }
\ No newline at end of file