/**
* Reads a block of files into memory.
* @param strs Array for data
* @param len Length of block
* @return number of lines read into memory
*/
int input_lines_read(hstring_t *strs, int len)
{
assert(strs && len > 0);
int read, i = 0, j = 0;
size_t size;
char buf[32], *line = NULL;
for (i = 0; i < len; i++) {
line = NULL;
read = gzgetline(&line, &size, in);
if (read == -1) {
free(line);
break;
}
/* Strip newline characters */
strip_newline(line, read);
/* Caution: May modify the line */
strs[j].label = get_label(line);
strs[j].str.c = line;
strs[j].type = TYPE_CHAR;
strs[j].len = strlen(line);
snprintf(buf, 32, "line%d", line_num++);
strs[j].src = strdup(buf);
j++;
}
return j;
}
/**
* Opens a file for reading text fasta.
* @param name File name
* @return number of fasta or -1 on error
*/
int input_fasta_open(char *name)
{
assert(name);
size_t read, size;
char *line = NULL;
const char *pattern;
/* Compile regular expression for label */
config_lookup_string(&cfg, "input.fasta_regex", &pattern);
if (regcomp(&re, pattern, REG_EXTENDED) != 0) {
error("Could not compile regex for label");
return -1;
}
in = gzopen(name, "r");
if (!in) {
error("Could not open '%s' for reading", name);
return -1;
}
int num, cont = FALSE;
while(!gzeof(in)) {
line = NULL;
read = gzgetline(&line, &size, in);
if (read > 0)
strtrim(line);
if (read > 1 && !cont && (line[0] == '>' || line[0] == ';')) {
num++;
cont = TRUE;
} else {
cont = FALSE;
}
free(line);
}
/* Prepare reading */
gzrewind(in);
return num;
}
/**
* Reads a block of files into memory.
* @param strs Array for data
* @param len Length of block
* @return number of lines read into memory
*/
int input_lines_read(string_t *strs, int len)
{
assert(strs && len > 0);
int read, i = 0, j = 0;
size_t size;
char buf[32], *line = NULL;
for (i = 0; i < len; i++) {
#ifdef ENABLE_EVALTIME
double t1 = time_stamp();
#endif
line = NULL;
read = gzgetline(&line, &size, in);
if (read == -1) {
free(line);
break;
}
/* Strip newline characters */
strip_newline(line, read);
strs[j].label = get_label(line);
strs[j].str = line;
strs[j].len = strlen(line);
snprintf(buf, 32, "line%d", line_num++);
strs[j].src = strdup(buf);
j++;
#ifdef ENABLE_EVALTIME
printf("strlen %d read %f\n", strs[j-1].len, time_stamp() - t1);
#endif
}
return j;
}
/**
* Reads a block of files into memory.
* @param strs Array for data
* @param len Length of block
* @return number of read files
*/
int input_fasta_read(string_t *strs, int len)
{
assert(strs && len > 0);
int read, i = 0, alloc = -1;
size_t size;
char *line = NULL, *seq = NULL;
while (i < len) {
/* Read line */
if (old_line) {
line = old_line;
read = strlen(line) + 1;
} else {
line = NULL;
read = gzgetline(&line, &size, in);
}
old_line = NULL;
/* Trim line */
if (read >= 0)
strtrim(line);
/* End of sequence */
if (alloc > 1 && (read == -1 || line[0] == ';' || line[0] == '>')) {
strs[i].str = seq;
strs[i].len = alloc - 1;
i++;
}
/* Stop on read error */
if (read == -1) {
free(line);
break;
}
/* Reset pointer for next chunk */
if (i == len) {
/* Save old line */
old_line = line;
#if 0
/* Alternative code with slow gzseek */
gzseek(in, -read, SEEK_CUR);
free(line);
#endif
break;
}
/* Check for comment char */
if (line[0] == ';' || line[0] == '>') {
/* Start of sequence */
if (alloc == -1 || alloc > 1) {
strs[i].src = strdup(line);
strs[i].label = get_label(line);
seq = calloc(sizeof(char), 1);
alloc = 1;
}
goto skip;
}
/* Skip text before first comment */
if (alloc == -1)
goto skip;
/* Append line to sequence */
alloc += strlen(line);
seq = realloc(seq, alloc * sizeof(char));
strncat(seq, line, strlen(line));
skip:
free(line);
}
return i;
}
struct matrix *load_sensing_matrix(const char *filename, unsigned int target_kmer) {
char *line = NULL;
char **headers = NULL;
double *matrix = NULL;
unsigned int kmer = 0;
unsigned long long i = 0;
unsigned long long *row = NULL;
unsigned long long sequences = 0;
unsigned long long width = 0;
struct matrix *ret = NULL;
size_t lineno = 0;
gzFile fh = NULL;
fh = gzopen(filename, "r");
if(fh == NULL) {
fprintf(stderr, "could not open %s", filename);
exit(EXIT_FAILURE);
}
line = malloc(1024 * sizeof(char));
check_malloc(line, NULL);
// Check for quikr
line = gzgets(fh, line, 1024);
lineno++;
if(strcmp(line, "quikr\n") != 0) {
fprintf(stderr, "This does not look like a quikr sensing matrix. Please check your path: %s\n", filename);
exit(EXIT_FAILURE);
}
// check version
line = gzgets(fh, line, 1024);
if(atoi(line) != MATRIX_REVISION) {
fprintf(stderr, "Sensing Matrix uses an unsupported version, please retrain your matrix\n");
exit(EXIT_FAILURE);
}
lineno++;
// get number of sequences
line = gzgets(fh, line, 1024);
sequences = strtoull(line, NULL, 10);
if(sequences == 0) {
fprintf(stderr, "Error parsing sensing matrix, sequence count is zero\n");
exit(EXIT_FAILURE);
}
lineno++;
// get kmer
gzgets(fh, line, 1024);
kmer = atoi(line);
if(kmer == 0) {
fprintf(stderr, "Error parsing sensing matrix, kmer is zero\n");
exit(EXIT_FAILURE);
}
lineno++;
if(kmer != target_kmer) {
fprintf(stderr, "The sensing_matrix was trained with a different kmer than your requested kmer\n");
exit(EXIT_FAILURE);
}
width = pow_four(kmer);
// allocate a +1 size for the extra row
matrix = malloc(sequences * (width) * sizeof(double));
check_malloc(matrix, NULL);
row = malloc((width) * sizeof(unsigned long long));
check_malloc(row, NULL);
headers = malloc(sequences * sizeof(char *));
check_malloc(headers, NULL);
char *buf = NULL;
size_t len = 0;
size_t read = 0;
for(i = 0; i < sequences; i++) {
unsigned long long j = 0;
// get header and add it to headers array
//
read = gzgetline(&buf, &len, fh);
if(read == 0) {
fprintf(stderr, "Error parsing sensing matrix, could not read header\n");
exit(EXIT_FAILURE);
}
char *header = malloc(sizeof(char) * read + 1);
check_malloc(header, NULL);
header = strncpy(header, buf, read - 1);
if(header[0] != '>') {
fprintf(stderr, "Error parsing sensing matrix, could not read header in line %llu\n", lineno);
exit(EXIT_FAILURE);
}
lineno++;
header[read - 1] = '\0';
headers[i] = header+1;
row = memset(row, 0, (width) * sizeof(unsigned long long));
for(j = 0; j < width; j++) {
line = gzgets(fh, line, 32);
lineno++;
if(line == NULL || line[0] == '>') {
fprintf(stderr, "Error parsing sensing matrix, line %zu does not look like a value\n", lineno);
exit(EXIT_FAILURE);
}
lineno++;
row[j] = strtoull(line, NULL, 10);
if(errno) {
printf("could not parse '%s'\n into a number", line);
exit(EXIT_FAILURE);
}
}
for(j = 0; j < width; j++) {
matrix[i*(width) + j] = ((double)row[j]);
}
}
// load the matrix of counts
gzclose(fh);
free(line);
free(row);
ret = malloc(sizeof(struct matrix));
(*ret).kmer = kmer;
(*ret).sequences = sequences;
(*ret).matrix = matrix;
(*ret).headers = headers;
return ret;
}
