Kaehvaman/OAIP_Mirror
1
0
Fork 0
mirror of https://github.com/Kaehvaman/OAIP.git synced 2025-01-18 16:49:11 +04:00
Code Issues Actions Packages Projects Releases Wiki Activity
f2428d8bb6
OAIP_Mirror/lab26/TextProcessingDict/TextProcessing.c

237 lines
5.7 KiB
C
Raw Normal View History

autotest lab26
2024-12-06 18:45:55 +04:00
#define _CRT_SECURE_NO_WARNINGS
#include <stdio.h>
#include <string.h>
#include <time.h>
#include <Windows.h>
#include "Dict.h"
int getNextDelim(FILE* fp, char token[]);
int getNextWord(FILE* fp, char token[], int maxLen);
int LoadDictionary(char* filename);
int TextProcessing(char* filenameIn, char* filenameOut);
char filenameDict[MAX_PATH] = "../Dictionaries/dict0.txt";
char filenameIn[MAX_PATH] = "../Texts/Alice.txt";
char filenameOut[MAX_PATH] = "out/Alice_out.html";
finish open addressing
2024-12-06 23:11:06 +04:00
double results[3][15];
autotest lab26
2024-12-06 18:45:55 +04:00
void test(int i, int j) {
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
printf("HTML = %s\ntext = %s\ndict = %s\n",
filenameOut, filenameIn, filenameDict);
LoadDictionary(filenameDict);
clock_t t0 = clock();
TextProcessing(filenameIn, filenameOut);
clock_t t1 = clock();
Destroy();
double runtime = (t1 - t0) / (double)CLOCKS_PER_SEC;
results[i][j] = runtime;
printf("t1 - t0 = %.3f sec (Run time of HTML generating)\n\n", runtime);
}
void test_dicts(int i) {
finish open addressing
2024-12-06 23:11:06 +04:00
for (int j = 0; j < 5; j++) {
autotest lab26
2024-12-06 18:45:55 +04:00
sprintf(filenameDict, "../Dictionaries/dict%d.txt", j);
test(i, j);
sprintf(filenameDict, "../Dictionaries/dict%da.txt", j);
finish open addressing
2024-12-06 23:11:06 +04:00
test(i, j + 5);
autotest lab26
2024-12-06 18:45:55 +04:00
sprintf(filenameDict, "../Dictionaries/dict%db.txt", j);
finish open addressing
2024-12-06 23:11:06 +04:00
test(i, j + 10);
autotest lab26
2024-12-06 18:45:55 +04:00
}
}
int main() {
strcpy(filenameIn, "../Texts/Alice.txt");
strcpy(filenameOut, "out/Alice_out.html");
test_dicts(0);
strcpy(filenameIn, "../Texts/Tolkien.txt");
strcpy(filenameOut, "out/Tolkien_out.html");
test_dicts(1);
strcpy(filenameIn, "../Texts/Tolkien2.txt");
strcpy(filenameOut, "out/Tolkien2_out.html");
test_dicts(2);
for (int i = 0; i < 3; i++) {
finish open addressing
2024-12-06 23:11:06 +04:00
for (int j = 0; j < 15; j++) {
autotest lab26
2024-12-06 18:45:55 +04:00
printf("%7.3lf ", results[i][j]);
}
printf("\n");
}
printf("\a");
return 0;
}
int LoadDictionary(char* filename) {
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD>
FILE* fin = fopen(filename, "r");
if (fin == NULL) {
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD>
printf("File %s didn't open!\n", filename);
return 0;
}
Create();
char token[MAX_LEN_WORD];
// <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD>
while (!feof(fin)) {
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD>
while (getNextDelim(fin, token)) {
}
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD>
if (getNextWord(fin, token, MAX_LEN_WORD)) {
Insert(token);
}
}
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
fclose(fin);
return 1;
}
int TextProcessing(char* filenameIn, char* filenameOut) {
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
FILE* fin = fopen(filenameIn, "r");
if (fin == NULL) {
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD>
printf("File %s doesn't opened!\n", filenameIn);
return 0;
}
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
FILE* fout = fopen(filenameOut, "w");
if (fout == NULL) {
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD>
printf("File %s doesn't opened!\n", filenameOut);
// <20> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD>
fclose(fin);
return 0;
}
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> HTML <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
fprintf(fout, "<!DOCTYPE html>");
fprintf(fout, "<html>");
fprintf(fout, "<head>");
finish tests of lab26
2024-12-06 20:08:12 +04:00
fprintf(fout, "<meta http-equiv = \"Content-Type\" content = \"text/html; charset=cp1251\" />");
autotest lab26
2024-12-06 18:45:55 +04:00
fprintf(fout, "<title>HTML Document</title>");
fprintf(fout, "</head>");
fprintf(fout, "<body>");
char token[MAX_LEN_WORD];
// <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD>
while (!feof(fin)) {
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD>
while (getNextDelim(fin, token)) {
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
if (strcmp(token, "<") == 0) {
fprintf(fout, "&lt");
}
else if (strcmp(token, ">") == 0) {
fprintf(fout, "&gt");
}
else {
if (strcmp(token, "\n") == 0) {
fprintf(fout, "<br>");
}
fprintf(fout, "%s", token);
}
}
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD>
if (getNextWord(fin, token, MAX_LEN_WORD)) {
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20> <20><> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD>
if (Member(token)) {
fprintf(fout, "<b>%s</b>", token);
}
else {
fprintf(fout, "%s", token);
}
}
}
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20> HTML <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> HTML
fprintf(fout, "</body>");
fprintf(fout, "</html>");
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD>
fclose(fin);
fclose(fout);
return 1;
}
int isalpha_my(unsigned char ch);
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> 1 - <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>.
// <20> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20> token <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD>, <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
// <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>.
// <20><><EFBFBD><EFBFBD> <20> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> 0.
// <20> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> token <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>.
int getNextDelim(FILE* fp, char token[])
{
int ch = getc(fp);
if (ch == EOF) {
return 0;
}
if (isalpha_my((unsigned char)ch)) {
ungetc(ch, fp);
return 0;
}
token[0] = (unsigned char)ch;
token[1] = '\0';
return 1;
}
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> 1 - <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD>.
// <20> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20> token <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD>, <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>
// <20><><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD>. <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> maxLen <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>.
// <20><><EFBFBD><EFBFBD> <20> <20><><EFBFBD><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD> - <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> 0.
// <20> <20><><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> token <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD>.
int getNextWord(FILE* fp, char token[], int maxLen)
{
int i = 0;
int ch;
while (((ch = getc(fp)) != EOF) && (i < maxLen - 1)) {
if (!isalpha_my((unsigned char)(ch))) {
break;
}
token[i++] = ch;
}
ungetc(ch, fp);
token[i] = '\0';
if (i == 0)
return 0;
return 1;
}
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> 0 - <20><><EFBFBD><EFBFBD> ch - <20><> <20><><EFBFBD><EFBFBD><EFBFBD>.
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> 1 - <20><><EFBFBD><EFBFBD> ch - <20><><EFBFBD><EFBFBD><EFBFBD>.
// <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> (<28> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD> < 128)
// <20> <20><><EFBFBD> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> <20><><EFBFBD><EFBFBD> <20><> <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD> ANSI
int isalpha_my(unsigned char ch) {
if (isalpha(ch))
return 1;
// ANSI <20><><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD><EFBFBD>!!!
if (ch >= 192 && ch <= 223)
return 1;
if (ch >= 224 && ch <= 255)
return 1;
/*
if (ch >= '<EFBFBD>' && ch <= '<EFBFBD>') return 1;
if (ch >= '<EFBFBD>' && ch <= '<EFBFBD>') return 1;
if (ch >= '<EFBFBD>' && ch <= '<EFBFBD>')return 1;
if (ch == '<EFBFBD>' ) return 1;
if (ch == '<EFBFBD>') return 1;*/
return 0;
}
Reference in New Issue Copy Permalink