89 lines
2 KiB
C
89 lines
2 KiB
C
|
#include <stdio.h>
|
||
|
|
#include <stdlib.h>
|
||
|
|
#include <string.h>
|
||
|
|
|
||
|
|
int main(void){
|
||
|
|
|
||
|
|
// Count nb of char
|
||
|
|
char cchar;
|
||
|
|
int size = 0;
|
||
|
|
FILE *fptr = fopen("test.txt", "r");
|
||
|
|
while ((cchar = fgetc(fptr)) != EOF){
|
||
|
|
size += 1;
|
||
|
|
}
|
||
|
|
|
||
|
|
// Load file
|
||
|
|
char raw_file[size];
|
||
|
|
fclose(fptr);
|
||
|
|
fptr = fopen("test.txt", "r");
|
||
|
|
int i = 0;
|
||
|
|
while ((cchar = fgetc(fptr)) != EOF){
|
||
|
|
raw_file[i] = cchar;
|
||
|
|
i++;
|
||
|
|
}
|
||
|
|
fclose(fptr);
|
||
|
|
|
||
|
|
// Count nb of words and longest word
|
||
|
|
int nb_words = 0;
|
||
|
|
int longest = 0;
|
||
|
|
int c_size;
|
||
|
|
for (i = 0; i < size - 1; i++){
|
||
|
|
if ( raw_file[i] == 0 && raw_file[i + 1] != ' ' ) {
|
||
|
|
nb_words += 1;
|
||
|
|
if (c_size > longest){
|
||
|
|
longest = c_size;
|
||
|
|
}
|
||
|
|
c_size = 0;
|
||
|
|
} else {
|
||
|
|
c_size++;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
// Load words
|
||
|
|
char words[nb_words][longest];
|
||
|
|
int word = 0;
|
||
|
|
int chr = 0;
|
||
|
|
for (i = 0; i < size - 1; i++){
|
||
|
|
if ( raw_file[i] == 0 && raw_file[i + 1] != ' ' ) {
|
||
|
|
word++;
|
||
|
|
chr = 0;
|
||
|
|
} else {
|
||
|
|
char words[word][chr] = raw_file[i];
|
||
|
|
chr++;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
// Count unique words
|
||
|
|
int freq_words[nb_words];
|
||
|
|
memset(freq_words, 0, nb_words * sizeof(int));
|
||
|
|
int uni_word = 0;
|
||
|
|
for (i = 0; i < nb_words; i++){
|
||
|
|
if (freq_words[i] == 0){
|
||
|
|
uni_word++;
|
||
|
|
freq_words[i] = uni_word;
|
||
|
|
for (int l = 0; l < nb_words; l++){
|
||
|
|
int k = 0;
|
||
|
|
while ( words[i][k] == words[l][k] ){
|
||
|
|
k += 1;
|
||
|
|
}
|
||
|
|
if (k == longest){
|
||
|
|
freq_words[l] = uni_word;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
// Init node
|
||
|
|
Node nodes[uni_word];
|
||
|
|
int c = 0;
|
||
|
|
for (i = 0; i < uni_word; i++){
|
||
|
|
nodes[uni_word].weights[uni_word];
|
||
|
|
memset(nodes[uni_word].weights, 0, uni_word * sizeof(int));
|
||
|
|
}
|
||
|
|
for (i = 0; i < nb_words; i++){
|
||
|
|
nodes[freq_words[i]].content = words[i];
|
||
|
|
}
|
||
|
|
|
||
|
|
return 0;
|
||
|
|
}
|