code

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#define MAX_WORD_LEN 32
#define MAX_LINE_LEN 1024
#define PHRASE_LEN 3
typedef struct TrieNode {
 int is_end;
 struct TrieNode *children[256];
} TrieNode;
TrieNode* createNode() {
 TrieNode* node = (TrieNode*)malloc(sizeof(TrieNode));
 node->is_end = 0;
 for (int i = 0; i < 256; i++) node->children[i] = NULL;
 return node;
}
void insert(TrieNode* root, char* phrase) {
 TrieNode* curr = root;
 for (int i = 0; phrase[i]; i++) {
 unsigned char c = (unsigned char)phrase[i];
 if (!curr->children[c])
 curr->children[c] = createNode();
 curr = curr->children[c];
 }
 curr->is_end = 1;
}
int search(TrieNode* root, char* phrase) {
 TrieNode* curr = root;
 for (int i = 0; phrase[i]; i++) {
 unsigned char c = (unsigned char)phrase[i];
 if (!curr->children[c])
 return 0;
 curr = curr->children[c];
 }
 return curr->is_end;
}
void sanitize(char* word) {
 int j = 0;
 for (int i = 0; word[i]; i++) {
 if (isalnum(word[i])) {
 word[j++] = tolower(word[i]);
 }
21
 }
 word[j] = '\0';
}
int tokenize(char* sentence, char words[][MAX_WORD_LEN]) {
 int count = 0;
 char* token = strtok(sentence, " ");
 while (token && count < 1000) {
 sanitize(token);
 if (strlen(token) > 0) {
 strncpy(words[count++], token, MAX_WORD_LEN);
 }
 token = strtok(NULL, " ");
 }
 return count;
}
void buildPhrases(char words[][MAX_WORD_LEN], int word_count, char
phrases[][MAX_LINE_LEN], int* phrase_count) {
 *phrase_count = 0;
 for (int i = 0; i <= word_count - PHRASE_LEN; i++) {
 snprintf(phrases[*phrase_count], MAX_LINE_LEN, "%s %s %s", words[i],
words[i+1], words[i+2]);
 (*phrase_count)++;
 }
}
void readFile(const char* filename, char* buffer) {
 FILE* fp = fopen(filename, "r");
 if (!fp) {
 perror("File read error");
 exit(1);
 }
 size_t len = fread(buffer, 1, 10000, fp);
 buffer[len] = '\0';
 fclose(fp);
}
int main() {
 char baseText[10000];
 char testText[10000];
 char words[1000][MAX_WORD_LEN];
 char phrases[1000][MAX_LINE_LEN];
 int phrase_count = 0, matched = 0;
 char baseFile[100], testFile[100];
 printf("Enter the path to the original file: ");
 scanf("%s", baseFile);
 printf("Enter the path to the test file: ");
22
 scanf("%s", testFile);
 TrieNode* root = createNode();
 readFile(baseFile, baseText);
 readFile(testFile, testText);
 int word_count = tokenize(baseText, words);
 buildPhrases(words, word_count, phrases, &phrase_count);
 for (int i = 0; i < phrase_count; i++)
 insert(root, phrases[i]);
 int test_phrase_count = 0;
 int test_word_count = tokenize(testText, words);
 buildPhrases(words, test_word_count, phrases, &test_phrase_count);
 for (int i = 0; i < test_phrase_count; i++) {
 if (search(root, phrases[i])) matched++;
 }
 float percent = (float)matched / test_phrase_count * 100.0;
 printf("\n=== Plagiarism Report ===\n");
 printf("Total Phrases Checked: %d\n", test_phrase_count);
 printf("Matched Phrases: %d\n", matched);
 printf("Plagiarism Percentage: %.2f%%\n", percent);
 if (percent > 30.0)
 printf("⚠️ Plagiarism Suspected!\n");
 else
 printf("✅ Likely Original Work.\n");
 return 0;
}