/*
* 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;
}
int main(int argc, char **argv) {
ztr_t *ztr;
mFILE *fp;
int i;
if (argc >= 2) {
if (NULL == (fp = mfopen(argv[1], "rb"))) {
perror(argv[1]);
return 1;
}
} else {
fp = mstdin();
}
if (NULL == (ztr = mfread_ztr(fp))) {
perror("fread_ztr");
return 1;
}
printf("Nchunks = %d\n", ztr->nchunks);
for (i = 0; i < ztr->nchunks; i++) {
char str[5];
int complen;
int rawlen;
char *val;
(void)ZTR_BE2STR(ztr->chunk[i].type, str);
complen = ztr->chunk[i].dlength;
val = ztr_lookup_mdata_value(ztr, &ztr->chunk[i], "TYPE");
if (val)
printf("-- %s (%s) --\n", str, val);
else
printf("-- %s --\n", str);
explode_chunk(ztr, &ztr->chunk[i]);
rawlen = ztr->chunk[i].dlength;
printf("SUMMARY %s mlen %3d, dlen %6d, rawlen %6d, ratio %f\n",
str, ztr->chunk[i].mdlength,
complen, rawlen, (double)complen/rawlen);
#if 0
fflush(stdout);
puts("\n========================================");
write(1, ztr->chunk[i].data, ztr->chunk[i].dlength);
puts("\n========================================");
#endif
}
delete_ztr(ztr);
return 0;
}
/*
* fread_ztr
*
* Reads a ZTR file from 'fp'. This checks for the correct magic number and
* major version number, but not minor version number.
*
* FIXME: Add automatic uncompression?
*
* Arguments:
* fp A readable FILE pointer
*
* Returns:
* Success: Pointer to a ztr_t structure (malloced)
* Failure: NULL
*/
ztr_t *fread_ztr(FILE *fp) {
ztr_t *ztr;
ztr_chunk_t *chunk;
int sections = read_sections(0);
/* Allocate */
if (NULL == (ztr = new_ztr()))
return NULL;
/* Read the header */
if (-1 == ztr_read_header(fp, &ztr->header))
return NULL;
/* Check magic number and version */
if (memcmp(ztr->header.magic, ZTR_MAGIC, 8) != 0)
return NULL;
if (ztr->header.version_major != ZTR_VERSION_MAJOR)
return NULL;
/* Load chunks */
while ((chunk = ztr_read_chunk_hdr(fp))) {
/*
char str[5];
fprintf(stderr, "Read chunk %.4s %08x length %d\n",
ZTR_BE2STR(chunk->type, str), chunk->type, chunk->dlength);
*/
switch(chunk->type) {
case ZTR_TYPE_HEADER:
/* End of file */
return ztr;
case ZTR_TYPE_SAMP:
case ZTR_TYPE_SMP4:
if (! (sections & READ_SAMPLES)) {
fseek(fp, chunk->dlength, SEEK_CUR);
xfree(chunk);
continue;
}
break;
case ZTR_TYPE_BASE:
case ZTR_TYPE_BPOS:
case ZTR_TYPE_CNF4:
case ZTR_TYPE_CNF1:
case ZTR_TYPE_CSID:
if (! (sections & READ_BASES)) {
fseek(fp, chunk->dlength, SEEK_CUR);
xfree(chunk);
continue;
}
break;
case ZTR_TYPE_TEXT:
if (! (sections & READ_COMMENTS)) {
fseek(fp, chunk->dlength, SEEK_CUR);
xfree(chunk);
continue;
}
break;
case ZTR_TYPE_CLIP:
case ZTR_TYPE_FLWO:
case ZTR_TYPE_FLWC:
break;
/*
default:
fprintf(stderr, "Unknown chunk type '%s': skipping\n",
ZTR_BE2STR(chunk->type,str));
fseek(fp, chunk->dlength, SEEK_CUR);
xfree(chunk);
continue;
*/
}
chunk->ztr_owns = 1;
chunk->data = (char *)xmalloc(chunk->dlength);
if (chunk->dlength != fread(chunk->data, 1, chunk->dlength, fp)) {
delete_ztr(ztr);
return NULL;
}
ztr->nchunks++;
ztr->chunk = (ztr_chunk_t *)xrealloc(ztr->chunk, ztr->nchunks *
sizeof(ztr_chunk_t));
memcpy(&ztr->chunk[ztr->nchunks-1], chunk, sizeof(*chunk));
xfree(chunk);
}
return ztr;
}
/*
* Write a sequence to a FILE *fp of format "format". If "format" is 0,
* we choose our favourite - SCF.
*
* Returns:
* 0 for success
* -1 for failure
*/
int mfwrite_reading(mFILE *fp, Read *read, int format) {
int r = -1;
int no_compress = 0;
#ifdef _WIN32
/*
* jkb 09/06/00 comment below
*
* On windows "prog > file.scf" will work wrongly (compared to
* "prog file.scf") because windows is rather stupid. It treats ascii
* and binary streams differently, it considers stdout to be ascii unless
* told otherwise, and it can only be told otherwise by using non-ansi
* windows-specific function calls.
*/
if (format != TT_EXP && format != TT_PLN && fp->fp)
_setmode(_fileno(fp->fp), _O_BINARY);
#endif
switch (format) {
default:
/* Defaults to ZTR type */
#ifdef IOLIB_ZTR
case TT_ZTR:
case TT_ZTR2: {
ztr_t *ztr;
ztr = read2ztr(read);
compress_ztr(ztr, 2);
r = mfwrite_ztr(fp, ztr);
delete_ztr(ztr);
no_compress = 1;
break;
}
case TT_ZTR1: {
ztr_t *ztr;
ztr = read2ztr(read);
compress_ztr(ztr, 1);
r = mfwrite_ztr(fp, ztr);
delete_ztr(ztr);
break;
}
case TT_ZTR3: {
ztr_t *ztr;
ztr = read2ztr(read);
compress_ztr(ztr, 3);
r = mfwrite_ztr(fp, ztr);
delete_ztr(ztr);
no_compress = 1;
break;
}
#endif
#ifdef IOLIB_SCF
case TT_SCF: {
Scf *scf;
scf = read2scf(read);
r = mfwrite_scf(scf, fp);
scf_deallocate(scf);
break;
}
#endif
#ifdef IOLIB_CTF
case TT_CTF:
r = mfwrite_ctf(fp, read);
break;
#endif
#ifdef IOLIB_ABI
case TT_ABI:
/*return mfwrite_abi(fp, read); */
break;
#endif
#ifdef IOLIB_ALF
case TT_ALF:
/* return mfwrite_alf(fp, read); */
break;
#endif
#ifdef IOLIB_EXP
case TT_EXP: {
Exp_info *e = read2exp(read, read->ident ? read->ident : "unknown");
if (NULL == e) {
fprintf(stderr, "Failed to create experiment file.\n");
r = -1;
} else {
exp_print_mfile(fp, e);
exp_destroy_info(e);
r = 0;
}
break;
}
#endif
#ifdef IOLIB_PLN
case TT_PLN:
r = mfwrite_pln(fp, read);
break;
#endif
}
mftruncate(fp, -1);
if (r == 0 && !no_compress) {
fcompress_file(fp);
}
mfflush(fp);
return r;
}
/*
* Read a sequence from a FILE *fp of format "format". If "format" is 0
* (ANY_FORMAT), we automatically determine the correct format.
* We still pass a filename 'fn' although this isn't used other than for
* filling in the read->trace_name field.
*
* NB this function should NOT be used when Biolims support is required
* (as biolims readings are not stored in a file)
*
* Returns:
* Read * for success
* NULLRead for failure
*/
Read *mfread_reading(mFILE *fp, char *fn, int format) {
Read *read;
mFILE *newfp;
if (!fn)
fn = "(unknown)";
newfp = freopen_compressed(fp, NULL);
if (newfp != fp) {
fp = newfp;
} else {
newfp = NULL;
}
#ifdef _WIN32
/*
* jkb 16/05/00 comment below
*
* On windows "prog < file.abi" will work wrongly (compared to
* "prog file.abi") because windows is rather stupid. It treats ascii
* and binary streams differently, it considers stdin to be ascii unless
* told otherwise, and it can only be told otherwise by using non-ansi
* windows-specific function calls.
*/
if (format != TT_EXP && format != TT_PLN && fp->fp)
_setmode(_fileno(fp->fp), _O_BINARY);
#endif
if (format == TT_ANY) {
format = fdetermine_trace_type(fp);
mrewind(fp);
}
switch (format) {
case TT_UNK:
case TT_ERR:
errout("File '%s' has unknown trace type\n", fn);
read = NULLRead;
break;
#ifdef IOLIB_SCF
case TT_SCF: {
Scf *scf;
scf = mfread_scf(fp);
if (scf) {
read = scf2read(scf);
scf_deallocate(scf);
} else
read = NULLRead;
break;
}
#endif
#ifdef IOLIB_CTF
case TT_CTF:
read = mfread_ctf(fp);
break;
#endif
#ifdef IOLIB_ZTR
case TT_ZTR:
case TT_ZTR1:
case TT_ZTR2:
case TT_ZTR3: {
ztr_t *ztr;
if ((ztr = mfread_ztr(fp))) {
uncompress_ztr(ztr);
read = ztr2read(ztr);
delete_ztr(ztr);
} else {
read = NULLRead;
}
break;
}
#endif
#ifdef IOLIB_ABI
case TT_ABI:
read = mfread_abi(fp);
break;
#endif
#ifdef IOLIB_ALF
case TT_ALF:
read = mfread_alf(fp);
break;
#endif
#ifdef IOLIB_EXP
case TT_EXP: {
/* FIXME: we shouldn't redirect like this */
Exp_info *e = exp_mfread_info(fp);
read = e ? exp2read(e,fn) : NULLRead;
break;
}
#endif
#ifdef IOLIB_PLN
case TT_PLN:
read = mfread_pln(fp);
break;
#endif
default:
errout("Unknown format %d specified to read_reading()\n", format);
read = NULLRead;
}
if (read != NULLRead && (read->trace_name = (char *)xmalloc(strlen(fn)+1)))
strcpy(read->trace_name, fn);
if (newfp) mfclose(newfp);
return read;
}
int check_trace_body(Settings *opts, Trace_reader *trc, srf_t *srf) {
char name[512];
Spot_info spot = { NULL, 0, 0, 0, 0, 0, 0 };
ztr_t *ztr = NULL;
ztr_chunk_t *chunk;
int start_chunk = 0;
int i;
if (-1 == construct_trace_name(srf->th.id_prefix,
(unsigned char *)srf->tb.read_id,
srf->tb.read_id_length,
name, sizeof(name))) {
printf("Couldn't construct read name\n");
return -1;
}
if (0 != decode_name(name, &spot)) {
printf("Couldn't decode read name.\n");
return -1;
}
if (opts->verbosity > 0) {
printf("%s\n", name);
}
if (0 != read_trace_data(opts, trc, &spot)) {
printf("Couldn't get trace data for %s\n", name);
return -1;
}
mfseek(srf->mf, srf->mf_end, SEEK_SET);
if (srf->tb.trace_size) {
mfwrite(srf->tb.trace, 1, srf->tb.trace_size, srf->mf);
free(srf->tb.trace);
srf->tb.trace = NULL;
}
mftruncate(srf->mf, mftell(srf->mf));
mfseek(srf->mf, srf->mf_pos, SEEK_SET);
if (srf->ztr) {
start_chunk = srf->ztr->nchunks;
ztr = ztr_dup(srf->ztr);
}
if (NULL == partial_decode_ztr(srf, srf->mf, ztr)) {
printf("Couldn't decode ZTR data for %s\n", name);
return -1;
}
for (i = start_chunk; i < ztr->nchunks; i++) {
chunk = &ztr->chunk[i];
if (opts->verbosity > 1) {
printf(" Chunk %d type %.4s\n", i,
(char *) &chunk->type);
}
switch (chunk->type) {
case ZTR_TYPE_BASE:
if (0 != uncompress_chunk(ztr, chunk)) {
printf("Couldn't uncompress BASE chunk\n");
return -1;
}
if (trc->seq_len + 1 != chunk->dlength
|| 0 != memcmp(trc->seq, chunk->data + 1, trc->seq_len)) {
printf("Sequence differs for %s\n", name);
return -1;
}
break;
case ZTR_TYPE_CNF1:
if (0 != uncompress_chunk(ztr, chunk)) {
printf("Couldn't uncompress CNF1 chunk\n");
return -1;
}
if (trc->seq_len + 1 != chunk->dlength
|| 0 != memcmp(trc->qual, chunk->data + 1, trc->seq_len)) {
printf("CNF1 confidence values differ for %s\n", name);
return -1;
}
break;
case ZTR_TYPE_CNF4:
if (0 != uncompress_chunk(ztr, chunk)) {
printf("Couldn't uncompress CNF4 chunk\n");
return -1;
}
if (0 != check_cnf4(ztr, chunk, trc)) {
printf("CNF4 confidence values differ for %s\n", name);
return -1;
}
break;
case ZTR_TYPE_SMP4:
if (0 != uncompress_chunk(ztr, chunk)) {
printf("Couldn't uncompress SMP4 chunk\n");
return -1;
}
if (0 != check_trace(ztr, chunk, trc)) {
printf("Trace doesn't match for %s\n", name);
return -1;
}
break;
default:
printf("Found unexpected chunk type in trace body for %s\n",
name);
return -1;
}
}
delete_ztr(ztr);
return 0;
}
int check_trace_header(Settings *opts, srf_t *srf, Previous_data *seen) {
int i;
int j;
int n;
uint8_t *data;
uint32_t val;
ztr_chunk_t *chunk;
if (NULL != srf->mf) {
mfrecreate(srf->mf, NULL, 0);
} else {
srf->mf = mfcreate(NULL, 0);
}
if (NULL == srf->mf) die("%s\n", strerror(errno));
if (srf->th.trace_hdr_size) {
if (1 != mfwrite(srf->th.trace_hdr, srf->th.trace_hdr_size, 1, srf->mf)) {
die("mfwrite failed: %s\n", strerror(errno));
}
}
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 {
/* We expect this to work for srfs made by illumina2srf */
printf("partial_decode_ztr failed for trace header\n");
srf->mf_pos = 0;
return -1;
}
mfseek(srf->mf, 0, SEEK_END);
srf->mf_end = mftell(srf->mf);
/* Go through chunks */
for (i = 0; i < srf->ztr->nchunks; i++) {
chunk = &srf->ztr->chunk[i];
if (opts->verbosity > 1) {
printf(" Chunk %d type %.4s\n", i,
(char *) &chunk->type);
}
switch (chunk->type) {
case ZTR_TYPE_HUFF:
break;
case ZTR_TYPE_BPOS:
if (0 != uncompress_chunk(srf->ztr, chunk)) {
printf("Couldn't uncompress BPOS chunk\n");
return -1;
}
n = (chunk->dlength - 4) / 4;
for (j = 0; j < n; j++) {
data = (uint8_t *) &chunk->data[j * 4 + 4];
val = ( ((uint32_t) data[0]) << 24
| ((uint32_t) data[1]) << 16
| ((uint32_t) data[2]) << 8
| ((uint32_t) data[3]));
if (val != j) {
printf("BPOS data misses cycles\n");
return -1;
}
}
break;
case ZTR_TYPE_REGN:
if (0 != uncompress_chunk(srf->ztr, chunk)) {
printf("Couldn't uncompress REGN chunk\n");
return -1;
}
if (NULL == seen->regn) {
/* Copy REGN chunk */
seen->regn_meta_sz = chunk->mdlength;
seen->regn_meta = smalloc(seen->regn_meta_sz);
memcpy(seen->regn_meta, chunk->mdata, seen->regn_meta_sz);
seen->regn_sz = chunk->dlength;
seen->regn = smalloc(seen->regn_sz);
memcpy(seen->regn, chunk->data, seen->regn_sz);
} else {
/* Compare with last copy */
if (seen->regn_meta_sz != chunk->mdlength
|| seen->regn_sz != chunk->dlength
|| 0 != memcmp(seen->regn_meta, chunk->mdata, seen->regn_meta_sz)
|| 0 != memcmp(seen->regn, chunk->data, seen->regn_sz)) {
printf("REGN chunk changed between header blocks\n");
return -1;
}
}
break;
case ZTR_TYPE_TEXT:
if (0 != uncompress_chunk(srf->ztr, chunk)) {
printf("Couldn't uncompress REGN chunk\n");
return -1;
}
if (NULL == seen->text) {
seen->text_sz = chunk->dlength;
seen->text = smalloc(seen->text_sz);
memcpy(seen->text, chunk->data, seen->text_sz);
if (0 != check_text(opts, seen)) return -1;
} else {
if (seen->text_sz != chunk->dlength
|| 0 != memcmp(seen->text, chunk->data, seen->text_sz)) {
printf("New TEXT chunk found\n");
seen->text_sz = chunk->dlength;
seen->text = srealloc(seen->text, seen->text_sz);
memcpy(seen->text, chunk->data, seen->text_sz);
if (0 != check_text(opts, seen)) return -1;
}
}
break;
default:
printf("Found unexpected chunk type in header block\n");
return -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;
}
/*
* As per partial_decode_ztr in srf.c, but without uncompress_ztr.
*/
ztr_t *partial_decode_ztr2(srf_t *srf, mFILE *mf, ztr_t *z) {
ztr_t *ztr;
ztr_chunk_t *chunk;
long pos = 0;
if (z) {
/* Use existing ZTR object => already loaded header */
ztr = z;
} else {
/* Allocate or use existing ztr */
if (NULL == (ztr = new_ztr()))
return NULL;
/* Read the header */
if (-1 == ztr_read_header(mf, &ztr->header)) {
if (!z)
delete_ztr(ztr);
mrewind(mf);
return NULL;
}
/* Check magic number and version */
if (memcmp(ztr->header.magic, ZTR_MAGIC, 8) != 0) {
if (!z)
delete_ztr(ztr);
mrewind(mf);
return NULL;
}
if (ztr->header.version_major != ZTR_VERSION_MAJOR) {
if (!z)
delete_ztr(ztr);
mrewind(mf);
return NULL;
}
}
/* Load chunks */
pos = mftell(mf);
while ((chunk = ztr_read_chunk_hdr(mf))) {
chunk->data = (char *)xmalloc(chunk->dlength);
if (chunk->dlength != mfread(chunk->data, 1, chunk->dlength, mf))
break;
ztr->nchunks++;
ztr->chunk = (ztr_chunk_t *)xrealloc(ztr->chunk, ztr->nchunks *
sizeof(ztr_chunk_t));
memcpy(&ztr->chunk[ztr->nchunks-1], chunk, sizeof(*chunk));
xfree(chunk);
pos = mftell(mf);
}
/*
* At this stage we're 'pos' into the mFILE mf with any remainder being
* a partial block.
*/
if (0 == ztr->nchunks) {
if (!z)
delete_ztr(ztr);
mrewind(mf);
return NULL;
}
/* Ensure we exit at the start of a ztr CHUNK */
mfseek(mf, pos, SEEK_SET);
/* If this is the header part, ensure we uncompress and init. data */
if (!z) {
/* Force caching of huffman code_sets */
ztr_find_hcode(ztr, CODE_USER);
/* And uncompress the rest */
uncompress_ztr(ztr);
}
return ztr;
}
/*
* 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;
}