feat: let KMP algorithm return index (#2713)

* fix: KMP algorithm (#2712)

* feat: let the KMP algorithm return index and add more tests.

* feat: format code

* Update strings/knuth_morris_pratt.cpp

* update knuth_morris_pratt.cpp

---------

Co-authored-by: realstealthninja <68815218+realstealthninja@users.noreply.github.com>
This commit is contained in:
Yancey
2024-09-01 12:17:49 +08:00
committed by GitHub
parent 9702903a12
commit 502a953e23

View File

@@ -1,95 +1,98 @@
/**
* \file
* \brief The [Knuth-Morris-Pratt
* @file
* @brief The [Knuth-Morris-Pratt
* Algorithm](https://en.wikipedia.org/wiki/KnuthMorrisPratt_algorithm) for
* finding a pattern within a piece of text with complexity O(n + m)
*
* @details
* 1. Preprocess pattern to identify any suffixes that are identical to
* prefixes. This tells us where to continue from if we get a mismatch between a
* character in our pattern and the text.
* 2. Step through the text one character at a time and compare it to a
* character in the pattern updating our location within the pattern if
* necessary
* @author [Yancey](https://github.com/Yancey2023)
*/
#include <iostream>
#ifdef _MSC_VER
#include <string> // use this for MS Visual C++
#else
#include <cstring>
#endif
#include <vector>
#include <cassert> /// for assert
#include <iostream> /// for IO operations
#include <string> /// for std::string
#include <vector> /// for std::vector
/** \namespace string_search
* \brief String search algorithms
/**
* @namespace string_search
* @brief String search algorithms
*/
namespace string_search {
/**
* Generate the partial match table aka failure function for a pattern to
* @brief Generate the partial match table aka failure function for a pattern to
* search.
* \param[in] pattern text for which to create the partial match table
* \returns the partial match table as a vector array
* @param pattern text for which to create the partial match table
* @returns the partial match table as a vector array
*/
std::vector<int> getFailureArray(const std::string &pattern) {
int pattern_length = pattern.size();
std::vector<int> failure(pattern_length + 1);
failure[0] = -1;
int j = -1;
std::vector<size_t> getFailureArray(const std::string &pattern) {
size_t pattern_length = pattern.size();
std::vector<size_t> failure(pattern_length + 1);
failure[0] = std::string::npos;
size_t j = std::string::npos;
for (int i = 0; i < pattern_length; i++) {
while (j != -1 && pattern[j] != pattern[i]) {
while (j != std::string::npos && pattern[j] != pattern[i]) {
j = failure[j];
}
j++;
failure[i + 1] = j;
failure[i + 1] = ++j;
}
return failure;
}
/**
* KMP algorithm to find a pattern in a text
* \param[in] pattern string pattern to search
* \param[in] text text in which to search
* \returns `true` if pattern was found
* \returns `false` if pattern was not found
* @brief KMP algorithm to find a pattern in a text
* @param pattern string pattern to search
* @param text text in which to search
* @returns the starting index of the pattern if found
* @returns `std::string::npos` if not found
*/
bool kmp(const std::string &pattern, const std::string &text) {
int text_length = text.size(), pattern_length = pattern.size();
std::vector<int> failure = getFailureArray(pattern);
int k = 0;
for (int j = 0; j < text_length; j++) {
while (k != -1 && pattern[k] != text[j]) {
size_t kmp(const std::string &pattern, const std::string &text) {
if (pattern.empty()) {
return 0;
}
std::vector<size_t> failure = getFailureArray(pattern);
size_t text_length = text.size();
size_t pattern_length = pattern.size();
size_t k = 0;
for (size_t j = 0; j < text_length; j++) {
while (k != std::string::npos && pattern[k] != text[j]) {
k = failure[k];
}
k++;
if (k == pattern_length)
return true;
if (++k == pattern_length) {
return j - k + 1;
}
}
return false;
return std::string::npos;
}
} // namespace string_search
using string_search::kmp;
/** Main function */
/**
* @brief self-test implementations
* @returns void
*/
static void tests() {
assert(kmp("abc1abc12l", "alskfjaldsabc1abc1abc12k2") == std::string::npos);
assert(kmp("bca", "abcabc") == 1);
assert(kmp("World", "helloWorld") == 5);
assert(kmp("c++", "his_is_c++") == 7);
assert(kmp("happy", "happy_coding") == 0);
assert(kmp("", "pattern is empty") == 0);
// this lets the user know that the tests have passed
std::cout << "All KMP algorithm tests have successfully passed!\n";
}
/*
* @brief Main function
* @returns 0 on exit
*/
int main() {
std::string text = "alskfjaldsabc1abc1abc12k23adsfabcabc";
std::string pattern = "abc1abc12l";
if (kmp(pattern, text) == true) {
std::cout << "Found" << std::endl;
} else {
std::cout << "Not Found" << std::endl;
}
text = "abcabc";
pattern = "bca";
if (kmp(pattern, text) == true) {
std::cout << "Found" << std::endl;
} else {
std::cout << "Not Found" << std::endl;
}
tests();
return 0;
}