Skip to content

Commit 60093d7

Browse files
authored
feat: add duval's algorithm (#2725)
* feat: Add Duval's algorithm for the lexicographically smallest rotation in a sequence. * fixes. * fixes.
1 parent 0ecb6bd commit 60093d7

File tree

1 file changed

+118
-0
lines changed

1 file changed

+118
-0
lines changed

strings/duval.cpp

+118
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,118 @@
1+
/**
2+
* @file duval.cpp
3+
* @brief Implementation of [Duval's algorithm](https://en.wikipedia.org/wiki/Lyndon_word).
4+
*
5+
* @details
6+
* Duval's algorithm is an algorithm to find the lexicographically smallest
7+
* rotation of a string. It is based on the concept of Lyndon words.
8+
* Lyndon words are defined as the lexicographically smallest string in a
9+
* rotation equivalence class. A rotation equivalence class is a set of strings
10+
* that can be obtained by rotating a string. For example, the rotation
11+
* equivalence class of "abc" is {"abc", "bca", "cab"}. The lexicographically
12+
* smallest string in this class is "abc".
13+
*
14+
* Duval's algorithm works by iterating over the string and finding the
15+
* smallest rotation of the string that is a Lyndon word. This is done by
16+
* comparing the string with its suffixes and finding the smallest suffix that
17+
* is lexicographically smaller than the string. This suffix is then added to
18+
* the result and the process is repeated with the remaining string.
19+
* The algorithm has a time complexity of O(n) where n is the length of the
20+
* string.
21+
*
22+
* @note While Lyndon words are described in the context of strings,
23+
* Duval's algorithm can be used to find the lexicographically smallest cyclic
24+
* shift of any sequence of comparable elements.
25+
*
26+
* @author [Amine Ghoussaini](https://github.com/aminegh20)
27+
*/
28+
29+
#include <array> /// for std::array
30+
#include <cassert> /// for assert
31+
#include <cstddef> /// for std::size_t
32+
#include <deque> /// for std::deque
33+
#include <iostream> /// for std::cout and std::endl
34+
#include <string> /// for std::string
35+
#include <vector> /// for std::vector
36+
37+
/**
38+
* @brief string manipulation algorithms
39+
* @namespace
40+
*/
41+
namespace string {
42+
/**
43+
* @brief Find the lexicographically smallest cyclic shift of a sequence.
44+
* @tparam T type of the sequence
45+
* @param s the sequence
46+
* @returns the 0-indexed position of the least cyclic shift of the sequence
47+
*/
48+
template <typename T>
49+
size_t duval(const T& s) {
50+
size_t n = s.size();
51+
size_t i = 0, ans = 0;
52+
while (i < n) {
53+
ans = i;
54+
size_t j = i + 1, k = i;
55+
while (j < (n + n) && s[j % n] >= s[k % n]) {
56+
if (s[k % n] < s[j % n]) {
57+
k = i;
58+
} else {
59+
k++;
60+
}
61+
j++;
62+
}
63+
while (i <= k) {
64+
i += j - k;
65+
}
66+
}
67+
return ans;
68+
// returns 0-indexed position of the least cyclic shift
69+
}
70+
71+
} // namespace string
72+
73+
/**
74+
* @brief self test implementation
75+
* returns void
76+
*/
77+
static void test() {
78+
using namespace string;
79+
80+
// Test 1
81+
std::string s1 = "abcab";
82+
assert(duval(s1) == 3);
83+
84+
// Test 2
85+
std::string s2 = "011100";
86+
assert(duval(s2) == 4);
87+
88+
// Test 3
89+
std::vector<int> v = {5, 2, 1, 3, 4};
90+
assert(duval(v) == 2);
91+
92+
// Test 4
93+
std::array<int, 5> a = {1, 2, 3, 4, 5};
94+
assert(duval(a) == 0);
95+
96+
// Test 5
97+
std::deque<char> d = {'a', 'z', 'c', 'a', 'b'};
98+
assert(duval(d) == 3);
99+
100+
// Test 6
101+
std::string s3;
102+
assert(duval(s3) == 0);
103+
104+
// Test 7
105+
std::vector<int> v2 = {5, 2, 1, 3, -4};
106+
assert(duval(v2) == 4);
107+
108+
std::cout << "All tests passed!" << std::endl;
109+
}
110+
111+
/**
112+
* @brief main function
113+
* @returns 0 on exit
114+
*/
115+
int main() {
116+
test(); // run self test implementations
117+
return 0;
118+
}

0 commit comments

Comments
 (0)