| 1 |
/* The MIT License |
| 2 |
|
| 3 |
Copyright (c) 2018- Dana-Farber Cancer Institute |
| 4 |
2009-2018 Broad Institute, Inc. |
| 5 |
2008-2009 Genome Research Ltd. (GRL) |
| 6 |
|
| 7 |
Permission is hereby granted, free of charge, to any person obtaining |
| 8 |
a copy of this software and associated documentation files (the |
| 9 |
"Software"), to deal in the Software without restriction, including |
| 10 |
without limitation the rights to use, copy, modify, merge, publish, |
| 11 |
distribute, sublicense, and/or sell copies of the Software, and to |
| 12 |
permit persons to whom the Software is furnished to do so, subject to |
| 13 |
the following conditions: |
| 14 |
|
| 15 |
The above copyright notice and this permission notice shall be |
| 16 |
included in all copies or substantial portions of the Software. |
| 17 |
|
| 18 |
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, |
| 19 |
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF |
| 20 |
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND |
| 21 |
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS |
| 22 |
BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN |
| 23 |
ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN |
| 24 |
CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE |
| 25 |
SOFTWARE. |
| 26 |
*/ |
| 27 |
#include <string.h> |
| 28 |
#include <stdio.h> |
| 29 |
#include <zlib.h> |
| 30 |
#include <assert.h> |
| 31 |
#include "bntseq.h" |
| 32 |
#include "bwa.h" |
| 33 |
#include "ksw.h" |
| 34 |
#include "utils.h" |
| 35 |
#include "kstring.h" |
| 36 |
#include "kvec.h" |
| 37 |
|
| 38 |
#ifdef USE_MALLOC_WRAPPERS |
| 39 |
# include "malloc_wrap.h" |
| 40 |
#endif |
| 41 |
|
| 42 |
int bwa_verbose = 3; |
| 43 |
int bwa_dbg = 0; |
| 44 |
char bwa_rg_id[256]; |
| 45 |
char *bwa_pg; |
| 46 |
|
| 47 |
/************************ |
| 48 |
* Batch FASTA/Q reader * |
| 49 |
************************/ |
| 50 |
|
| 51 |
#include "kseq.h" |
| 52 |
KSEQ_DECLARE(gzFile) |
| 53 |
|
| 54 |
static inline void trim_readno(kstring_t *s) |
| 55 |
{ |
| 56 |
if (s->l > 2 && s->s[s->l-2] == '/' && isdigit(s->s[s->l-1])) |
| 57 |
s->l -= 2, s->s[s->l] = 0; |
| 58 |
} |
| 59 |
|
| 60 |
static inline char *dupkstring(const kstring_t *str, int dupempty) |
| 61 |
{ |
| 62 |
char *s = (str->l > 0 || dupempty)? malloc(str->l + 1) : NULL; |
| 63 |
if (!s) return NULL; |
| 64 |
|
| 65 |
memcpy(s, str->s, str->l); |
| 66 |
s[str->l] = '\0'; |
| 67 |
return s; |
| 68 |
} |
| 69 |
|
| 70 |
static inline void kseq2bseq1(const kseq_t *ks, bseq1_t *s) |
| 71 |
{ // TODO: it would be better to allocate one chunk of memory, but probably it does not matter in practice |
| 72 |
s->name = dupkstring(&ks->name, 1); |
| 73 |
s->comment = dupkstring(&ks->comment, 0); |
| 74 |
s->seq = dupkstring(&ks->seq, 1); |
| 75 |
s->qual = dupkstring(&ks->qual, 0); |
| 76 |
s->l_seq = ks->seq.l; |
| 77 |
} |
| 78 |
|
| 79 |
bseq1_t *bseq_read(int chunk_size, int *n_, void *ks1_, void *ks2_) |
| 80 |
{ |
| 81 |
kseq_t *ks = (kseq_t*)ks1_, *ks2 = (kseq_t*)ks2_; |
| 82 |
int size = 0, m, n; |
| 83 |
bseq1_t *seqs; |
| 84 |
m = n = 0; seqs = 0; |
| 85 |
while (kseq_read(ks) >= 0) { |
| 86 |
if (ks2 && kseq_read(ks2) < 0) { // the 2nd file has fewer reads |
| 87 |
fprintf(stderr, "[W::%s] the 2nd file has fewer sequences.\n", __func__); |
| 88 |
break; |
| 89 |
} |
| 90 |
if (n >= m) { |
| 91 |
m = m? m<<1 : 256; |
| 92 |
seqs = realloc(seqs, m * sizeof(bseq1_t)); |
| 93 |
} |
| 94 |
trim_readno(&ks->name); |
| 95 |
kseq2bseq1(ks, &seqs[n]); |
| 96 |
seqs[n].id = n; |
| 97 |
size += seqs[n++].l_seq; |
| 98 |
if (ks2) { |
| 99 |
trim_readno(&ks2->name); |
| 100 |
kseq2bseq1(ks2, &seqs[n]); |
| 101 |
seqs[n].id = n; |
| 102 |
size += seqs[n++].l_seq; |
| 103 |
} |
| 104 |
if (size >= chunk_size && (n&1) == 0) break; |
| 105 |
} |
| 106 |
if (size == 0) { // test if the 2nd file is finished |
| 107 |
if (ks2 && kseq_read(ks2) >= 0) |
| 108 |
fprintf(stderr, "[W::%s] the 1st file has fewer sequences.\n", __func__); |
| 109 |
} |
| 110 |
*n_ = n; |
| 111 |
return seqs; |
| 112 |
} |
| 113 |
|
| 114 |
void bseq_classify(int n, bseq1_t *seqs, int m[2], bseq1_t *sep[2]) |
| 115 |
{ |
| 116 |
int i, has_last; |
| 117 |
kvec_t(bseq1_t) a[2] = {{0,0,0}, {0,0,0}}; |
| 118 |
for (i = 1, has_last = 1; i < n; ++i) { |
| 119 |
if (has_last) { |
| 120 |
if (strcmp(seqs[i].name, seqs[i-1].name) == 0) { |
| 121 |
kv_push(bseq1_t, a[1], seqs[i-1]); |
| 122 |
kv_push(bseq1_t, a[1], seqs[i]); |
| 123 |
has_last = 0; |
| 124 |
} else kv_push(bseq1_t, a[0], seqs[i-1]); |
| 125 |
} else has_last = 1; |
| 126 |
} |
| 127 |
if (has_last) kv_push(bseq1_t, a[0], seqs[i-1]); |
| 128 |
sep[0] = a[0].a, m[0] = a[0].n; |
| 129 |
sep[1] = a[1].a, m[1] = a[1].n; |
| 130 |
} |
| 131 |
|
| 132 |
/***************** |
| 133 |
* CIGAR related * |
| 134 |
*****************/ |
| 135 |
|
| 136 |
void bwa_fill_scmat(int a, int b, int8_t mat[25]) |
| 137 |
{ |
| 138 |
int i, j, k; |
| 139 |
for (i = k = 0; i < 4; ++i) { |
| 140 |
for (j = 0; j < 4; ++j) |
| 141 |
mat[k++] = i == j? a : -b; |
| 142 |
mat[k++] = -1; // ambiguous base |
| 143 |
} |
| 144 |
for (j = 0; j < 5; ++j) mat[k++] = -1; |
| 145 |
} |
| 146 |
|
| 147 |
// Generate CIGAR when the alignment end points are known |
| 148 |
uint32_t *bwa_gen_cigar2(const int8_t mat[25], int o_del, int e_del, int o_ins, int e_ins, int w_, int64_t l_pac, const uint8_t *pac, int l_query, uint8_t *query, int64_t rb, int64_t re, int *score, int *n_cigar, int *NM) |
| 149 |
{ |
| 150 |
uint32_t *cigar = 0; |
| 151 |
uint8_t tmp, *rseq; |
| 152 |
int i; |
| 153 |
int64_t rlen; |
| 154 |
kstring_t str; |
| 155 |
const char *int2base; |
| 156 |
|
| 157 |
if (n_cigar) *n_cigar = 0; |
| 158 |
if (NM) *NM = -1; |
| 159 |
if (l_query <= 0 || rb >= re || (rb < l_pac && re > l_pac)) return 0; // reject if negative length or bridging the forward and reverse strand |
| 160 |
rseq = bns_get_seq(l_pac, pac, rb, re, &rlen); |
| 161 |
if (re - rb != rlen) goto ret_gen_cigar; // possible if out of range |
| 162 |
if (rb >= l_pac) { // then reverse both query and rseq; this is to ensure indels to be placed at the leftmost position |
| 163 |
for (i = 0; i < l_query>>1; ++i) |
| 164 |
tmp = query[i], query[i] = query[l_query - 1 - i], query[l_query - 1 - i] = tmp; |
| 165 |
for (i = 0; i < rlen>>1; ++i) |
| 166 |
tmp = rseq[i], rseq[i] = rseq[rlen - 1 - i], rseq[rlen - 1 - i] = tmp; |
| 167 |
} |
| 168 |
if (l_query == re - rb && w_ == 0) { // no gap; no need to do DP |
| 169 |
// UPDATE: we come to this block now... FIXME: due to an issue in mem_reg2aln(), we never come to this block. This does not affect accuracy, but it hurts performance. |
| 170 |
if (n_cigar) { |
| 171 |
cigar = malloc(4); |
| 172 |
cigar[0] = l_query<<4 | 0; |
| 173 |
*n_cigar = 1; |
| 174 |
} |
| 175 |
for (i = 0, *score = 0; i < l_query; ++i) |
| 176 |
*score += mat[rseq[i]*5 + query[i]]; |
| 177 |
} else { |
| 178 |
int w, max_gap, max_ins, max_del, min_w; |
| 179 |
// set the band-width |
| 180 |
max_ins = (int)((double)(((l_query+1)>>1) * mat[0] - o_ins) / e_ins + 1.); |
| 181 |
max_del = (int)((double)(((l_query+1)>>1) * mat[0] - o_del) / e_del + 1.); |
| 182 |
max_gap = max_ins > max_del? max_ins : max_del; |
| 183 |
max_gap = max_gap > 1? max_gap : 1; |
| 184 |
w = (max_gap + abs((int)rlen - l_query) + 1) >> 1; |
| 185 |
w = w < w_? w : w_; |
| 186 |
min_w = abs((int)rlen - l_query) + 3; |
| 187 |
w = w > min_w? w : min_w; |
| 188 |
// NW alignment |
| 189 |
if (bwa_verbose >= 4) { |
| 190 |
printf("* Global bandwidth: %d\n", w); |
| 191 |
printf("* Global ref: "); for (i = 0; i < rlen; ++i) putchar("ACGTN"[(int)rseq[i]]); putchar('\n'); |
| 192 |
printf("* Global query: "); for (i = 0; i < l_query; ++i) putchar("ACGTN"[(int)query[i]]); putchar('\n'); |
| 193 |
} |
| 194 |
*score = ksw_global2(l_query, query, rlen, rseq, 5, mat, o_del, e_del, o_ins, e_ins, w, n_cigar, &cigar); |
| 195 |
} |
| 196 |
if (NM && n_cigar) {// compute NM and MD |
| 197 |
int k, x, y, u, n_mm = 0, n_gap = 0; |
| 198 |
str.l = str.m = *n_cigar * 4; str.s = (char*)cigar; // append MD to CIGAR |
| 199 |
int2base = rb < l_pac? "ACGTN" : "TGCAN"; |
| 200 |
for (k = 0, x = y = u = 0; k < *n_cigar; ++k) { |
| 201 |
int op, len; |
| 202 |
cigar = (uint32_t*)str.s; |
| 203 |
op = cigar[k]&0xf, len = cigar[k]>>4; |
| 204 |
if (op == 0) { // match |
| 205 |
for (i = 0; i < len; ++i) { |
| 206 |
if (query[x + i] != rseq[y + i]) { |
| 207 |
kputw(u, &str); |
| 208 |
kputc(int2base[rseq[y+i]], &str); |
| 209 |
++n_mm; u = 0; |
| 210 |
} else ++u; |
| 211 |
} |
| 212 |
x += len; y += len; |
| 213 |
} else if (op == 2) { // deletion |
| 214 |
if (k > 0 && k < *n_cigar - 1) { // don't do the following if D is the first or the last CIGAR |
| 215 |
kputw(u, &str); kputc('^', &str); |
| 216 |
for (i = 0; i < len; ++i) |
| 217 |
kputc(int2base[rseq[y+i]], &str); |
| 218 |
u = 0; n_gap += len; |
| 219 |
} |
| 220 |
y += len; |
| 221 |
} else if (op == 1) x += len, n_gap += len; // insertion |
| 222 |
} |
| 223 |
kputw(u, &str); kputc(0, &str); |
| 224 |
*NM = n_mm + n_gap; |
| 225 |
cigar = (uint32_t*)str.s; |
| 226 |
} |
| 227 |
if (rb >= l_pac) // reverse back query |
| 228 |
for (i = 0; i < l_query>>1; ++i) |
| 229 |
tmp = query[i], query[i] = query[l_query - 1 - i], query[l_query - 1 - i] = tmp; |
| 230 |
|
| 231 |
ret_gen_cigar: |
| 232 |
free(rseq); |
| 233 |
return cigar; |
| 234 |
} |
| 235 |
|
| 236 |
uint32_t *bwa_gen_cigar(const int8_t mat[25], int q, int r, int w_, int64_t l_pac, const uint8_t *pac, int l_query, uint8_t *query, int64_t rb, int64_t re, int *score, int *n_cigar, int *NM) |
| 237 |
{ |
| 238 |
return bwa_gen_cigar2(mat, q, r, q, r, w_, l_pac, pac, l_query, query, rb, re, score, n_cigar, NM); |
| 239 |
} |
| 240 |
|
| 241 |
/********************* |
| 242 |
* Full index reader * |
| 243 |
*********************/ |
| 244 |
|
| 245 |
char *bwa_idx_infer_prefix(const char *hint) |
| 246 |
{ |
| 247 |
char *prefix; |
| 248 |
int l_hint; |
| 249 |
FILE *fp; |
| 250 |
l_hint = strlen(hint); |
| 251 |
prefix = malloc(l_hint + 3 + 4 + 1); |
| 252 |
strcpy(prefix, hint); |
| 253 |
strcpy(prefix + l_hint, ".64.bwt"); |
| 254 |
if ((fp = fopen(prefix, "rb")) != 0) { |
| 255 |
fclose(fp); |
| 256 |
prefix[l_hint + 3] = 0; |
| 257 |
return prefix; |
| 258 |
} else { |
| 259 |
strcpy(prefix + l_hint, ".bwt"); |
| 260 |
if ((fp = fopen(prefix, "rb")) == 0) { |
| 261 |
free(prefix); |
| 262 |
return 0; |
| 263 |
} else { |
| 264 |
fclose(fp); |
| 265 |
prefix[l_hint] = 0; |
| 266 |
return prefix; |
| 267 |
} |
| 268 |
} |
| 269 |
} |
| 270 |
|
| 271 |
bwt_t *bwa_idx_load_bwt(const char *hint) |
| 272 |
{ |
| 273 |
char *tmp, *prefix; |
| 274 |
bwt_t *bwt; |
| 275 |
prefix = bwa_idx_infer_prefix(hint); |
| 276 |
if (prefix == 0) { |
| 277 |
if (bwa_verbose >= 1) fprintf(stderr, "[E::%s] fail to locate the index files\n", __func__); |
| 278 |
return 0; |
| 279 |
} |
| 280 |
tmp = calloc(strlen(prefix) + 5, 1); |
| 281 |
strcat(strcpy(tmp, prefix), ".bwt"); // FM-index |
| 282 |
bwt = bwt_restore_bwt(tmp); |
| 283 |
strcat(strcpy(tmp, prefix), ".sa"); // partial suffix array (SA) |
| 284 |
bwt_restore_sa(tmp, bwt); |
| 285 |
free(tmp); free(prefix); |
| 286 |
return bwt; |
| 287 |
} |
| 288 |
|
| 289 |
bwaidx_t *bwa_idx_load_from_disk(const char *hint, int which) |
| 290 |
{ |
| 291 |
bwaidx_t *idx; |
| 292 |
char *prefix; |
| 293 |
prefix = bwa_idx_infer_prefix(hint); |
| 294 |
if (prefix == 0) { |
| 295 |
if (bwa_verbose >= 1) fprintf(stderr, "[E::%s] fail to locate the index files\n", __func__); |
| 296 |
return 0; |
| 297 |
} |
| 298 |
idx = calloc(1, sizeof(bwaidx_t)); |
| 299 |
if (which & BWA_IDX_BWT) idx->bwt = bwa_idx_load_bwt(hint); |
| 300 |
if (which & BWA_IDX_BNS) { |
| 301 |
int i, c; |
| 302 |
idx->bns = bns_restore(prefix); |
| 303 |
for (i = c = 0; i < idx->bns->n_seqs; ++i) |
| 304 |
if (idx->bns->anns[i].is_alt) ++c; |
| 305 |
if (bwa_verbose >= 3) |
| 306 |
fprintf(stderr, "[M::%s] read %d ALT contigs\n", __func__, c); |
| 307 |
if (which & BWA_IDX_PAC) { |
| 308 |
idx->pac = calloc(idx->bns->l_pac/4+1, 1); |
| 309 |
err_fread_noeof(idx->pac, 1, idx->bns->l_pac/4+1, idx->bns->fp_pac); // concatenated 2-bit encoded sequence |
| 310 |
err_fclose(idx->bns->fp_pac); |
| 311 |
idx->bns->fp_pac = 0; |
| 312 |
} |
| 313 |
} |
| 314 |
free(prefix); |
| 315 |
return idx; |
| 316 |
} |
| 317 |
|
| 318 |
bwaidx_t *bwa_idx_load(const char *hint, int which) |
| 319 |
{ |
| 320 |
return bwa_idx_load_from_disk(hint, which); |
| 321 |
} |
| 322 |
|
| 323 |
void bwa_idx_destroy(bwaidx_t *idx) |
| 324 |
{ |
| 325 |
if (idx == 0) return; |
| 326 |
if (idx->mem == 0) { |
| 327 |
if (idx->bwt) bwt_destroy(idx->bwt); |
| 328 |
if (idx->bns) bns_destroy(idx->bns); |
| 329 |
if (idx->pac) free(idx->pac); |
| 330 |
} else { |
| 331 |
free(idx->bwt); free(idx->bns->anns); free(idx->bns); |
| 332 |
if (!idx->is_shm) free(idx->mem); |
| 333 |
} |
| 334 |
free(idx); |
| 335 |
} |
| 336 |
|
| 337 |
int bwa_mem2idx(int64_t l_mem, uint8_t *mem, bwaidx_t *idx) |
| 338 |
{ |
| 339 |
int64_t k = 0, x; |
| 340 |
int i; |
| 341 |
|
| 342 |
// generate idx->bwt |
| 343 |
x = sizeof(bwt_t); idx->bwt = malloc(x); memcpy(idx->bwt, mem + k, x); k += x; |
| 344 |
x = idx->bwt->bwt_size * 4; idx->bwt->bwt = (uint32_t*)(mem + k); k += x; |
| 345 |
x = idx->bwt->n_sa * sizeof(bwtint_t); idx->bwt->sa = (bwtint_t*)(mem + k); k += x; |
| 346 |
|
| 347 |
// generate idx->bns and idx->pac |
| 348 |
x = sizeof(bntseq_t); idx->bns = malloc(x); memcpy(idx->bns, mem + k, x); k += x; |
| 349 |
x = idx->bns->n_holes * sizeof(bntamb1_t); idx->bns->ambs = (bntamb1_t*)(mem + k); k += x; |
| 350 |
x = idx->bns->n_seqs * sizeof(bntann1_t); idx->bns->anns = malloc(x); memcpy(idx->bns->anns, mem + k, x); k += x; |
| 351 |
for (i = 0; i < idx->bns->n_seqs; ++i) { |
| 352 |
idx->bns->anns[i].name = (char*)(mem + k); k += strlen(idx->bns->anns[i].name) + 1; |
| 353 |
idx->bns->anns[i].anno = (char*)(mem + k); k += strlen(idx->bns->anns[i].anno) + 1; |
| 354 |
} |
| 355 |
idx->pac = (uint8_t*)(mem + k); k += idx->bns->l_pac/4+1; |
| 356 |
assert(k == l_mem); |
| 357 |
|
| 358 |
idx->l_mem = k; idx->mem = mem; |
| 359 |
return 0; |
| 360 |
} |
| 361 |
|
| 362 |
int bwa_idx2mem(bwaidx_t *idx) |
| 363 |
{ |
| 364 |
int i; |
| 365 |
int64_t k, x, tmp; |
| 366 |
uint8_t *mem; |
| 367 |
|
| 368 |
// copy idx->bwt |
| 369 |
x = idx->bwt->bwt_size * 4; |
| 370 |
mem = realloc(idx->bwt->bwt, sizeof(bwt_t) + x); idx->bwt->bwt = 0; |
| 371 |
memmove(mem + sizeof(bwt_t), mem, x); |
| 372 |
memcpy(mem, idx->bwt, sizeof(bwt_t)); k = sizeof(bwt_t) + x; |
| 373 |
x = idx->bwt->n_sa * sizeof(bwtint_t); mem = realloc(mem, k + x); memcpy(mem + k, idx->bwt->sa, x); k += x; |
| 374 |
free(idx->bwt->sa); |
| 375 |
free(idx->bwt); idx->bwt = 0; |
| 376 |
|
| 377 |
// copy idx->bns |
| 378 |
tmp = idx->bns->n_seqs * sizeof(bntann1_t) + idx->bns->n_holes * sizeof(bntamb1_t); |
| 379 |
for (i = 0; i < idx->bns->n_seqs; ++i) // compute the size of heap-allocated memory |
| 380 |
tmp += strlen(idx->bns->anns[i].name) + strlen(idx->bns->anns[i].anno) + 2; |
| 381 |
mem = realloc(mem, k + sizeof(bntseq_t) + tmp); |
| 382 |
x = sizeof(bntseq_t); memcpy(mem + k, idx->bns, x); k += x; |
| 383 |
x = idx->bns->n_holes * sizeof(bntamb1_t); memcpy(mem + k, idx->bns->ambs, x); k += x; |
| 384 |
free(idx->bns->ambs); |
| 385 |
x = idx->bns->n_seqs * sizeof(bntann1_t); memcpy(mem + k, idx->bns->anns, x); k += x; |
| 386 |
for (i = 0; i < idx->bns->n_seqs; ++i) { |
| 387 |
x = strlen(idx->bns->anns[i].name) + 1; memcpy(mem + k, idx->bns->anns[i].name, x); k += x; |
| 388 |
x = strlen(idx->bns->anns[i].anno) + 1; memcpy(mem + k, idx->bns->anns[i].anno, x); k += x; |
| 389 |
free(idx->bns->anns[i].name); free(idx->bns->anns[i].anno); |
| 390 |
} |
| 391 |
free(idx->bns->anns); |
| 392 |
|
| 393 |
// copy idx->pac |
| 394 |
x = idx->bns->l_pac/4+1; |
| 395 |
mem = realloc(mem, k + x); |
| 396 |
memcpy(mem + k, idx->pac, x); k += x; |
| 397 |
free(idx->bns); idx->bns = 0; |
| 398 |
free(idx->pac); idx->pac = 0; |
| 399 |
|
| 400 |
return bwa_mem2idx(k, mem, idx); |
| 401 |
} |
| 402 |
|
| 403 |
/*********************** |
| 404 |
* SAM header routines * |
| 405 |
***********************/ |
| 406 |
|
| 407 |
void bwa_print_sam_hdr(const bntseq_t *bns, const char *hdr_line) |
| 408 |
{ |
| 409 |
int i, n_HD = 0, n_SQ = 0; |
| 410 |
extern char *bwa_pg; |
| 411 |
|
| 412 |
if (hdr_line) { |
| 413 |
// check for HD line |
| 414 |
const char *p = hdr_line; |
| 415 |
if ((p = strstr(p, "@HD")) != 0) { |
| 416 |
++n_HD; |
| 417 |
} |
| 418 |
// check for SQ lines |
| 419 |
p = hdr_line; |
| 420 |
while ((p = strstr(p, "@SQ\t")) != 0) { |
| 421 |
if (p == hdr_line || *(p-1) == '\n') ++n_SQ; |
| 422 |
p += 4; |
| 423 |
} |
| 424 |
} |
| 425 |
if (n_SQ == 0) { |
| 426 |
for (i = 0; i < bns->n_seqs; ++i) { |
| 427 |
err_printf("@SQ\tSN:%s\tLN:%d", bns->anns[i].name, bns->anns[i].len); |
| 428 |
if (bns->anns[i].is_alt) err_printf("\tAH:*\n"); |
| 429 |
else err_fputc('\n', stdout); |
| 430 |
} |
| 431 |
} else if (n_SQ != bns->n_seqs && bwa_verbose >= 2) |
| 432 |
fprintf(stderr, "[W::%s] %d @SQ lines provided with -H; %d sequences in the index. Continue anyway.\n", __func__, n_SQ, bns->n_seqs); |
| 433 |
if (n_HD == 0) { |
| 434 |
err_printf("@HD\tVN:1.5\tSO:unsorted\tGO:query\n"); |
| 435 |
} |
| 436 |
if (hdr_line) err_printf("%s\n", hdr_line); |
| 437 |
if (bwa_pg) err_printf("%s\n", bwa_pg); |
| 438 |
} |
| 439 |
|
| 440 |
static char *bwa_escape(char *s) |
| 441 |
{ |
| 442 |
char *p, *q; |
| 443 |
for (p = q = s; *p; ++p) { |
| 444 |
if (*p == '\\') { |
| 445 |
++p; |
| 446 |
if (*p == 't') *q++ = '\t'; |
| 447 |
else if (*p == 'n') *q++ = '\n'; |
| 448 |
else if (*p == 'r') *q++ = '\r'; |
| 449 |
else if (*p == '\\') *q++ = '\\'; |
| 450 |
} else *q++ = *p; |
| 451 |
} |
| 452 |
*q = '\0'; |
| 453 |
return s; |
| 454 |
} |
| 455 |
|
| 456 |
char *bwa_set_rg(const char *s) |
| 457 |
{ |
| 458 |
char *p, *q, *r, *rg_line = 0; |
| 459 |
memset(bwa_rg_id, 0, 256); |
| 460 |
if (strstr(s, "@RG") != s) { |
| 461 |
if (bwa_verbose >= 1) fprintf(stderr, "[E::%s] the read group line is not started with @RG\n", __func__); |
| 462 |
goto err_set_rg; |
| 463 |
} |
| 464 |
if (strstr(s, "\t") != NULL) { |
| 465 |
if (bwa_verbose >= 1) fprintf(stderr, "[E::%s] the read group line contained literal <tab> characters -- replace with escaped tabs: \\t\n", __func__); |
| 466 |
goto err_set_rg; |
| 467 |
} |
| 468 |
rg_line = strdup(s); |
| 469 |
bwa_escape(rg_line); |
| 470 |
if ((p = strstr(rg_line, "\tID:")) == 0) { |
| 471 |
if (bwa_verbose >= 1) fprintf(stderr, "[E::%s] no ID within the read group line\n", __func__); |
| 472 |
goto err_set_rg; |
| 473 |
} |
| 474 |
p += 4; |
| 475 |
for (q = p; *q && *q != '\t' && *q != '\n'; ++q); |
| 476 |
if (q - p + 1 > 256) { |
| 477 |
if (bwa_verbose >= 1) fprintf(stderr, "[E::%s] @RG:ID is longer than 255 characters\n", __func__); |
| 478 |
goto err_set_rg; |
| 479 |
} |
| 480 |
for (q = p, r = bwa_rg_id; *q && *q != '\t' && *q != '\n'; ++q) |
| 481 |
*r++ = *q; |
| 482 |
return rg_line; |
| 483 |
|
| 484 |
err_set_rg: |
| 485 |
free(rg_line); |
| 486 |
return 0; |
| 487 |
} |
| 488 |
|
| 489 |
char *bwa_insert_header(const char *s, char *hdr) |
| 490 |
{ |
| 491 |
int len = 0; |
| 492 |
if (s == 0 || s[0] != '@') return hdr; |
| 493 |
if (hdr) { |
| 494 |
len = strlen(hdr); |
| 495 |
hdr = realloc(hdr, len + strlen(s) + 2); |
| 496 |
hdr[len++] = '\n'; |
| 497 |
strcpy(hdr + len, s); |
| 498 |
} else hdr = strdup(s); |
| 499 |
bwa_escape(hdr + len); |
| 500 |
return hdr; |
| 501 |
} |