/*
* Reads a read header (including variable length components) from an mFILE.
*
* Returns the a pointer to the header on success
* NULL on failure
*/
sff_read_header *read_sff_read_header(mFILE *mf) {
sff_read_header *h;
unsigned char rhdr[16];
if (16 != mfread(rhdr, 1, 16, mf))
return NULL;
h = decode_sff_read_header(rhdr);
if (h->name_len != mfread(h->name, 1, h->name_len, mf))
return free_sff_read_header(h), NULL;
/* Pad to 8 chars */
mfseek(mf, (mftell(mf) + 7)& ~7, SEEK_SET);
return h;
}
/*
* Override the sff.c functions to use FILE pointers instead. This means
* we don't have to load the entire archive into memory, which is optimal when
* dealing with a single file (ie in sff/sff.c), but not when indexing it.
*
* Done with minimal error checking I'll admit...
*/
static sff_read_header *fread_sff_read_header(FILE *fp) {
sff_read_header *h;
unsigned char rhdr[16];
if (16 != fread(rhdr, 1, 16, fp))
return NULL;
h = decode_sff_read_header(rhdr);
if (h->name_len != fread(h->name, 1, h->name_len, fp))
return free_sff_read_header(h), NULL;
/* Pad to 8 chars */
fseek(fp, (ftell(fp) + 7)& ~7, SEEK_SET);
return h;
}
/*
* Unpacks the 16-byte fixed size part of the SFF read header.
* It allocates memory for this and for the base calls, but does not
* unpack these.
*
* Returns sff_read_header* on success
* NULL on failure
*/
sff_read_header *decode_sff_read_header(unsigned char *buf) {
sff_read_header *h;
if (NULL == (h = (sff_read_header *)xcalloc(1, sizeof(*h))))
return NULL;
h->header_len = be_int2(*(uint16_t *)(buf+0));
h->name_len = be_int2(*(uint16_t *)(buf+2));
h->nbases = be_int4(*(uint32_t *)(buf+4));
h->clip_qual_left = be_int2(*(uint16_t *)(buf+8));
h->clip_qual_right = be_int2(*(uint16_t *)(buf+10));
h->clip_adapter_left = be_int2(*(uint16_t *)(buf+12));
h->clip_adapter_right = be_int2(*(uint16_t *)(buf+14));
if (NULL == (h->name = (char *)xmalloc(h->name_len)))
return free_sff_read_header(h), NULL;
return h;
}
/* F U N C T I O N S *********************************************************/
void process_sff_to_fastq(char *sff_file, int trim_flag) {
sff_read_header rh;
sff_read_data rd;
FILE *sff_fp, *fastq_fp;
if ( (sff_fp = fopen(sff_file, "r")) == NULL ) {
fprintf(stderr,
"[err] Could not open file '%s' for reading.\n", sff_file);
exit(1);
}
get_sff_file_size(sff_fp);
read_sff_common_header(sff_fp, &h);
verify_sff_common_header(&h);
if ( keep_fastq_orig == true ) {
vector<string> tmp_rep;
split_str(string(sff_file), tmp_rep, "//");
if ( ( fastq_fp = fopen( (tmp_rep[tmp_rep.size() - 1].substr(0,tmp_rep[tmp_rep.size() - 1].length()-4) + ".fastq").c_str(), "w") ) == NULL ) {
fprintf(stderr,
"[err] Could not open file '%s' for writing.\n",
"");
exit(1);
}
}
int left_clip = 0, right_clip = 0, nbases = 0;
char *name;
char *bases;
uint8_t *quality;
//register int i;
unsigned int numreads = h.nreads;
for (int i = 0; i < numreads; i++) { //cout << i << " " << numreads << endl;
read_sff_read_header(sff_fp, &rh);
read_sff_read_data(sff_fp, &rd, h.flow_len, rh.nbases);
//rheaders.push_back(rh);
// get clipping points
get_clip_values(rh, trim_flag, &left_clip, &right_clip);
nbases = right_clip - left_clip;
// create bases string
bases = get_read_bases(rd, left_clip, right_clip);
// create quality array
quality = get_read_quality_values(rd, left_clip, right_clip);
//Create new read
Read *read = new Read();
read->initial_length = nbases;
read->read = string(bases);
uint8_t quality_char;
read->quality = (uint8_t*)malloc(sizeof(uint8_t)*nbases);
for (int j = 0; j < nbases; j++)
{
quality_char = (quality[j] <= 93 ? quality[j] : 93) + 33;
read->quality[j] = quality_char;
}
//read->rd = rd;
read->flowgram = new uint16_t[h.flow_len];
for(int j=0; j<h.flow_len; j++) {
read->flowgram[j] = rd.flowgram[j];
//cout << rd.flowgram[j] << " " << endl;
}
read->flow_index = (uint8_t*)malloc(sizeof(uint8_t)*nbases);
for(int j=0; j<nbases; j++) {
read->flow_index[j] = rd.flow_index[j];
}
read->roche_left_clip = (int) max(1, max(rh.clip_qual_left, rh.clip_adapter_left)) - 1;
read->roche_right_clip = (int) min( (rh.clip_qual_right == 0 ? rh.nbases : rh.clip_qual_right ), (rh.clip_adapter_right == 0 ? rh.nbases : rh.clip_adapter_right) );
reads.push_back(read);
string tstr = string(rh.name) + " " + string(itoa(rh.clip_adapter_left,new char[5],10)) + " " + string(itoa(rh.clip_adapter_right,new char[5],10))+ " " + string(itoa(rh.clip_qual_left,new char[5],10)) + " " + string(itoa(rh.clip_qual_right,new char[5],10)) + " " + string(itoa(rh.clip_qual_right,new char[5],10));
int t_len = tstr.length();
// create read name string
int name_length = (int) t_len + 1; // account for NULL termination
name = (char *) malloc( name_length * sizeof(char) );
if (!name) {
fprintf(stderr, "Out of memory! For read name string!\n");
exit(1);
}
memset(name, '\0', (size_t) name_length);
read->readID = (char *) malloc( rh.name_len * sizeof(char) );
//read->readID = rh.name;
memcpy( read->readID, rh.name, (size_t) rh.name_len );
//strncpy(name, rh.name, (size_t) rh.name_len);
strncpy(name, tstr.c_str(), (size_t)t_len);
if ( keep_fastq_orig == true )
construct_fastq_entry(fastq_fp, name, bases, quality, nbases);
//printf("%d\n",rh.name_len);
free(name);
free(bases);
free(quality);
free_sff_read_header(&rh);
free_sff_read_data(&rd);
}
read_manifest(sff_fp);
//free_sff_common_header(&h);
if ( keep_fastq_orig == true )
fclose(fastq_fp);
fclose(sff_fp);
}
/*
* Reads an SFF file from an mFILE and decodes it to a Read struct.
*
* Returns Read* on success
* NULL on failure
*/
Read *mfread_sff(mFILE *mf) {
int i, bpos;
Read *r;
sff_common_header *ch;
sff_read_header *rh;
sff_read_data *rd;
/* Load the SFF contents */
if (NULL == (ch = read_sff_common_header(mf)))
return NULL;
if (NULL == (rh = read_sff_read_header(mf))) {
free_sff_common_header(ch);
return NULL;
}
if (NULL == (rd = read_sff_read_data(mf, ch->flow_len, rh->nbases))) {
free_sff_common_header(ch);
free_sff_read_header(rh);
return NULL;
}
/* Convert to Read struct */
r = read_allocate(0,0);
if (r->basePos) free(r->basePos);
if (r->base) free(r->base);
if (r->prob_A) free(r->prob_A);
if (r->prob_C) free(r->prob_C);
if (r->prob_G) free(r->prob_G);
if (r->prob_T) free(r->prob_T);
r->nflows = ch->flow_len;
r->flow_order = ch->flow; ch->flow = NULL;
r->flow_raw = NULL;
r->flow = (float *)malloc(r->nflows * sizeof(float));
for (i = 0; i < r->nflows; i++) {
r->flow[i] = rd->flowgram[i] / 100.0;
}
r->NBases = rh->nbases;
r->basePos = (uint_2 *)calloc(r->NBases, 2);
r->base = rd->bases; rd->bases = NULL;
r->prob_A = (char *)calloc(r->NBases, 1);
r->prob_C = (char *)calloc(r->NBases, 1);
r->prob_G = (char *)calloc(r->NBases, 1);
r->prob_T = (char *)calloc(r->NBases, 1);
bpos = 0;
for (i=0; i < r->NBases; i++) {
r->prob_A[i] = 0;
r->prob_C[i] = 0;
r->prob_G[i] = 0;
r->prob_T[i] = 0;
switch (r->base[i]) {
case 'A':
case 'a':
r->prob_A[i] = rd->quality[i];
break;
case 'C':
case 'c':
r->prob_C[i] = rd->quality[i];
break;
case 'G':
case 'g':
r->prob_G[i] = rd->quality[i];
break;
case 'T':
case 't':
r->prob_T[i] = rd->quality[i];
break;
}
bpos += rd->flow_index[i];
r->basePos[i] = bpos;
}
r->leftCutoff = MAX(rh->clip_qual_left, rh->clip_adapter_left);
r->rightCutoff = MIN(rh->clip_qual_right
? rh->clip_qual_right
: r->NBases+1,
rh->clip_adapter_right
? rh->clip_adapter_right
: r->NBases+1);
free_sff_common_header(ch);
free_sff_read_header(rh);
free_sff_read_data(rd);
return r;
}
void
process_sff_to_fastq(char *sff_file, char *fastq_file, int trim_flag) {
sff_common_header h;
sff_read_header rh;
sff_read_data rd;
FILE *sff_fp, *fastq_fp;
if ( !strlen(sff_file) ) {
sff_fp = stdin;
}
else if ( (sff_fp = fopen(sff_file, "r")) == NULL ) {
fprintf(stderr,
"[err] Could not open file '%s' for reading.\n", sff_file);
exit(1);
}
read_sff_common_header(sff_fp, &h);
verify_sff_common_header(PRG_NAME, VERSION, &h);
// printf("size of header: %d \n", sizeof(sff_common_header));
// printf("\tmagic : 0x%x\n" , h.magic);
// printf("\tindex_offset : 0x%llx\n", h.index_offset);
// printf("\tindex_len : 0x%x\n" , h.index_len);
// printf("\tnumreads : 0x%x\n" , h.nreads);
// printf("\theader_len : 0x%x\n" , h.header_len);
// printf("\tkey_len : 0x%x\n" , h.key_len);
// printf("\tflow_len : 0x%x\n" , h.flow_len);
// printf("\tflowgram_fmt : 0x%x\n" , h.flowgram_format);
// printf("\tflow : %s\n " , h.flow);
// printf("\tkey : %s\n " , h.key);
// printf("\n\n");
if ( !strlen(fastq_file) ) {
fastq_fp = stdout;
}
else {
if ( (fastq_fp = fopen(fastq_file, "w")) == NULL ) {
fprintf(stderr,
"[err] Could not open file '%s' for writing.\n",
fastq_file);
exit(1);
}
}
int left_clip = 0, right_clip = 0, nbases = 0;
char *name;
char *bases;
uint8_t *quality;
register int i;
int numreads = (int) h.nreads;
for (i = 0; i < numreads; i++) {
read_sff_read_header(sff_fp, &rh);
read_sff_read_data(sff_fp, &rd, h.flow_len, rh.nbases);
/* get clipping points */
get_clip_values(rh, trim_flag, &left_clip, &right_clip);
nbases = right_clip - left_clip;
/* create bases string */
bases = get_read_bases(rd, left_clip, right_clip);
/* create quality array */
quality = get_read_quality_values(rd, left_clip, right_clip);
/* create read name string */
int name_length = (int) rh.name_len + 1; // account for NULL termination
name = (char *) malloc( name_length * sizeof(char) );
if (!name) {
fprintf(stderr, "Out of memory! For read name string!\n");
exit(1);
}
memset(name, '\0', (size_t) name_length);
strncpy(name, rh.name, (size_t) rh.name_len);
construct_fastq_entry(fastq_fp, name, bases, quality, nbases);
free(name);
free(bases);
free(quality);
free_sff_read_header(&rh);
free_sff_read_data(&rd);
}
free_sff_common_header(&h);
fclose(fastq_fp);
fclose(sff_fp);
}