/*
* Counts the contents of an SRF file.
* If the hash index exists it uses this instead.
*
* Returns num_reads for success
* -1 for failure
*/
int64_t count_file(char *fname, opts *opts) {
srf_t *srf;
srf_index_hdr_t hdr;
off_t ipos, skip;
int item_sz = 9;
if (NULL == (srf = srf_open(fname, "r"))) {
perror(fname);
return -1;
}
/* Read the index header */
if (0 != srf_read_index_hdr(srf, &hdr, 0)) {
srf_destroy(srf, 1);
return list_file(fname, opts);
}
ipos = ftello(srf->fp);
/* Compute the remaining size of the index and divide by item_sz */
if (hdr.dbh_pos_stored_sep)
item_sz += 4;
skip = hdr.index_hdr_sz
+ hdr.n_container * 8
+ hdr.n_data_block_hdr * 8
+ hdr.n_buckets * 8;
srf_destroy(srf, 1);
return (hdr.size - skip - 16/* footer*/) / item_sz;
}
/*
* Lists the contents of an SRF file.
*
* Returns num_reads for success
* -1 for failure
*/
int64_t list_file(char *fname, opts *opts) {
srf_t *srf;
char name[512];
int64_t count = 0;
int type;
uint64_t pos;
if (NULL == (srf = srf_open(fname, "r"))) {
perror(fname);
return -1;
}
/* Scan through file gathering the details to index in memory */
while ((type = srf_next_block_details(srf, &pos, name)) >= 0) {
if (type == SRFB_TRACE_BODY) {
count++;
if (!opts->count_only) {
if (opts->long_format)
printf("%-30s %10"PRId64" + %4d + %5d\n",
name, pos,
srf->tb.trace_size,
srf->th.trace_hdr_size);
else
puts(name);
}
}
}
srf_destroy(srf, 1);
return count;
}
int check_srf(char *srf_name, Settings *opts) {
FILE *fp = fopen(srf_name, "rb");
srf_t *srf;
int res;
if (NULL == fp) {
printf("Couldn't open %s: %s\n", srf_name, strerror(errno));
return -1;
}
srf = srf_create(fp);
if (NULL == srf) {
printf("srf_create failed for %s\n", srf_name);
fclose(fp);
return -1;
}
res = read_srf(srf, srf_name, opts);
srf_destroy(srf, 0);
if (0 != fclose(fp)) {
printf("Error closing %s: %s\n", srf_name, strerror(errno));
return -1;
}
return res;
}
/*
* Given the archive name (ar_name), the chunk types to output (chunk_mode),
* and some other parameters such as the read filter, open one file and dumps
* text formatted output to that file. The number of reads is updated in the
* \"num_reads\" parameter.
*
* Returns 0 on success.
*/
int process_srf_to_text_files(char *ar_name, char chunk_mode, int num_reads_only_mode, int filter_mode, read_filter_t *read_filter, long *num_reads) {
srf_t *srf;
char name[1024], dir2[1024];
ztr_t *ztr;
int last_lane = 0, last_tile = 0;
/* dump */
FILE *files[] = {NULL};
if (NULL == (srf = srf_open(ar_name, "rb"))) {
perror(ar_name);
return 1;
}
char *cp = strrchr(ar_name, '.');
if (cp) *cp = 0;
sprintf(dir2, "%s.run", ar_name);
mkdir(dir2, 0777);
while (NULL != (ztr = srf_next_ztr(srf, name, 0))) {
int lane, tile, x, y;
parse_name(name, &lane, &tile, &x, &y);
if (last_lane != lane || last_tile != tile) {
fprintf(stderr, "New tile: %d/%d\n", lane, tile);
last_lane = lane;
last_tile = tile;
if (files[0]) { fclose(files[0]); files[0] = NULL; }
if (chunk_mode) {
files[0] = fopen_text(dir2, lane, tile);
}
}
if(!num_reads_only_mode) {
if(!filter_mode || (filter_mode && check_read_name(read_filter, name))) {
dump_text(ztr, name, chunk_mode, files);
++(*num_reads);
}
} else {
++(*num_reads);
}
delete_ztr(ztr);
}
if (files[0]) fclose(files[0]);
srf_destroy(srf, 1);
return 0;
}
/*
* Given the archive name and the level_mode
* generate information about the archive
*
* Note the generated srf file is NOT indexed
*
* Returns 0 on success.
*/
int srf_info(char *input, int level_mode, long *read_count, long *chunk_count,
uint64_t *chunk_size, long key_count[NCHUNKS][NKEYS],
long type_count[NCHUNKS][NTYPES], HashTable *regn_hash,
uint64_t *base_count) {
srf_t *srf;
off_t pos;
int type;
int count = 0;
long trace_body_count = 0;
char name[1024];
if (NULL == (srf = srf_open(input, "rb"))) {
perror(input);
return 1;
}
while ((type = srf_next_block_type(srf)) >= 0) {
switch (type) {
case SRFB_CONTAINER:
if( trace_body_count ){
if( level_mode & LEVEL_NAME )
printf( " ... %s x%ld\n", name+strlen(srf->th.id_prefix), trace_body_count);
trace_body_count = 0;
}
if (0 != srf_read_cont_hdr(srf, &srf->ch)) {
fprintf(stderr, "Error reading container header.\nExiting.\n");
exit(1);
}
break;
case SRFB_XML:
if( trace_body_count ){
if( level_mode & LEVEL_NAME )
printf( " ... %s x%ld\n", name+strlen(srf->th.id_prefix), trace_body_count);
trace_body_count = 0;
}
if (0 != srf_read_xml(srf, &srf->xml)) {
fprintf(stderr, "Error reading XML.\nExiting.\n");
exit(1);
}
break;
case SRFB_TRACE_HEADER:
if( trace_body_count ){
if( level_mode & LEVEL_NAME )
printf( " ... %s x%ld\n", name+strlen(srf->th.id_prefix), trace_body_count);
trace_body_count = 0;
}
if (0 != srf_read_trace_hdr(srf, &srf->th)) {
fprintf(stderr, "Error reading trace header.\nExiting.\n");
exit(1);
}
if( 0 == (level_mode & (LEVEL_CHUNK | LEVEL_BASE)) )
break;
/* Decode ZTR chunks in the header */
if (srf->mf)
mfdestroy(srf->mf);
srf->mf = mfcreate(NULL, 0);
if (srf->th.trace_hdr_size)
mfwrite(srf->th.trace_hdr, 1, srf->th.trace_hdr_size, srf->mf);
if (srf->ztr)
delete_ztr(srf->ztr);
mrewind(srf->mf);
if (NULL != (srf->ztr = partial_decode_ztr(srf, srf->mf, NULL))) {
srf->mf_pos = mftell(srf->mf);
} else {
/* Maybe not enough to decode or no headerBlob. */
/* So delay until decoding the body. */
srf->mf_pos = 0;
}
mfseek(srf->mf, 0, SEEK_END);
srf->mf_end = mftell(srf->mf);
break;
case SRFB_TRACE_BODY: {
srf_trace_body_t old_tb;
ztr_t *ztr_tmp;
int no_trace = (level_mode & (LEVEL_CHUNK | LEVEL_BASE) ? 0 : 1);
if (0 != srf_read_trace_body(srf, &old_tb, no_trace)) {
fprintf(stderr, "Error reading trace body.\nExiting.\n");
exit(1);
}
if (-1 == construct_trace_name(srf->th.id_prefix,
(unsigned char *)old_tb.read_id,
old_tb.read_id_length,
name, 512)) {
fprintf(stderr, "Error constructing trace name.\nExiting.\n");
exit(1);
}
trace_body_count++;
if( 1 == trace_body_count ){
if( level_mode & LEVEL_NAME )
printf( "trace_name: %s + %s", srf->th.id_prefix, name+strlen(srf->th.id_prefix));
}
read_count[READ_TOTAL]++;
if (old_tb.flags & SRF_READ_FLAG_BAD_MASK ){
read_count[READ_BAD]++;
} else {
read_count[READ_GOOD]++;
}
if( 0 == (level_mode & (LEVEL_CHUNK | LEVEL_BASE)) )
break;
if (!srf->mf) {
fprintf(stderr, "Error reading trace body.\nExiting.\n");
exit(1);
}
mfseek(srf->mf, srf->mf_end, SEEK_SET);
if (old_tb.trace_size) {
mfwrite(old_tb.trace, 1, old_tb.trace_size, srf->mf);
free(old_tb.trace);
old_tb.trace = NULL;
}
mftruncate(srf->mf, mftell(srf->mf));
mfseek(srf->mf, srf->mf_pos, SEEK_SET);
if (srf->ztr)
ztr_tmp = ztr_dup(srf->ztr); /* inefficient, but simple */
else
ztr_tmp = NULL;
if ((ztr_tmp = partial_decode_ztr(srf, srf->mf, ztr_tmp))) {
int i;
for (i=0; i<ztr_tmp->nchunks; i++) {
int ichunk = -1;
switch (ztr_tmp->chunk[i].type) {
case ZTR_TYPE_BASE:
ichunk = CHUNK_BASE;
chunk_size[ichunk] += ztr_tmp->chunk[i].dlength;
if( parse_base(ztr_tmp, &ztr_tmp->chunk[i], base_count) ){
delete_ztr(ztr_tmp);
return 1;
}
break;
case ZTR_TYPE_CNF1:
ichunk = CHUNK_CNF1;
chunk_size[ichunk] += ztr_tmp->chunk[i].dlength;
break;
case ZTR_TYPE_CNF4:
ichunk = CHUNK_CNF4;
chunk_size[ichunk] += ztr_tmp->chunk[i].dlength;
break;
case ZTR_TYPE_SAMP:
ichunk = CHUNK_SAMP;
chunk_size[ichunk] += ztr_tmp->chunk[i].dlength;
break;
case ZTR_TYPE_SMP4:
ichunk = CHUNK_SMP4;
chunk_size[ichunk] += ztr_tmp->chunk[i].dlength;
break;
case ZTR_TYPE_REGN:
ichunk = CHUNK_REGN;
chunk_size[ichunk] += ztr_tmp->chunk[i].dlength;
if( NULL == parse_regn(ztr_tmp, &ztr_tmp->chunk[i], regn_hash) ){
delete_ztr(ztr_tmp);
return 1;
}
break;
default:
break;
}
if( ichunk > -1 ) {
chunk_count[ichunk]++;
count_mdata_keys(ztr_tmp, &ztr_tmp->chunk[i], ichunk, key_count, type_count);
}
}
}
if( ztr_tmp )
delete_ztr(ztr_tmp);
count++;
if( (level_mode == LEVEL_CHECK) && (count == 10) ){
printf( " ... %s x%ld\n", name+strlen(srf->th.id_prefix), trace_body_count);
srf_destroy(srf, 1);
return 0;
}
break;
}
case SRFB_INDEX: {
off_t pos = ftell(srf->fp);
if( trace_body_count ){
if( level_mode & LEVEL_NAME )
printf( " ... %s x%ld\n", name+strlen(srf->th.id_prefix), trace_body_count);
trace_body_count = 0;
}
printf( "Reading srf index block\n");
if (0 != srf_read_index_hdr(srf, &srf->hdr, 1)) {
srf_destroy(srf, 1);
fprintf(stderr, "Error reading srf index block header.\nExiting.\n");
exit(1);
}
/* Skip the index body */
fseeko(srf->fp, pos + srf->hdr.size, SEEK_SET);
break;
}
case SRFB_NULL_INDEX: {
uint64_t ilen;
if( trace_body_count ){
if( level_mode & LEVEL_NAME )
printf( " ... %s x%ld\n", name+strlen(srf->th.id_prefix), trace_body_count);
trace_body_count = 0;
}
printf( "Reading srf null index block\n");
/*
* Maybe the last 8 bytes of a the file (or previously was
* last 8 bytes prior to concatenating SRF files together).
* If so it's the index length and should always be 8 zeros.
*/
if (1 != fread(&ilen, 8, 1, srf->fp)) {
srf_destroy(srf, 1);
fprintf(stderr, "Error reading srf null index block.\nExiting.\n");
exit(1);
}
if (ilen != 0) {
srf_destroy(srf, 1);
fprintf(stderr, "Invalid srf null index block.\nExiting.\n");
exit(1);
}
break;
}
default:
srf_destroy(srf, 1);
fprintf(stderr, "Block of unknown type '%c'\nExiting.\n", type);
exit(1);
}
}
if( trace_body_count ){
if( level_mode & LEVEL_NAME )
printf( " ... %s x%ld\n", name+strlen(srf->th.id_prefix), trace_body_count);
trace_body_count = 0;
}
/* the type should be -1 (EOF) */
if( type != -1 ) {
fprintf(stderr, "Block of unknown type '%c'\nExiting.\n", type);
exit(1);
}
/* are we really at the end of the srf file */
pos = ftell(srf->fp);
fseek(srf->fp, 0, SEEK_END);
if( pos != ftell(srf->fp) ){
fprintf(stderr, "srf file is corrupt\n");
exit(1);
}
srf_destroy(srf, 1);
return 0;
}
/*
* Main method.
*/
int main(int argc, char **argv) {
int i;
long num_reads = 0;
int num_reads_mode = 0;
int num_reads_only_mode = 0;
int filter_mode = 0;
char *ar_name = NULL;
srf_t *srf = NULL;
char name[512];
ztr_t *ztr = NULL;
read_filter_t *read_filter = NULL;
char *filter_value = NULL;
int c;
int errflg = 0;
extern char *optarg;
extern int optind, optopt;
char chunk_mode = ALL;
char type_mode = TEXT;
char destination_mode = CONSOLE_DEST;
FILE **files = NULL;
int verbose = 0;
if (argc < 2) {
fprintf(stderr, "Please specify an archive name.\n");
usage(1);
}
while ((c = getopt(argc, argv, ":c:d:f:not:v")) != -1) {
switch (c) {
case 'c':
chunk_mode = 0;
if(get_chunk_types(optarg, &chunk_mode) || !chunk_mode) {
fprintf(stderr,
"Invalid value \"%s\" given to option -%c.\n", optarg, c);
errflg++;
}
break;
case 'd':
destination_mode = 0;
if(get_destination_types(optarg, &destination_mode) || !destination_mode) {
fprintf(stderr,
"Invalid value \"%s\" given to option -%c.\n", optarg, c);
errflg++;
}
break;
case 'f':
if (num_reads_only_mode) {
fprintf(stderr,
"Option -%c is exclusing with option -o.\n", c);
errflg++;
} else {
filter_mode++;
filter_value = optarg;
}
break;
case 'n':
if (num_reads_only_mode) {
fprintf(stderr,
"Option -%c is exclusing with option -o.\n", c);
errflg++;
}
else
num_reads_mode++;
break;
case 'o':
if (num_reads_mode) {
fprintf(stderr,
"Option -%c is exclusing with option -n.\n", c);
errflg++;
} else if (filter_mode) {
fprintf(stderr,
"Option -%c is exclusing with option -f.\n", c);
errflg++;
} else
num_reads_only_mode++;
break;
case 't':
type_mode=0;
if(get_type_of_output(optarg, &type_mode) || !type_mode) {
fprintf(stderr,
"Invalid value \"%s\" given to option -%c.\n", optarg, c);
errflg++;
}
break;
case 'v':
verbose++;
break;
case ':': /* -? without operand */
fprintf(stderr,
"Option -%c requires an operand\n", optopt);
errflg++;
break;
case '?':
fprintf(stderr,
"Unrecognised option: -%c\n", optopt);
errflg++;
}
}
if (errflg) {
usage(1);
}
if(optind + 1 != argc) {
fprintf(stderr, "The archive name must be the last argument.\n");
usage(1);
} else {
ar_name = argv[optind];
}
if(filter_mode) {
read_filter = get_read_filter(filter_value);
if(verbose) {
dump_read_filter(read_filter);
}
}
if(chunk_mode && verbose) {
dump_chunk_mode(chunk_mode);
}
if(type_mode && verbose) {
dump_type_mode(type_mode);
}
if(destination_mode && verbose) {
dump_destination_mode(destination_mode);
}
if (!access(ar_name, R_OK)) {
if(verbose) {
printf("Dumping from archive %s.\n", ar_name);
}
} else {
fprintf(stderr, "Archive %s not found.\n", ar_name);
}
if(destination_mode & NONE_DEST) {
if (NULL == (srf = srf_open(ar_name, "rb"))) {
perror(ar_name);
return 1;
}
while (NULL != (ztr = srf_next_ztr(srf, name, 0))) {
if(!num_reads_only_mode) {
if(!filter_mode || (filter_mode && check_read_name(read_filter, name))) {
++num_reads;
}
} else {
++num_reads;
}
delete_ztr(ztr);
}
}
if(destination_mode & CONSOLE_DEST) {
/* Cosole means stdout. */
files = malloc((sizeof(FILE *))*5);
for(i = 0; i < 5; i++) {
files[i] = stdout;
}
if (NULL == (srf = srf_open(ar_name, "rb"))) {
perror(ar_name);
return 1;
}
while (NULL != (ztr = srf_next_ztr(srf, name, 0))) {
if(!num_reads_only_mode) {
if(!filter_mode || (filter_mode && check_read_name(read_filter, name))) {
if(type_mode & SOLEXA) {
dump_solexa(ztr, name, chunk_mode, files);
} else if(type_mode & TEXT) {
dump_text(ztr, name, chunk_mode, files);
} else {
fprintf(stderr, "Assertion error on type_mode (%c).\nExiting.\n", type_mode);
exit(1);
}
++num_reads;
}
} else {
++num_reads;
}
delete_ztr(ztr);
}
}
if(destination_mode & FILE_DEST) {
if(type_mode & SOLEXA) {
process_srf_to_solexa_files(ar_name, chunk_mode, num_reads_only_mode, filter_mode, read_filter, &num_reads);
} else if(type_mode & TEXT) {
process_srf_to_text_files(ar_name, chunk_mode, num_reads_only_mode, filter_mode, read_filter, &num_reads);
} else {
fprintf(stderr, "Assertion error on type_mode (%c).\nExiting.\n", type_mode);
exit(1);
}
}
if(num_reads_mode || num_reads_only_mode) {
printf("\nReads: %ld\n", num_reads);
}
srf_destroy(srf, 1);
return 0;
}