Find a substring within a string using processes - c

Question

I have a large file (around 1,000,000 characters) in the format "AATACGTAGCTA" and a subsequent file, such as "CGTATC" (10,240 characters). I want to find the largest match of the subsequence within the main sequence. A full, 100% subsequence match may not exist, this is not guaranteed. For the sake of a smaller example, the above would output: Longest match is 4/6 starting at position 5.
I'm working on my C basics, and would like to implement it like so:
The user chooses how many processes they would like to split the work
into.
Each process does 1/nth of the work and updates the shared memory
values located in the struct.
The longest match (it may not be all characters) is reflected in the
struct, as well as it's starting position, and how many
characters were matched. See output below.
Code
#define _GNU_SOURCE
#include <limits.h>
#include <stdio.h>
#include <errno.h>
#include <semaphore.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/wait.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <sys/types.h>
#include <sys/shm.h>
typedef struct memoryNeeded {
int start_pos, total_correct;
char sequence[1038336];
char subsequence[10240];
sem_t *sem;
} memoryNeeded;
// Used to check all arguments for validity
int checkArguments(char* p, int argc) {
char *prcs;
errno = 0;
int num;
long conv = strtol(p, &prcs, 10);
if (errno!= 0 || *prcs != '\0' || conv > INT_MAX || conv > 50) {
puts("Please input a valid integer for number of processes. (1-50)");
exit(1);
} else {
num = conv;
if (argc != 4) {
puts("\nPlease input the correct amount of command line arguments (4) in"
"the format: \n./DNA (processes) (sequence) (subsequence)\n");
exit(1);
} else
printf("Looking for string using %d processes...\n", num);
return(num);
}
}
int main (int argc, char* argv[]) {
int processes = checkArguments(argv[1], argc);
key_t shmkey;
int procNumber, shmid, pid;
FILE *sequence;
FILE *subsequence;
char *buf1, *buf2;
// Create shared memory
size_t region_size = sizeof(memoryNeeded);
shmkey = ftok("ckozeny", 5);
shmid = shmget(shmkey, region_size, 0644 | IPC_CREAT);
if (shmid < 0) {
perror("shmget\n");
exit(1);
}
// Create structure in shared memory, attach memory and open semaphore
memoryNeeded *mn;
mn = (memoryNeeded *)shmat(shmid, NULL, 0);
mn->sem = sem_open("sem", O_CREAT | O_EXCL, 0644, 1);
sequence = fopen(argv[2], "r");
subsequence = fopen(argv[3], "r");
// Get file sizes
fseek(sequence, 0L, SEEK_END);
int sz1 = ftell(sequence);
rewind(sequence);
fseek(subsequence, 0L, SEEK_END);
int sz2 = ftell(subsequence);
rewind(subsequence);
// Read files into 2 buffers, which are put into struct mn
buf1 = malloc(sz1);
buf2 = malloc(sz2);
if (sz1 != fread(buf1, 1, sz1, sequence)) {
free(buf1);
}
if (sz2 != fread(buf2, 1, sz2, subsequence)) {
free(buf2);
}
// Initialize struct with necessary values
mn->start_pos = 0;
mn->total_correct = 0;
strncpy(mn->sequence, buf1, sz1);
strncpy(mn->subsequence, buf2, sz2);
fclose(sequence);
fclose(subsequence);
// Begin n forks
for (procNumber = 0; procNumber < processes; procNumber++) {
pid = fork();
if (pid < 0) {
sem_unlink("sem");
sem_close(mn->sem);
printf ("Fork error.\n");
} else if (pid == 0)
break;
}
if (pid != 0) {
while ((pid = waitpid (-1, NULL, 0))){
if (errno == ECHILD)
break;
}
printf("Best match is at position %d with %d/10240 correct.", mn->start_pos, mn->total_correct);
printf ("\nParent: All children have exited.\n");
sem_unlink("sem");
sem_close(mn->sem);
shmdt(mn);
shmctl(shmid, IPC_RMID, 0);
exit(0);
} else {
// this child process will do its 1/nth of the work
sem_wait(mn->sem);
printf ("Child(%d) is in critical section.\n", procNumber);
sleep(1);
int i = 0;
int longest, count = 0;
for (i = 0; i < sz1; i += processes) {
for (int j = 0; j < sz2; j += processes) {
count = 0;
while (mn->sequence[i+j] == mn->subsequence[j]) {
count++;
j++;
}
if (count > longest) {
longest = count;
}
}
}
// If local match is longer than that of the struct, update and unlock
if (longest > mn->total_correct) {
mn->total_correct = count;
mn->start_pos = (i - count);
sem_post(mn->sem);
} else
// If not - unlock and let next process go
sem_post(mn->sem);
exit(0);
}
return 1;
}
The current child code is more or less "pseudocode". I've put it together how it makes sense in my head. (I'm aware this may not be correct or function as intended.) My question is in regard to the child code algorithm near the bottom.
How do I implement this so each child does 1/nth of the work, and finds the longest match, even though it may not match 100%?
Final output would be:
./DNA 6 sequence1 subsequence1
Looking for string using 6 processes...
Best match is at position 123456 with 9876/10240 correct.
Thanks.