void
wells_data_print(FILE *fp, wells_header_t *header, wells_data_t *data, int32_t nonzero)
{
int32_t i, output;
output = 1;
if(1 == nonzero) {
output = 0;
for(i=0;i<header->num_flows;i++) {
if(0 < data->flow_values[i]) {
output = 1;
break;
}
}
}
if(1 == output) {
if(fprintf(fp, "%d,%d,%d", data->rank, data->x, data->y) < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
for(i=0;i<header->num_flows;i++) {
if(fprintf(fp, ",%f", data->flow_values[i]) < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
}
if(fprintf(fp, "\n") < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
}
}
uint32_t
sff_read_header_write(FILE *fp, sff_read_header_t *rh)
{
uint32_t n = 0;
sff_read_header_calc_bytes(rh);
// convert values to big-endian
sff_read_header_hton(rh);
if(1 != fwrite(&rh->rheader_length, sizeof(uint16_t), 1, fp)
|| 1 != fwrite(&rh->name_length, sizeof(uint16_t), 1, fp)
|| 1 != fwrite(&rh->n_bases, sizeof(uint32_t), 1, fp)
|| 1 != fwrite(&rh->clip_qual_left, sizeof(uint16_t), 1, fp)
|| 1 != fwrite(&rh->clip_qual_right, sizeof(uint16_t), 1, fp)
|| 1 != fwrite(&rh->clip_adapter_left, sizeof(uint16_t), 1, fp)
|| 1 != fwrite(&rh->clip_adapter_right, sizeof(uint16_t), 1, fp)) {
ion_error(__func__, "fwrite", Exit, WriteFileError);
}
n += sizeof(uint32_t) + 6*sizeof(uint16_t);
// convert values from big-endian
sff_read_header_ntoh(rh);
if(rh->name_length != fwrite(rh->name->s, sizeof(char), rh->name_length, fp)) {
ion_error(__func__, "fwrite", Exit, ReadFileError);
}
n += sizeof(char)*rh->name_length;
n += ion_write_padding(fp, n);
return n;
}
void
sff_sort(sff_file_t *fp_in, sff_file_t *fp_out)
{
int32_t i, row, col;
sff_t *sff;
int32_t requires_sort = 0;
sff_sort_t *sffs = NULL;
int32_t sffs_mem = 0, sffs_len = 0;
// initialize memory
sffs_mem = 1024;
sffs = ion_malloc(sizeof(sff_sort_t) * sffs_mem, __func__, "sffs");
// go through the input file
while(NULL != (sff = sff_read(fp_in))) {
// get the row/col co-ordinates
if(0 == ion_readname_to_rowcol(sff->rheader->name->s, &row, &col)) {
ion_error(__func__, "could not understand the read name", Exit, OutOfRange);
}
// copy over
while(sffs_mem <= sffs_len) {
sffs_mem <<= 1; // double
sffs = ion_realloc(sffs, sizeof(sff_sort_t) * sffs_mem, __func__, "sffs");
}
sffs[sffs_len].row = row;
sffs[sffs_len].col = col;
sffs[sffs_len].sff = sff;
sff = NULL;
// check if we need to sort, for later
if(0 < sffs_len && __sff_sort_lt(sffs[sffs_len], sffs[sffs_len-1])) {
requires_sort = 1;
}
sffs_len++;
}
// resize
sffs_mem = sffs_len;
sffs = ion_realloc(sffs, sizeof(sff_sort_t) * sffs_mem, __func__, "sffs");
if(1 == requires_sort) {
// sort
ion_sort_introsort(sff_sort, sffs_len, sffs);
}
// write
for(i=0;i<sffs_len;i++) {
if(0 == sff_write(fp_out, sffs[i].sff)) {
ion_error(__func__, "sff_write", Exit, WriteFileError);
}
}
// destroy
for(i=0;i<sffs_len;i++) {
sff_destroy(sffs[i].sff);
}
free(sffs);
}
uint32_t
sff_index_write(FILE *fp, sff_index_t *idx)
{
int32_t i;
uint32_t n = 0;
uint64_t len = 0;
if(NULL == idx) {
return 0;
}
// convert values to big-endian
sff_index_hton(idx);
// index header
if(1 != fwrite(&idx->index_magic_number, sizeof(uint32_t), 1, fp)
|| 1 != fwrite(&idx->index_version, sizeof(uint32_t), 1, fp)
|| 1 != fwrite(&idx->num_rows, sizeof(int32_t), 1, fp)
|| 1 != fwrite(&idx->num_cols, sizeof(int32_t), 1, fp)
|| 1 != fwrite(&idx->type, sizeof(int32_t), 1, fp)) {
ion_error(__func__, "fwrite", Exit, WriteFileError);
}
n += sizeof(uint32_t)*2 + sizeof(int32_t)*3;
// convert values from big-endian
sff_index_ntoh(idx);
// offsets
if(SFF_INDEX_ROW_ONLY == idx->type) {
len = 1 + idx->num_rows;
}
else if(SFF_INDEX_ALL == idx->type) {
len = 1 + (idx->num_rows * idx->num_cols);
}
else {
ion_error(__func__, "could not understand index type", Exit, OutOfRange);
}
// convert values to big-endian
for(i=0;i<len;i++) {
idx->offset[i] = htonll(idx->offset[i]);
}
// write
if(len != fwrite(idx->offset, sizeof(uint64_t), len, fp)) {
ion_error(__func__, "fwrite", Exit, WriteFileError);
}
n += sizeof(uint64_t) * len;
// convert values from big-endian
for(i=0;i<len;i++) {
idx->offset[i] = ntohll(idx->offset[i]);
}
// padding
n += ion_write_padding(fp, n);
return n;
}
int
wells_mask_view_main(int argc, char *argv[])
{
FILE *fp_in;
int32_t i;
int c;
int32_t min_row, max_row, min_col, max_col;
wells_mask_t *mask = NULL;
min_row = max_row = min_col = max_col = -1;
while((c = getopt(argc, argv, "r:c:h")) >= 0) {
switch(c) {
case 'r':
if(ion_parse_range(optarg, &min_row, &max_row) < 0) {
ion_error(__func__, "-r : format not recognized", Exit, OutOfRange);
}
break;
case 'c':
if(ion_parse_range(optarg, &min_col, &max_col) < 0) {
ion_error(__func__, "-c : format not recognized", Exit, OutOfRange);
}
break;
case 'h':
default:
return usage();
}
}
if(argc == optind) {
return usage();
}
else {
for(i=optind;i<argc;i++) {
if(!(fp_in = fopen(argv[i], "rb"))) {
fprintf(stderr, "** Could not open %s for reading. **\n", argv[i]);
ion_error(__func__, argv[i], Exit, OpenFileError);
}
if(NULL == (mask = wells_mask_read(fp_in))) {
ion_error(__func__, argv[i], Exit, ReadFileError);
}
fclose(fp_in);
fp_in = NULL;
wells_mask_print(stdout, mask);
wells_mask_destroy(mask);
}
}
return 0;
}
sff_index_t *
sff_index_read(FILE *fp)
{
int32_t i;
uint32_t n = 0;
uint64_t len = 0;
sff_index_t *idx;
idx = sff_index_init();
// index header
if(1 != fread(&idx->index_magic_number, sizeof(uint32_t), 1, fp)
|| 1 != fread(&idx->index_version, sizeof(uint32_t), 1, fp)
|| 1 != fread(&idx->num_rows, sizeof(int32_t), 1, fp)
|| 1 != fread(&idx->num_cols, sizeof(int32_t), 1, fp)
|| 1 != fread(&idx->type, sizeof(int32_t), 1, fp)) {
ion_error(__func__, "fread", Exit, WriteFileError);
}
n += sizeof(uint32_t)*2 + sizeof(int32_t)*3;
// convert values from big-endian
sff_index_ntoh(idx);
// offsets
if(SFF_INDEX_ROW_ONLY == idx->type) {
len = 1 + idx->num_rows;
}
else if(SFF_INDEX_ALL == idx->type) {
len = 1 + (idx->num_rows * idx->num_cols);
}
else {
ion_error(__func__, "could not understand index type", Exit, OutOfRange);
}
// alloc
idx->offset = ion_malloc(sizeof(uint64_t) * len, __func__, "idx->offset");
// read
if(len != fread(idx->offset, sizeof(uint64_t), len, fp)) {
ion_error(__func__, "fread", Exit, WriteFileError);
}
// convert values from big-endian
for(i=0;i<len;i++) {
idx->offset[i] = ntohll(idx->offset[i]);
}
n += sizeof(uint64_t) * len;
// padding
n += ion_read_padding(fp, n);
return idx;
}
void
wells_mask_write(FILE *fp, wells_mask_t *mask)
{
int32_t i;
if(fwrite(&mask->num_rows, sizeof(int32_t), 1, fp) != 1
|| fwrite(&mask->num_cols, sizeof(int32_t), 1, fp) != 1) {
ion_error(__func__, "fwrite", Exit, WriteFileError);
}
for(i=0;i<mask->num_rows;i++) {
if(fwrite(mask->masks[i], sizeof(uint16_t), mask->num_cols, fp) != mask->num_cols) {
ion_error(__func__, "fwrite", Exit, WriteFileError);
}
}
}
uint32_t
sff_read_write(FILE *fp, sff_header_t *gh, sff_read_header_t *rh, sff_read_t *r)
{
uint32_t i, n = 0;
// convert flowgram to network order
for(i=0;i<gh->flow_length;i++) {
r->flowgram[i] = htons(r->flowgram[i]);
}
if(gh->flow_length != fwrite(r->flowgram, sizeof(uint16_t), gh->flow_length, fp)
|| rh->n_bases != fwrite(r->flow_index, sizeof(uint8_t), rh->n_bases, fp)
|| rh->n_bases != fwrite(r->bases->s, sizeof(char), rh->n_bases, fp)
|| rh->n_bases != fwrite(r->quality->s, sizeof(char), rh->n_bases, fp)) {
ion_error(__func__, "fread", Exit, ReadFileError);
}
n += sizeof(uint16_t)*gh->flow_length + 3*sizeof(uint8_t)*rh->n_bases;
// convert flowgram to host order
for(i=0;i<gh->flow_length;i++) {
r->flowgram[i] = ntohs(r->flowgram[i]);
}
n += ion_write_padding(fp, n);
return n;
}
int
dat_frame_write(FILE *fp, dat_frame_t *cur, dat_frame_t *prev, dat_header_t *header)
{
int32_t i;
uint32_t compressed=0;
// first frame or uninterlaced
if(NULL == prev || DAT_HEADER_UNINTERLACED == header->interlace_type) {
// timestamp and comprssion
compressed = 0;
if(fwrite_big_endian_uint32_t(fp, &cur->timestamp) != 1
|| fwrite_big_endian_uint32_t(fp, &compressed) != 1) {
return EOF;
}
// manually convert to big-endian
for(i=0;i<header->rows*header->cols;i++) {
cur->data[i] = htons(cur->data[i]);
}
// write the data
if(fwrite(cur->data, sizeof(uint16_t), header->rows*header->cols, fp) != header->rows*header->cols) { // image data
return EOF;
}
// manually convert from big-endian
for(i=0;i<header->rows*header->cols;i++) {
cur->data[i] = ntohs(cur->data[i]);
}
}
else {
// Currently not supported
// TODO
ion_error(__func__, "Writing interlaced DAT files currently not supported", Warn, OutOfRange);
return EOF;
}
return 1;
}
uint32_t
sff_index_length(sff_index_t *idx)
{
uint32_t n = 0;
uint64_t len = 0;
// index header
n += sizeof(uint32_t)*2 + sizeof(int32_t)*3;
// offsets
if(SFF_INDEX_ROW_ONLY == idx->type) {
len = 1 + idx->num_rows;
}
else if(SFF_INDEX_ALL == idx->type) {
len = 1 + (idx->num_rows * idx->num_cols);
}
else {
ion_error(__func__, "could not understand index type", Exit, OutOfRange);
}
n += sizeof(uint64_t) * len;
// padding
n += ion_write_padding(NULL, n);
return n;
}
sff_read_header_t *
sff_read_header_read(FILE *fp)
{
sff_read_header_t *rh = NULL;
uint32_t n = 0;
rh = sff_read_header_init();
if(1 != fread(&rh->rheader_length, sizeof(uint16_t), 1, fp)
|| 1 != fread(&rh->name_length, sizeof(uint16_t), 1, fp)
|| 1 != fread(&rh->n_bases, sizeof(uint32_t), 1, fp)
|| 1 != fread(&rh->clip_qual_left, sizeof(uint16_t), 1, fp)
|| 1 != fread(&rh->clip_qual_right, sizeof(uint16_t), 1, fp)
|| 1 != fread(&rh->clip_adapter_left, sizeof(uint16_t), 1, fp)
|| 1 != fread(&rh->clip_adapter_right, sizeof(uint16_t), 1, fp)) {
free(rh);
return NULL;
}
n += sizeof(uint32_t) + 6*sizeof(uint16_t);
// convert values from big-endian
sff_read_header_ntoh(rh);
rh->name = ion_string_init(rh->name_length+1);
if(rh->name_length != fread(rh->name->s, sizeof(char), rh->name_length, fp)) {
// truncated file, error
ion_error(__func__, "fread", Exit, ReadFileError);
}
n += sizeof(char)*rh->name_length;
// set read name length and null-terminator
rh->name->l = rh->name_length;
rh->name->s[rh->name->l]='\0';
n += ion_read_padding(fp, n);
#ifdef ION_SFF_DEBUG
sff_read_header_print(stderr, rh);
#endif
if(rh->rheader_length != n) {
ion_error(__func__, "SFF read header length did not match", Exit, ReadFileError);
}
return rh;
}
dat_header_t *
dat_header_read(FILE *fp)
{
dat_header_t *h=NULL;
h = ion_malloc(sizeof(dat_header_t), __func__, "h");
if(fread_big_endian_uint32_t(fp, &h->signature) != 1
|| fread_big_endian_uint32_t(fp, &h->version) != 1
|| fread_big_endian_uint32_t(fp, &h->header_size) != 1
|| fread_big_endian_uint32_t(fp, &h->data_size) != 1
|| fread_big_endian_uint32_t(fp, &h->wall_time) != 1
|| fread_big_endian_uint16_t(fp, &h->rows) != 1
|| fread_big_endian_uint16_t(fp, &h->cols) != 1
|| fread_big_endian_uint16_t(fp, &h->channels) != 1
|| fread_big_endian_uint16_t(fp, &h->interlace_type) != 1
|| fread_big_endian_uint16_t(fp, &h->frames_in_file) != 1
|| fread_big_endian_uint16_t(fp, &h->reserved) != 1
|| fread_big_endian_uint32_t(fp, &h->sample_rate) != 1
|| fread_big_endian_uint16_t(fp, &h->full_scale_voltage[0]) != 1
|| fread_big_endian_uint16_t(fp, &h->full_scale_voltage[1]) != 1
|| fread_big_endian_uint16_t(fp, &h->full_scale_voltage[2]) != 1
|| fread_big_endian_uint16_t(fp, &h->full_scale_voltage[3]) != 1
|| fread_big_endian_uint16_t(fp, &h->channel_offsets[0]) != 1
|| fread_big_endian_uint16_t(fp, &h->channel_offsets[1]) != 1
|| fread_big_endian_uint16_t(fp, &h->channel_offsets[2]) != 1
|| fread_big_endian_uint16_t(fp, &h->channel_offsets[3]) != 1
|| fread_big_endian_uint16_t(fp, &h->ref_electrode_offset) != 1
|| fread_big_endian_uint16_t(fp, &h->frame_interval) != 1) {
free(h);
return NULL;
}
if(3 != h->version) { // from Image.cpp
ion_error(__func__, "h->version", Exit, OutOfRange);
}
// Check signature
if(h->signature != DAT_HEADER_SIGNATURE) {
ion_error(__func__, "h->signature", Exit, OutOfRange);
}
return h;
}
void
wells_header_print(FILE *fp, wells_header_t *h)
{
if(fprintf(fp, "@HD,num_wells=%d,num_flows=%d,flow_order=%s\n",
h->num_wells,
h->num_flows,
h->flow_order) < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
}
void
wells_mask_print(FILE *fp, wells_mask_t *mask)
{
int32_t i, j, ctr;
for(i=ctr=0;i<mask->num_rows;i++) { // row-major
for(j=0;j<mask->num_cols;j++,ctr++) {
if(0 < i && j < 0) {
if(EOF == fputc(',', fp)) {
ion_error(__func__, "fputc", Exit, WriteFileError);
}
}
// TODO: speed up fprintf
if(fprintf(fp, "%d", mask->masks[i][j]) < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
}
if(EOF == fputc('\n', fp)) {
ion_error(__func__, "fputc", Exit, WriteFileError);
}
}
}
void
sff_read_print(FILE *fp, sff_read_t *r, sff_header_t *gh, sff_read_header_t *rh)
{
uint32_t i;
// flowgram
for(i=0;i<gh->flow_length;i++) {
if(0 < i) {
if(-1L == fputc(',', fp)) {
ion_error(__func__, "fputc", Exit, WriteFileError);
}
}
if(fprintf(fp, "%u", r->flowgram[i]) < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
}
if(-1L == fputc('\n', fp)) {
ion_error(__func__, "fputc", Exit, WriteFileError);
}
// flow index
for(i=0;i<rh->n_bases;i++) {
if(0 < i) {
if(-1L == fputc(',', fp)) {
ion_error(__func__, "fputc", Exit, WriteFileError);
}
}
if(fprintf(fp, "%u", r->flow_index[i]) < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
}
if(-1L == fputc('\n', fp)) {
ion_error(__func__, "fputc", Exit, WriteFileError);
}
// bases
if(fprintf(fp, "%s\n", r->bases->s) < 0) {
ion_error(__func__, "fprintf", Exit, WriteFileError);
}
// quality
for(i=0;i<r->quality->l;i++) {
if(EOF == fputc(QUAL2CHAR(r->quality->s[i]), fp)) {
ion_error(__func__, "fputc", Exit, WriteFileError);
}
}
fputc('\n', fp);
}
sff_read_t *
sff_read_read(FILE *fp, sff_header_t *gh, sff_read_header_t *rh)
{
sff_read_t *r = NULL;
uint32_t i, n = 0;
r = sff_read_init();
r->flowgram = ion_malloc(sizeof(uint16_t)*gh->flow_length, __func__, "r->flowgram");
r->flow_index = ion_malloc(sizeof(uint8_t)*rh->n_bases, __func__, "r->flow_index");
r->bases = ion_string_init(rh->n_bases+1);
r->quality = ion_string_init(rh->n_bases+1);
if(gh->flow_length != fread(r->flowgram, sizeof(uint16_t), gh->flow_length, fp)
|| rh->n_bases != fread(r->flow_index, sizeof(uint8_t), rh->n_bases, fp)
|| rh->n_bases != fread(r->bases->s, sizeof(char), rh->n_bases, fp)
|| rh->n_bases != fread(r->quality->s, sizeof(char), rh->n_bases, fp)) {
// truncated file, error
ion_error(__func__, "fread", Exit, ReadFileError);
}
n += sizeof(uint16_t)*gh->flow_length + 3*sizeof(uint8_t)*rh->n_bases;
// set length and null-terminators
r->bases->l = rh->n_bases;
r->quality->l = rh->n_bases;
r->bases->s[r->bases->l]='\0';
r->quality->s[r->quality->l]='\0';
// convert flowgram to host order
for(i=0;i<gh->flow_length;i++) {
r->flowgram[i] = ntohs(r->flowgram[i]);
}
n += ion_read_padding(fp, n);
#ifdef ION_SFF_DEBUG
sff_read_print(stderr, r, gh, rh);
#endif
return r;
}
int
sff_index_create_main(int argc, char *argv[])
{
int c;
sff_file_t *fp_in, *fp_out;
int32_t num_rows, num_cols, type;
sff_header_t *fp_out_header;
sff_index_t* index;
sff_t *sff;
num_rows = num_cols = -1;
type = SFF_INDEX_ALL;
while((c = getopt(argc, argv, "r:c:C:Rh")) >= 0) {
switch(c) {
case 'r':
num_rows = atoi(optarg);
break;
case 'c':
num_cols = atoi(optarg);
break;
case 'C':
switch(atoi(optarg)) {
case 0:
num_rows = 1152;
num_cols = 1280;
break;
case 1:
num_rows = 2640;
num_cols = 2736;
break;
case 2:
num_rows = 3792;
num_cols = 3392;
break;
default:
break;
}
case 'R':
type = SFF_INDEX_ROW_ONLY;
break;
case 'h':
default:
return usage();
}
}
if(argc != 1+optind) {
return usage();
}
else {
// check cmd line args
if(num_rows < 0) {
ion_error(__func__, "-r must be specified and greater than zero", Exit, CommandLineArgument);
}
if(num_cols < 0) {
ion_error(__func__, "-c must be specified and greater than zero", Exit, CommandLineArgument);
}
switch(type) {
case SFF_INDEX_ROW_ONLY:
case SFF_INDEX_ALL:
break;
default:
ion_error(__func__, "bug encountered", Exit, OutOfRange);
break;
}
fp_in = sff_fopen(argv[optind], "rb", NULL, NULL);
fp_out_header = sff_header_clone(fp_in->header);
index = sff_index_create(fp_in, fp_out_header, num_rows, num_cols, type);
fp_out = sff_fdopen(fileno(stdout), "wbi", fp_out_header, index);
// seek the input file to the beginning of the the entries, which is the same
// location as where the index begins in the output file.
if(0 != fseek(fp_in->fp, fp_out_header->index_offset, SEEK_SET)) {
ion_error(__func__, "fseek", Exit, ReadFileError);
}
// write the sff entries
while(NULL != (sff = sff_read(fp_in))) {
sff_write(fp_out, sff);
sff_destroy(sff);
}
// destroy the header. Don't destroy index, sff_fclose does that
sff_header_destroy(fp_out_header);
// sff_index_destroy(index);
sff_fclose(fp_in);
sff_fclose(fp_out);
}
return 0;
}
// TODO: should we change the header:
// - must trake index_length
// - assumes row-major order
sff_index_t*
sff_index_create(sff_file_t *fp_in, sff_header_t *fp_out_header, int32_t num_rows, int32_t num_cols, int32_t type)
{
int64_t len = 0;
int32_t i, prev_row, prev_col, row, col;
sff_index_t *idx;
sff_t *sff;
uint64_t fp_in_start, prev_pos;
idx = sff_index_init();
idx->num_rows = num_rows;
idx->num_cols = num_cols;
idx->type = type;
// alloc
switch(type) {
case SFF_INDEX_ROW_ONLY:
len = 1 + idx->num_rows;
idx->offset = ion_malloc(len * sizeof(uint64_t), __func__, "idx->offset");
break;
case SFF_INDEX_ALL:
len = 1 + (idx->num_rows * idx->num_cols);
idx->offset = ion_malloc(len * sizeof(uint64_t), __func__, "idx->offset");
break;
default:
ion_error(__func__, "this index type is currently not supported", Exit, OutOfRange);
}
// save where the sff entries started
prev_pos = fp_in_start = ftell(fp_in->fp);
if(-1L == fp_in_start) {
ion_error(__func__, "ftell", Exit, ReadFileError);
}
// go through the input file
i = 0;
prev_row = prev_col = 0;
while(NULL != (sff = sff_read(fp_in))) {
// out of range
if(len-1 <= i) {
ion_error(__func__, "bug encountered", Exit, OutOfRange);
}
// get the row/col co-ordinates
if(0 == ion_readname_to_rowcol(sff->rheader->name->s, &row, &col)) {
ion_error(__func__, "could not understand the read name", Exit, OutOfRange);
}
// assumes row-major order, skips over reads that are not present
if(row < prev_row || (row == prev_row && col < prev_col)) {
ion_error(__func__, "SFF file was not sorted in row-major order", Exit, OutOfRange);
}
while(row != prev_row || col != prev_col) {
// add in empty entry
switch(type) {
case SFF_INDEX_ROW_ONLY:
if(0 == prev_col) { // first column
idx->offset[i] = UINT64_MAX;
// do not increment i, since we only do this when moving to a new row
}
break;
case SFF_INDEX_ALL:
// all rows and columns
idx->offset[i] = UINT64_MAX;
i++;
break;
default:
ion_error(__func__, "this index type is currently not supported", Exit, OutOfRange);
}
if(len-1 <= i) {
ion_error(__func__, "x/y was out of range", Exit, OutOfRange);
}
prev_col++;
if(prev_col == idx->num_cols) {
// new row
prev_col = 0;
prev_row++;
if(SFF_INDEX_ROW_ONLY == type) {
i++;
}
}
}
// add to the index
switch(type) {
case SFF_INDEX_ROW_ONLY:
if(0 == col) { // first column
idx->offset[i] = prev_pos;
}
else if(0 < col && UINT64_MAX == idx->offset[i]) {
idx->offset[i] = prev_pos;
// do not move onto the next
}
break;
case SFF_INDEX_ALL:
// all rows and columns
idx->offset[i] = prev_pos;
i++;
break;
default:
ion_error(__func__, "this index type is currently not supported", Exit, OutOfRange);
}
prev_row = row;
prev_col = col;
// destroy
sff_destroy(sff);
// next
prev_col++;
if(prev_col == idx->num_cols) {
// new row
prev_col = 0;
prev_row++;
if(SFF_INDEX_ROW_ONLY == type) {
i++;
}
}
prev_pos = ftell(fp_in->fp);
if(-1L == prev_pos) {
ion_error(__func__, "ftell", Exit, ReadFileError);
}
}
// get the last offset
idx->offset[len-1] = prev_pos;
// update the index offset in the header
fp_out_header->index_offset = fp_in_start; // insert between the header and sff entries
// update the index length in the header
fp_out_header->index_length = sff_index_length(idx);
// update the offsets based on the index length
for(i=0;i<len;i++) {
if(UINT64_MAX != idx->offset[i]) {
idx->offset[i] += fp_out_header->index_length;
}
}
return idx;
}
int
sff_view_main(int argc, char *argv[])
{
int i, c;
sff_file_t *sff_file_in=NULL, *sff_file_out=NULL;
sff_iter_t *sff_iter = NULL;
sff_t *sff = NULL;
char *fn_names = NULL;
char **names = NULL;
int32_t names_num = 0, names_mem = 0;
int32_t out_mode, min_row, max_row, min_col, max_col;
out_mode = 0;
min_row = max_row = min_col = max_col = -1;
while((c = getopt(argc, argv, "r:c:R:bqh")) >= 0) {
switch(c) {
case 'r':
if(ion_parse_range(optarg, &min_row, &max_row) < 0) {
ion_error(__func__, "-r : format not recognized", Exit, OutOfRange);
}
break;
case 'c':
if(ion_parse_range(optarg, &min_col, &max_col) < 0) {
ion_error(__func__, "-c : format not recognized", Exit, OutOfRange);
}
break;
case 'R':
free(fn_names);
fn_names = strdup(optarg); break;
case 'q':
out_mode |= 1;
break;
case 'b':
out_mode |= 2;
break;
case 'h':
default:
return usage();
}
}
if(argc != 1+optind) {
return usage();
}
else {
sff_header_t *header = NULL;
if(3 == out_mode) {
ion_error(__func__, "options -b and -q cannot be used together", Exit, CommandLineArgument);
}
// open the input SFF
if(-1 != min_row || -1 != max_row || -1 != min_col || -1 != max_col) {
sff_file_in = sff_fopen(argv[optind], "rbi", NULL, NULL);
}
else {
sff_file_in = sff_fopen(argv[optind], "rb", NULL, NULL);
}
header = sff_header_clone(sff_file_in->header); /* copy header, but update n_reads if using index or names */
// read in the names
if(NULL != fn_names) {
FILE *fp = NULL;
char name[1024]="\0"; // lets hope we don't exceed this length
names_num = names_mem = 0;
names = NULL;
if(!(fp = fopen(fn_names, "rb"))) {
fprintf(stderr, "** Could not open %s for reading. **\n", fn_names);
ion_error(__func__, fn_names, Exit, OpenFileError);
}
while(EOF != fscanf(fp, "%s", name)) {
while(names_num == names_mem) {
if(0 == names_mem) names_mem = 4;
else names_mem *= 2;
names = ion_realloc(names, sizeof(char*) * names_mem, __func__, "names");
}
names[names_num] = strdup(name);
if(NULL == names[names_num]) {
ion_error(__func__, name, Exit, MallocMemory);
}
names_num++;
}
names = ion_realloc(names, sizeof(char*) * names_num, __func__, "names");
fclose(fp);
header->n_reads = names_num;
}
else {
// if using index, then iterate once through the index to count the entries
// so we can set the count correctly in the header
if (-1 != min_row || -1 != max_row || -1 != min_col || -1 != max_col) {
int entries = 0;
sff_iter = sff_iter_query(sff_file_in, min_row, max_row, min_col, max_col);
while (NULL != (sff = sff_iter_read(sff_file_in, sff_iter)))
entries++;
header->n_reads = entries;
/* reset sff_iter */
sff_iter_destroy(sff_iter);
sff_iter = sff_iter_query(sff_file_in, min_row, max_row, min_col, max_col);
}
}
// print the header
switch(out_mode) {
case 0:
sff_header_print(stdout, header);
break;
case 1:
// do nothing: FASTQ
break;
case 2:
sff_file_out = sff_fdopen(fileno(stdout), "wb", header, NULL);
break;
}
while(1) {
int32_t to_print = 1;
if(-1 != min_row || -1 != max_row || -1 != min_col || -1 != max_col) {
if(NULL == (sff = sff_iter_read(sff_file_in, sff_iter))) {
break;
}
}
else {
if(NULL == (sff = sff_read(sff_file_in))) {
break;
}
}
if(0 < names_mem) {
to_print = 0;
for(i=0;i<names_num;i++) {
if(0 == strcmp(names[i], sff_name(sff))) {
to_print = 1;
break;
}
}
// shift down
if(1 == to_print) { // i < names_num
free(names[i]);
names[i] = NULL;
for(;i<names_num-1;i++) {
names[i] = names[i+1];
names[i+1] = NULL;
}
names_num--;
}
}
if(1 == to_print) {
switch(out_mode) {
case 0:
sff_print(stdout, sff);
break;
case 1:
if(fprintf(stdout, "@%s\n%s\n+\n",
sff->rheader->name->s,
sff->read->bases->s + sff->gheader->key_length) < 0) {
ion_error(__func__, "stdout", Exit, WriteFileError);
}
for(i=sff->gheader->key_length;i<sff->read->quality->l;i++) {
if(fputc(QUAL2CHAR(sff->read->quality->s[i]), stdout) < 0) {
ion_error(__func__, "stdout", Exit, WriteFileError);
}
}
if(fputc('\n', stdout) < 0) {
ion_error(__func__, "stdout", Exit, WriteFileError);
}
break;
case 2:
sff_write(sff_file_out, sff);
break;
}
}
sff_destroy(sff);
}
sff_fclose(sff_file_in);
if(2 == out_mode) {
sff_fclose(sff_file_out);
}
if(-1 != min_row || -1 != max_row || -1 != min_col || -1 != max_col) {
sff_iter_destroy(sff_iter);
}
if(0 != names_num) {
fprintf(stderr, "** Did not find all the reads with (-R). **\n");
ion_error(__func__, fn_names, Exit, OutOfRange);
}
sff_header_destroy(header);
}
if(NULL != names && 0 < names_num) {
free(names);
}
free(fn_names);
return 0;
}
int
wells_combine_main(int argc, char *argv[])
{
FILE **fps_in = NULL;
wells_chip_t **chips = NULL;
int c;
int32_t i, j, k, l, n;
int32_t nonzero, min_row, max_row, min_col, max_col;
min_row = max_row = min_col = max_col = -1;
nonzero=0;
while((c = getopt(argc, argv, "r:c:zh")) >= 0) {
switch(c) {
case 'r':
if(ion_parse_range(optarg, &min_row, &max_row) < 0) {
ion_error(__func__, "-r : format not recognized", Exit, OutOfRange);
}
break;
case 'c':
if(ion_parse_range(optarg, &min_col, &max_col) < 0) {
ion_error(__func__, "-c : format not recognized", Exit, OutOfRange);
}
break;
case 'z':
nonzero = 1;
break;
case 'h':
default:
return usage();
}
}
if(argc - optind < 2) {
return usage();
}
else {
n = argc - optind;
// open the files
fps_in = ion_calloc(n, sizeof(FILE*), __func__, "fps_in");
for(i=optind;i<argc;i++) {
if(!(fps_in[i-optind] = fopen(argv[i], "rb"))) {
fprintf(stderr, "** Could not open %s for reading. **\n", argv[i]);
ion_error(__func__, argv[i], Exit, OpenFileError);
}
}
// read in the headers
chips = ion_calloc(n, sizeof(wells_chip_t*), __func__, "chips");
for(i=0;i<n;i++) {
// NB: reads in both wells files
if(NULL == (chips[i] = wells_chip_read1(fps_in[i]))) {
ion_error(__func__, argv[i+optind], Exit, ReadFileError);
}
}
// check we can combine the data
for(i=1;i<n;i++) {
if(chips[i-1]->header->num_wells != chips[i]->header->num_wells) {
ion_error(__func__, "# of wells did not match", Exit, OutOfRange);
}
else if(chips[i-1]->header->num_flows != chips[i]->header->num_flows) {
ion_error(__func__, "# of flows did not match", Exit, OutOfRange);
}
else if(0 != strcmp(chips[i-1]->header->flow_order, chips[i]->header->flow_order)) {
ion_error(__func__, "flow order did not match", Exit, OutOfRange);
}
else if(chips[i-1]->num_rows != chips[i]->num_rows) {
ion_error(__func__, "# of rows did not match", Exit, OutOfRange);
}
else if(chips[i-1]->num_cols != chips[i]->num_cols) {
ion_error(__func__, "# of columns did not match", Exit, OutOfRange);
}
}
// write the header
if(0 == wells_header_write(stdout, chips[0]->header)) {
ion_error(__func__, "Could not write to the output file", Exit, WriteFileError);
}
// combine the data
// NB: modifies the wells for the first chip
for(k=0;k<chips[0]->num_cols;k++) { // cols
for(j=0;j<chips[0]->num_rows;j++) { // rows
wells_data_t data_out;
// make room for the new flow values
data_out.flow_values = ion_calloc(chips[0]->header->num_flows, sizeof(float), __func__, "data_out->flow_values");
for(i=0;i<n;i++) { // chip
wells_data_t *data_in = NULL;
// read data in
if(NULL == (data_in = wells_data_read(fps_in[i], chips[i]->header))) {
ion_error(__func__, argv[i+optind], Exit, ReadFileError);
}
// copy over relevant data
if(0 == i) {
data_out.x = data_in->x;
data_out.y = data_in->y;
data_out.rank = data_in->rank;
}
for(l=0;l<chips[i]->header->num_flows;l++) { // flows
// update sum
data_out.flow_values[l] += data_in->flow_values[l];
}
// destroy
wells_data_destroy(data_in);
}
// update
for(l=0;l<chips[0]->header->num_flows;l++) { // flows
data_out.flow_values[l] /= n;
}
// write
if(0 == wells_data_write(stdout, chips[0]->header, &data_out)) {
ion_error(__func__, "Could not write to the output file", Exit, WriteFileError);
}
// destroy
free(data_out.flow_values);
}
}
// close the files
for(i=0;i<n;i++) {
fclose(fps_in[i]);
}
// free
free(fps_in);
free(chips);
}
return 0;
}
dat_frame_t *
dat_frame_read1(FILE *fp, dat_frame_t *cur, dat_frame_t *prev, dat_header_t *header)
{
int32_t i, mode;
uint32_t ctr, compressed;
int8_t *tmp_data8=NULL;
//FILE *fp_debug = stdout; // for debuggin
// read the timestamp and compression flag for this frame
if(fread_big_endian_uint32_t(fp, &cur->timestamp) != 1
|| fread_big_endian_uint32_t(fp, &compressed) != 1) {
return NULL;
}
// the first frame is always uncompressed...
if(NULL == prev || DAT_HEADER_UNINTERLACED == header->interlace_type) {
if(0 != compressed) {
ion_error(__func__, "compressed", Exit, OutOfRange);
}
// read in the data
if(fread(cur->data, sizeof(uint16_t), header->rows*header->cols, fp) != header->rows*header->cols) { // image data
return NULL;
}
// manually convert from big-endian
for(i=0;i<header->rows*header->cols;i++) {
cur->data[i] = ntohs(cur->data[i]) & DAT_FRAME_DATA_MASK; // unmask data
}
}
else {
uint32_t len, transitions, total, sentinel;
uint32_t observed_transitions = 0;
// read in the length of the frame, the # of transitions, the
// total sum of hte pixel values, and hte sentinel
if(fread_big_endian_uint32_t(fp, &len) != 1
|| fread_big_endian_uint32_t(fp, &transitions) != 1
|| fread_big_endian_uint32_t(fp, &total) != 1
|| fread_big_endian_uint32_t(fp, &sentinel) != 1) {
return NULL;
}
// check the sentinel
if(sentinel != DAT_HEADER_SIGNATURE) {
ion_error(__func__, "cur->sentinel", Exit, OutOfRange);
}
// subtract len, transitions, total, sentinel
len -= sizeof(uint32_t)*4;
assert(0 < len);
// initialize memory
cur->data = ion_calloc(header->rows*header->cols, sizeof(uint16_t), __func__, "cur->data");
tmp_data8 = ion_malloc(len*sizeof(int8_t), __func__, "tmp_data8");
// read in the whole frame
// NOTE: could read in byte-by-byte and process to reduce memory overhead
if(len != fread(tmp_data8, sizeof(int8_t), len, fp)) {
free(tmp_data8);
return NULL;
}
// de-interlace the data
i=mode=ctr=0;
while(ctr < header->rows*header->cols) {
if(len <= i) { // not enough bytes read
ion_error(__func__, "len <= i", Exit, OutOfRange);
}
// switch to 8-bit mode, or 16-bit mode where appropriate
if(i < len-1 && DAT_FRAME_KEY_0 == (uint8_t)tmp_data8[i]) {
if(DAT_FRAME_KEY_8_1 == (uint8_t)tmp_data8[i+1]) {
// 16-bit to 8-bit
observed_transitions++;
/*
fprintf(stderr, "[%d-%d] from %d-bit mode to 8-bit mode #%d/%d ctr=%d\n", i, i+1, mode,
observed_transitions, transitions, ctr);
*/
mode = 8;
i+=2;
}
else if(DAT_FRAME_KEY_16_1 == (uint8_t)tmp_data8[i+1]) {
// 8-bit to 16-bit
observed_transitions++;
/*
fprintf(stderr, "[%d-%d] from %d-bit mode to 16-bit mode #%d/%d ctr=%d\n", i, i+1, mode,
observed_transitions, transitions, ctr);
*/
mode = 16;
i+=2;
}
}
// Note: assumes we must have data read between mode switches
// read in data
switch(mode) {
case 8:
// 8-bit mode
cur->data[ctr] = tmp_data8[i] + prev->data[ctr];
ctr++;
i++;
break;
case 16:
// 16-bit mode
cur->data[ctr] = (ntohs((int16_t)((tmp_data8[i] << 8) | tmp_data8[i+1])) & DAT_FRAME_DATA_MASK) + prev->data[ctr];
ctr++;
i+=2;
break;
default:
// mode?
ion_error(__func__, "mode", Exit, OutOfRange);
break;
}
}
if(((i+3) & ~0x3) != len) { // check that the data was quad-word aligned
ion_error(__func__, "quad word alignment", Exit, OutOfRange);
}
// free tmp_data8
free(tmp_data8);
// check that the observed # of transitions equals the state # of
// transitions
if(transitions != observed_transitions) {
ion_error(__func__, "transitions != observed_transitions", Exit, OutOfRange);
}
}
return cur;
}