zzh
/
minimap2
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

添加了一些统计运行时间的变量，测试性能

This commit is contained in:
zzh 2023-08-03 03:12:38 +08:00
parent 1641a66d15
commit 023c18ba32
10 changed files with 988 additions and 456 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
.vscode/extensions.json vendored 100644
View File

@ -0,0 +1,5 @@
{
"recommendations": [
"mhutchie.git-graph"
]
}

16
.vscode/launch.json vendored
View File

@ -10,12 +10,16 @@
"type": "cppdbg", "type": "cppdbg",
"request": "launch", "request": "launch",
"program": "${workspaceRoot}/minimap2", "program": "${workspaceRoot}/minimap2",
"args": ["-x", "args": [
"ava-ont", "-x",
"-t", "ava-ont",
"1", "-t",
"/public/home/zzh/work/3gseq/TGM-2021YFF/Acinetobacter_pittii.fastq", "1",
"/public/home/zzh/work/3gseq/TGM-2021YFF/Acinetobacter_pittii.fastq"], "/public/home/zzh/work/3gseq/TGM-2021YFF/Acinetobacter_pittii.fastq",
"/public/home/zzh/work/3gseq/TGM-2021YFF/Acinetobacter_pittii.fastq",
"-o",
"reads.paf"
],
"cwd": "${workspaceFolder}", // 当前工作路径：当前文件所在的工作空间 "cwd": "${workspaceFolder}", // 当前工作路径：当前文件所在的工作空间
} }
] ]

6
.vscode/settings.json vendored 100644
View File

@ -0,0 +1,6 @@
{
"files.associations": {
"minimap.h": "c",
"time.h": "c"
}
}

3
Makefile
View File

@ -1,4 +1,5 @@
CFLAGS= -g -Wall -O2 -Wc++-compat #-Wextra CFLAGS= -g -Wall -Wc++-compat #-Wextra
#CFLAGS= -g -Wall -O2 -Wc++-compat #-Wextra
CPPFLAGS= -DHAVE_KALLOC CPPFLAGS= -DHAVE_KALLOC
INCLUDES= INCLUDES=
OBJS= kthread.o kalloc.o misc.o bseq.o sketch.o sdust.o options.o index.o \ OBJS= kthread.o kalloc.o misc.o bseq.o sketch.o sdust.o options.o index.o \

11
index.c
View File

@ -15,6 +15,11 @@
#include "kvec.h" #include "kvec.h"
#include "khash.h" #include "khash.h"
#ifdef ANALYSIS_PERF
extern int64_t get_mseconds();
extern int64_t time_mm_idx_reader_read;
#endif
#define idx_hash(a) ((a)>>1) #define idx_hash(a) ((a)>>1)
#define idx_eq(a, b) ((a)>>1 == (b)>>1) #define idx_eq(a, b) ((a)>>1 == (b)>>1)
KHASH_INIT(idx, uint64_t, uint64_t, 1, idx_hash, idx_eq) KHASH_INIT(idx, uint64_t, uint64_t, 1, idx_hash, idx_eq)
@ -608,6 +613,9 @@ void mm_idx_reader_close(mm_idx_reader_t *r)
mm_idx_t *mm_idx_reader_read(mm_idx_reader_t *r, int n_threads) mm_idx_t *mm_idx_reader_read(mm_idx_reader_t *r, int n_threads)
{ {
mm_idx_t *mi; mm_idx_t *mi;
#ifdef ANALYSIS_PERF
int64_t tmp_cur_time = get_mseconds();
#endif
if (r->is_idx) { if (r->is_idx) {
mi = mm_idx_load(r->fp.idx); mi = mm_idx_load(r->fp.idx);
if (mi && mm_verbose >= 2 && (mi->k != r->opt.k || mi->w != r->opt.w || (mi->flag&MM_I_HPC) != (r->opt.flag&MM_I_HPC))) if (mi && mm_verbose >= 2 && (mi->k != r->opt.k || mi->w != r->opt.w || (mi->flag&MM_I_HPC) != (r->opt.flag&MM_I_HPC)))
@ -618,6 +626,9 @@ mm_idx_t *mm_idx_reader_read(mm_idx_reader_t *r, int n_threads)
if (r->fp_out) mm_idx_dump(r->fp_out, mi); if (r->fp_out) mm_idx_dump(r->fp_out, mi);
mi->index = r->n_parts++; mi->index = r->n_parts++;
} }
#ifdef ANALYSIS_PERF
time_mm_idx_reader_read += get_mseconds() - tmp_cur_time;
#endif
return mi; return mi;
} }

22
lchain.c
View File

@ -5,7 +5,11 @@
#include "mmpriv.h" #include "mmpriv.h"
#include "kalloc.h" #include "kalloc.h"
#include "krmq.h" #include "krmq.h"
#ifdef ANALYSIS_PERF
extern int64_t get_mseconds();
extern int64_t time_mg_lchain_dp,
time_mg_chain_backtrack;
#endif
static int64_t mg_chain_bk_end(int32_t max_drop, const mm128_t *z, const int32_t *f, const int64_t *p, int32_t *t, int64_t k) static int64_t mg_chain_bk_end(int32_t max_drop, const mm128_t *z, const int32_t *f, const int64_t *p, int32_t *t, int64_t k)
{ {
int64_t i = z[k].y, end_i = -1, max_i = i; int64_t i = z[k].y, end_i = -1, max_i = i;
@ -151,7 +155,9 @@ mm128_t *mg_lchain_dp(int max_dist_x, int max_dist_y, int bw, int max_skip, int
int32_t *f, *t, *v, n_u, n_v, mmax_f = 0, max_drop = bw; int32_t *f, *t, *v, n_u, n_v, mmax_f = 0, max_drop = bw;
int64_t *p, i, j, max_ii, st = 0, n_iter = 0; int64_t *p, i, j, max_ii, st = 0, n_iter = 0;
uint64_t *u; uint64_t *u;
#ifdef ANALYSIS_PERF
int64_t tmp_cur_time = get_mseconds(), tmp_diff = 0;
#endif
if (_u) *_u = 0, *n_u_ = 0; if (_u) *_u = 0, *n_u_ = 0;
if (n == 0 || a == 0) { if (n == 0 || a == 0) {
kfree(km, a); kfree(km, a);
@ -205,10 +211,20 @@ mm128_t *mg_lchain_dp(int max_dist_x, int max_dist_y, int bw, int max_skip, int
max_ii = i; max_ii = i;
if (mmax_f < max_f) mmax_f = max_f; if (mmax_f < max_f) mmax_f = max_f;
} }
#ifdef ANALYSIS_PERF
int64_t tmp_inner_time = get_mseconds();
#endif
u = mg_chain_backtrack(km, n, f, p, v, t, min_cnt, min_sc, max_drop, &n_u, &n_v); u = mg_chain_backtrack(km, n, f, p, v, t, min_cnt, min_sc, max_drop, &n_u, &n_v);
#ifdef ANALYSIS_PERF
tmp_diff = get_mseconds() - tmp_inner_time;
__sync_fetch_and_add(&time_mg_chain_backtrack, tmp_diff);
#endif
*n_u_ = n_u, *_u = u; // NB: note that u[] may not be sorted by score here *n_u_ = n_u, *_u = u; // NB: note that u[] may not be sorted by score here
kfree(km, p); kfree(km, f); kfree(km, t); kfree(km, p); kfree(km, f); kfree(km, t);
#ifdef ANALYSIS_PERF
tmp_diff = get_mseconds() - tmp_cur_time;
__sync_fetch_and_add(&time_mg_lchain_dp, tmp_diff);
#endif
if (n_u == 0) { if (n_u == 0) {
kfree(km, a); kfree(km, v); kfree(km, a); kfree(km, v);
return 0; return 0;

727
main.c
View File

@ -7,6 +7,38 @@
#include "mmpriv.h" #include "mmpriv.h"
#include "ketopt.h" #include "ketopt.h"
// 用来调试，计算感兴趣部分的运行时间
#include "sys/time.h"
// 获取当前毫秒数
int64_t get_mseconds()
{
struct timeval tv;
gettimeofday(&tv, NULL);
return (int64_t)1000 * (tv.tv_sec + ((1e-6) * tv.tv_usec));
}
// 记录运行时间的变量
#ifdef ANALYSIS_PERF
int64_t time_mm_idx_reader_read,
time_mm_map_file_frag,
time_map_work_for_block_1,
time_map_work_for_block_2,
time_map_work_for_block_3,
time_mm_map_frag_b1,
time_mm_map_frag_b2,
time_mm_map_frag_b3,
time_mm_map_frag_b4,
time_mm_map_frag_b5,
time_mm_map_frag_b6,
time_mg_lchain_dp = 0,
time_collect_seed_hits_heap = 0,
time_collect_seed_hits = 0,
time_mg_chain_backtrack = 0;
#endif
//////////////////////////////////
#ifdef __linux__ #ifdef __linux__
#include <sys/resource.h> #include <sys/resource.h>
#include <sys/time.h> #include <sys/time.h>
@ -18,85 +50,90 @@ void liftrlimit()
setrlimit(RLIMIT_AS, &r); setrlimit(RLIMIT_AS, &r);
} }
#else #else
void liftrlimit() {} void liftrlimit()
{
}
#endif #endif
static ko_longopt_t long_options[] = { static ko_longopt_t long_options[] = {
{ "bucket-bits", ko_required_argument, 300 }, {"bucket-bits", ko_required_argument, 300},
{ "mb-size", ko_required_argument, 'K' }, {"mb-size", ko_required_argument, 'K'},
{ "seed", ko_required_argument, 302 }, {"seed", ko_required_argument, 302},
{ "no-kalloc", ko_no_argument, 303 }, {"no-kalloc", ko_no_argument, 303},
{ "print-qname", ko_no_argument, 304 }, {"print-qname", ko_no_argument, 304},
{ "no-self", ko_no_argument, 'D' }, {"no-self", ko_no_argument, 'D'},
{ "print-seeds", ko_no_argument, 306 }, {"print-seeds", ko_no_argument, 306},
{ "max-chain-skip", ko_required_argument, 307 }, {"max-chain-skip", ko_required_argument, 307},
{ "min-dp-len", ko_required_argument, 308 }, {"min-dp-len", ko_required_argument, 308},
{ "print-aln-seq", ko_no_argument, 309 }, {"print-aln-seq", ko_no_argument, 309},
{ "splice", ko_no_argument, 310 }, {"splice", ko_no_argument, 310},
{ "cost-non-gt-ag", ko_required_argument, 'C' }, {"cost-non-gt-ag", ko_required_argument, 'C'},
{ "no-long-join", ko_no_argument, 312 }, {"no-long-join", ko_no_argument, 312},
{ "sr", ko_no_argument, 313 }, {"sr", ko_no_argument, 313},
{ "frag", ko_required_argument, 314 }, {"frag", ko_required_argument, 314},
{ "secondary", ko_required_argument, 315 }, {"secondary", ko_required_argument, 315},
{ "cs", ko_optional_argument, 316 }, {"cs", ko_optional_argument, 316},
{ "end-bonus", ko_required_argument, 317 }, {"end-bonus", ko_required_argument, 317},
{ "no-pairing", ko_no_argument, 318 }, {"no-pairing", ko_no_argument, 318},
{ "splice-flank", ko_required_argument, 319 }, {"splice-flank", ko_required_argument, 319},
{ "idx-no-seq", ko_no_argument, 320 }, {"idx-no-seq", ko_no_argument, 320},
{ "end-seed-pen", ko_required_argument, 321 }, {"end-seed-pen", ko_required_argument, 321},
{ "for-only", ko_no_argument, 322 }, {"for-only", ko_no_argument, 322},
{ "rev-only", ko_no_argument, 323 }, {"rev-only", ko_no_argument, 323},
{ "heap-sort", ko_required_argument, 324 }, {"heap-sort", ko_required_argument, 324},
{ "all-chain", ko_no_argument, 'P' }, {"all-chain", ko_no_argument, 'P'},
{ "dual", ko_required_argument, 326 }, {"dual", ko_required_argument, 326},
{ "max-clip-ratio", ko_required_argument, 327 }, {"max-clip-ratio", ko_required_argument, 327},
{ "min-occ-floor", ko_required_argument, 328 }, {"min-occ-floor", ko_required_argument, 328},
{ "MD", ko_no_argument, 329 }, {"MD", ko_no_argument, 329},
{ "lj-min-ratio", ko_required_argument, 330 }, {"lj-min-ratio", ko_required_argument, 330},
{ "score-N", ko_required_argument, 331 }, {"score-N", ko_required_argument, 331},
{ "eqx", ko_no_argument, 332 }, {"eqx", ko_no_argument, 332},
{ "paf-no-hit", ko_no_argument, 333 }, {"paf-no-hit", ko_no_argument, 333},
{ "split-prefix", ko_required_argument, 334 }, {"split-prefix", ko_required_argument, 334},
{ "no-end-flt", ko_no_argument, 335 }, {"no-end-flt", ko_no_argument, 335},
{ "hard-mask-level",ko_no_argument, 336 }, {"hard-mask-level", ko_no_argument, 336},
{ "cap-sw-mem", ko_required_argument, 337 }, {"cap-sw-mem", ko_required_argument, 337},
{ "max-qlen", ko_required_argument, 338 }, {"max-qlen", ko_required_argument, 338},
{ "max-chain-iter", ko_required_argument, 339 }, {"max-chain-iter", ko_required_argument, 339},
{ "junc-bed", ko_required_argument, 340 }, {"junc-bed", ko_required_argument, 340},
{ "junc-bonus", ko_required_argument, 341 }, {"junc-bonus", ko_required_argument, 341},
{ "sam-hit-only", ko_no_argument, 342 }, {"sam-hit-only", ko_no_argument, 342},
{ "chain-gap-scale",ko_required_argument, 343 }, {"chain-gap-scale", ko_required_argument, 343},
{ "alt", ko_required_argument, 344 }, {"alt", ko_required_argument, 344},
{ "alt-drop", ko_required_argument, 345 }, {"alt-drop", ko_required_argument, 345},
{ "mask-len", ko_required_argument, 346 }, {"mask-len", ko_required_argument, 346},
{ "rmq", ko_optional_argument, 347 }, {"rmq", ko_optional_argument, 347},
{ "qstrand", ko_no_argument, 348 }, {"qstrand", ko_no_argument, 348},
{ "cap-kalloc", ko_required_argument, 349 }, {"cap-kalloc", ko_required_argument, 349},
{ "q-occ-frac", ko_required_argument, 350 }, {"q-occ-frac", ko_required_argument, 350},
{ "chain-skip-scale",ko_required_argument,351 }, {"chain-skip-scale", ko_required_argument, 351},
{ "print-chains", ko_no_argument, 352 }, {"print-chains", ko_no_argument, 352},
{ "no-hash-name", ko_no_argument, 353 }, {"no-hash-name", ko_no_argument, 353},
{ "secondary-seq", ko_no_argument, 354 }, {"secondary-seq", ko_no_argument, 354},
{ "help", ko_no_argument, 'h' }, {"help", ko_no_argument, 'h'},
{ "max-intron-len", ko_required_argument, 'G' }, {"max-intron-len", ko_required_argument, 'G'},
{ "version", ko_no_argument, 'V' }, {"version", ko_no_argument, 'V'},
{ "min-count", ko_required_argument, 'n' }, {"min-count", ko_required_argument, 'n'},
{ "min-chain-score",ko_required_argument, 'm' }, {"min-chain-score", ko_required_argument, 'm'},
{ "mask-level", ko_required_argument, 'M' }, {"mask-level", ko_required_argument, 'M'},
{ "min-dp-score", ko_required_argument, 's' }, {"min-dp-score", ko_required_argument, 's'},
{ "sam", ko_no_argument, 'a' }, {"sam", ko_no_argument, 'a'},
{ 0, 0, 0 } {0, 0, 0}};
};
static inline int64_t mm_parse_num2(const char *str, char **q) static inline int64_t mm_parse_num2(const char *str, char **q)
{ {
double x; double x;
char *p; char *p;
x = strtod(str, &p); x = strtod(str, &p);
if (*p == 'G' || *p == 'g') x *= 1e9, ++p; if (*p == 'G' || *p == 'g')
else if (*p == 'M' || *p == 'm') x *= 1e6, ++p; x *= 1e9, ++p;
else if (*p == 'K' || *p == 'k') x *= 1e3, ++p; else if (*p == 'M' || *p == 'm')
if (q) *q = p; x *= 1e6, ++p;
else if (*p == 'K' || *p == 'k')
x *= 1e3, ++p;
if (q)
*q = p;
return (int64_t)(x + .499); return (int64_t)(x + .499);
} }
@ -107,14 +144,23 @@ static inline int64_t mm_parse_num(const char *str)
static inline void yes_or_no(mm_mapopt_t *opt, int64_t flag, int long_idx, const char *arg, int yes_to_set) static inline void yes_or_no(mm_mapopt_t *opt, int64_t flag, int long_idx, const char *arg, int yes_to_set)
{ {
if (yes_to_set) { if (yes_to_set)
if (strcmp(arg, "yes") == 0 || strcmp(arg, "y") == 0) opt->flag |= flag; {
else if (strcmp(arg, "no") == 0 || strcmp(arg, "n") == 0) opt->flag &= ~flag; if (strcmp(arg, "yes") == 0 || strcmp(arg, "y") == 0)
else fprintf(stderr, "[WARNING]\033[1;31m option '--%s' only accepts 'yes' or 'no'.\033[0m\n", long_options[long_idx].name); opt->flag |= flag;
} else { else if (strcmp(arg, "no") == 0 || strcmp(arg, "n") == 0)
if (strcmp(arg, "yes") == 0 || strcmp(arg, "y") == 0) opt->flag &= ~flag; opt->flag &= ~flag;
else if (strcmp(arg, "no") == 0 || strcmp(arg, "n") == 0) opt->flag |= flag; else
else fprintf(stderr, "[WARNING]\033[1;31m option '--%s' only accepts 'yes' or 'no'.\033[0m\n", long_options[long_idx].name); fprintf(stderr, "[WARNING]\033[1;31m option '--%s' only accepts 'yes' or 'no'.\033[0m\n", long_options[long_idx].name);
}
else
{
if (strcmp(arg, "yes") == 0 || strcmp(arg, "y") == 0)
opt->flag &= ~flag;
else if (strcmp(arg, "no") == 0 || strcmp(arg, "n") == 0)
opt->flag |= flag;
else
fprintf(stderr, "[WARNING]\033[1;31m option '--%s' only accepts 'yes' or 'no'.\033[0m\n", long_options[long_idx].name);
} }
} }
@ -135,194 +181,360 @@ int main(int argc, char *argv[])
mm_realtime0 = realtime(); mm_realtime0 = realtime();
mm_set_opt(0, &ipt, &opt); mm_set_opt(0, &ipt, &opt);
while ((c = ketopt(&o, argc, argv, 1, opt_str, long_options)) >= 0) { // test command line options and apply option -x/preset first #ifdef ANALYSIS_PERF
if (c == 'x') {
if (mm_set_opt(o.arg, &ipt, &opt) < 0) { time_mm_idx_reader_read = 0;
time_mm_map_file_frag = 0;
time_map_work_for_block_1 = 0;
time_map_work_for_block_2 = 0;
time_map_work_for_block_3 = 0;
time_mm_map_frag_b1 = 0;
time_mm_map_frag_b2 = 0;
time_mm_map_frag_b3 = 0;
time_mm_map_frag_b4 = 0;
time_mm_map_frag_b5 = 0;
time_mm_map_frag_b6 = 0;
#endif
while ((c = ketopt(&o, argc, argv, 1, opt_str, long_options)) >= 0)
{ // test command line options and apply option -x/preset first
if (c == 'x')
{
if (mm_set_opt(o.arg, &ipt, &opt) < 0)
{
fprintf(stderr, "[ERROR] unknown preset '%s'\n", o.arg); fprintf(stderr, "[ERROR] unknown preset '%s'\n", o.arg);
return 1; return 1;
} }
} else if (c == ':') { }
else if (c == ':')
{
fprintf(stderr, "[ERROR] missing option argument\n"); fprintf(stderr, "[ERROR] missing option argument\n");
return 1; return 1;
} else if (c == '?') { }
else if (c == '?')
{
fprintf(stderr, "[ERROR] unknown option in \"%s\"\n", argv[o.i - 1]); fprintf(stderr, "[ERROR] unknown option in \"%s\"\n", argv[o.i - 1]);
return 1; return 1;
} }
} }
o = KETOPT_INIT; o = KETOPT_INIT;
while ((c = ketopt(&o, argc, argv, 1, opt_str, long_options)) >= 0) { while ((c = ketopt(&o, argc, argv, 1, opt_str, long_options)) >= 0)
if (c == 'w') ipt.w = atoi(o.arg); {
else if (c == 'k') ipt.k = atoi(o.arg); if (c == 'w')
else if (c == 'H') ipt.flag |= MM_I_HPC; ipt.w = atoi(o.arg);
else if (c == 'd') fnw = o.arg; // the above are indexing related options, except -I else if (c == 'k')
else if (c == 't') n_threads = atoi(o.arg); ipt.k = atoi(o.arg);
else if (c == 'v') mm_verbose = atoi(o.arg); else if (c == 'H')
else if (c == 'g') opt.max_gap = (int)mm_parse_num(o.arg); ipt.flag |= MM_I_HPC;
else if (c == 'G') mm_mapopt_max_intron_len(&opt, (int)mm_parse_num(o.arg)); else if (c == 'd')
else if (c == 'F') opt.max_frag_len = (int)mm_parse_num(o.arg); fnw = o.arg; // the above are indexing related options, except -I
else if (c == 'N') old_best_n = opt.best_n, opt.best_n = atoi(o.arg); else if (c == 't')
else if (c == 'p') opt.pri_ratio = atof(o.arg); n_threads = atoi(o.arg);
else if (c == 'M') opt.mask_level = atof(o.arg); else if (c == 'v')
else if (c == 'c') opt.flag |= MM_F_OUT_CG | MM_F_CIGAR; mm_verbose = atoi(o.arg);
else if (c == 'D') opt.flag |= MM_F_NO_DIAG; else if (c == 'g')
else if (c == 'P') opt.flag |= MM_F_ALL_CHAINS; opt.max_gap = (int)mm_parse_num(o.arg);
else if (c == 'X') opt.flag |= MM_F_ALL_CHAINS | MM_F_NO_DIAG | MM_F_NO_DUAL | MM_F_NO_LJOIN; // -D -P --no-long-join --dual=no else if (c == 'G')
else if (c == 'a') opt.flag |= MM_F_OUT_SAM | MM_F_CIGAR; mm_mapopt_max_intron_len(&opt, (int)mm_parse_num(o.arg));
else if (c == 'Q') opt.flag |= MM_F_NO_QUAL; else if (c == 'F')
else if (c == 'Y') opt.flag |= MM_F_SOFTCLIP; opt.max_frag_len = (int)mm_parse_num(o.arg);
else if (c == 'L') opt.flag |= MM_F_LONG_CIGAR; else if (c == 'N')
else if (c == 'y') opt.flag |= MM_F_COPY_COMMENT; old_best_n = opt.best_n, opt.best_n = atoi(o.arg);
else if (c == 'T') opt.sdust_thres = atoi(o.arg); else if (c == 'p')
else if (c == 'n') opt.min_cnt = atoi(o.arg); opt.pri_ratio = atof(o.arg);
else if (c == 'm') opt.min_chain_score = atoi(o.arg); else if (c == 'M')
else if (c == 'A') opt.a = atoi(o.arg); opt.mask_level = atof(o.arg);
else if (c == 'B') opt.b = atoi(o.arg); else if (c == 'c')
else if (c == 'b') opt.transition = atoi(o.arg); opt.flag |= MM_F_OUT_CG | MM_F_CIGAR;
else if (c == 's') opt.min_dp_max = atoi(o.arg); else if (c == 'D')
else if (c == 'C') opt.noncan = atoi(o.arg); opt.flag |= MM_F_NO_DIAG;
else if (c == 'I') ipt.batch_size = mm_parse_num(o.arg); else if (c == 'P')
else if (c == 'K') opt.mini_batch_size = mm_parse_num(o.arg); opt.flag |= MM_F_ALL_CHAINS;
else if (c == 'e') opt.occ_dist = mm_parse_num(o.arg); else if (c == 'X')
else if (c == 'R') rg = o.arg; opt.flag |= MM_F_ALL_CHAINS | MM_F_NO_DIAG | MM_F_NO_DUAL | MM_F_NO_LJOIN; // -D -P --no-long-join --dual=no
else if (c == 'h') fp_help = stdout; else if (c == 'a')
else if (c == '2') opt.flag |= MM_F_2_IO_THREADS; opt.flag |= MM_F_OUT_SAM | MM_F_CIGAR;
else if (c == 'J') { else if (c == 'Q')
opt.flag |= MM_F_NO_QUAL;
else if (c == 'Y')
opt.flag |= MM_F_SOFTCLIP;
else if (c == 'L')
opt.flag |= MM_F_LONG_CIGAR;
else if (c == 'y')
opt.flag |= MM_F_COPY_COMMENT;
else if (c == 'T')
opt.sdust_thres = atoi(o.arg);
else if (c == 'n')
opt.min_cnt = atoi(o.arg);
else if (c == 'm')
opt.min_chain_score = atoi(o.arg);
else if (c == 'A')
opt.a = atoi(o.arg);
else if (c == 'B')
opt.b = atoi(o.arg);
else if (c == 'b')
opt.transition = atoi(o.arg);
else if (c == 's')
opt.min_dp_max = atoi(o.arg);
else if (c == 'C')
opt.noncan = atoi(o.arg);
else if (c == 'I')
ipt.batch_size = mm_parse_num(o.arg);
else if (c == 'K')
opt.mini_batch_size = mm_parse_num(o.arg);
else if (c == 'e')
opt.occ_dist = mm_parse_num(o.arg);
else if (c == 'R')
rg = o.arg;
else if (c == 'h')
fp_help = stdout;
else if (c == '2')
opt.flag |= MM_F_2_IO_THREADS;
else if (c == 'J')
{
int t; int t;
t = atoi(o.arg); t = atoi(o.arg);
if (t == 0) opt.flag |= MM_F_SPLICE_OLD; if (t == 0)
else if (t == 1) opt.flag &= ~MM_F_SPLICE_OLD; opt.flag |= MM_F_SPLICE_OLD;
} else if (c == 'o') { else if (t == 1)
if (strcmp(o.arg, "-") != 0) { opt.flag &= ~MM_F_SPLICE_OLD;
if (freopen(o.arg, "wb", stdout) == NULL) { }
else if (c == 'o')
{
if (strcmp(o.arg, "-") != 0)
{
if (freopen(o.arg, "wb", stdout) == NULL)
{
fprintf(stderr, "[ERROR]\033[1;31m failed to write the output to file '%s'\033[0m: %s\n", o.arg, strerror(errno)); fprintf(stderr, "[ERROR]\033[1;31m failed to write the output to file '%s'\033[0m: %s\n", o.arg, strerror(errno));
exit(1); exit(1);
} }
} }
} }
else if (c == 300) ipt.bucket_bits = atoi(o.arg); // --bucket-bits else if (c == 300)
else if (c == 302) opt.seed = atoi(o.arg); // --seed ipt.bucket_bits = atoi(o.arg); // --bucket-bits
else if (c == 303) mm_dbg_flag |= MM_DBG_NO_KALLOC; // --no-kalloc else if (c == 302)
else if (c == 304) mm_dbg_flag |= MM_DBG_PRINT_QNAME; // --print-qname opt.seed = atoi(o.arg); // --seed
else if (c == 306) mm_dbg_flag |= MM_DBG_PRINT_QNAME | MM_DBG_PRINT_SEED, n_threads = 1; // --print-seed else if (c == 303)
else if (c == 307) opt.max_chain_skip = atoi(o.arg); // --max-chain-skip mm_dbg_flag |= MM_DBG_NO_KALLOC; // --no-kalloc
else if (c == 339) opt.max_chain_iter = atoi(o.arg); // --max-chain-iter else if (c == 304)
else if (c == 308) opt.min_ksw_len = atoi(o.arg); // --min-dp-len mm_dbg_flag |= MM_DBG_PRINT_QNAME; // --print-qname
else if (c == 309) mm_dbg_flag |= MM_DBG_PRINT_QNAME | MM_DBG_PRINT_ALN_SEQ, n_threads = 1; // --print-aln-seq else if (c == 306)
else if (c == 310) opt.flag |= MM_F_SPLICE; // --splice mm_dbg_flag |= MM_DBG_PRINT_QNAME | MM_DBG_PRINT_SEED, n_threads = 1; // --print-seed
else if (c == 312) opt.flag |= MM_F_NO_LJOIN; // --no-long-join else if (c == 307)
else if (c == 313) opt.flag |= MM_F_SR; // --sr opt.max_chain_skip = atoi(o.arg); // --max-chain-skip
else if (c == 317) opt.end_bonus = atoi(o.arg); // --end-bonus else if (c == 339)
else if (c == 318) opt.flag |= MM_F_INDEPEND_SEG; // --no-pairing opt.max_chain_iter = atoi(o.arg); // --max-chain-iter
else if (c == 320) ipt.flag |= MM_I_NO_SEQ; // --idx-no-seq else if (c == 308)
else if (c == 321) opt.anchor_ext_shift = atoi(o.arg); // --end-seed-pen opt.min_ksw_len = atoi(o.arg); // --min-dp-len
else if (c == 322) opt.flag |= MM_F_FOR_ONLY; // --for-only else if (c == 309)
else if (c == 323) opt.flag |= MM_F_REV_ONLY; // --rev-only mm_dbg_flag |= MM_DBG_PRINT_QNAME | MM_DBG_PRINT_ALN_SEQ, n_threads = 1; // --print-aln-seq
else if (c == 327) opt.max_clip_ratio = atof(o.arg); // --max-clip-ratio else if (c == 310)
else if (c == 328) opt.min_mid_occ = atoi(o.arg); // --min-occ-floor opt.flag |= MM_F_SPLICE; // --splice
else if (c == 329) opt.flag |= MM_F_OUT_MD; // --MD else if (c == 312)
else if (c == 331) opt.sc_ambi = atoi(o.arg); // --score-N opt.flag |= MM_F_NO_LJOIN; // --no-long-join
else if (c == 332) opt.flag |= MM_F_EQX; // --eqx else if (c == 313)
else if (c == 333) opt.flag |= MM_F_PAF_NO_HIT; // --paf-no-hit opt.flag |= MM_F_SR; // --sr
else if (c == 334) opt.split_prefix = o.arg; // --split-prefix else if (c == 317)
else if (c == 335) opt.flag |= MM_F_NO_END_FLT; // --no-end-flt opt.end_bonus = atoi(o.arg); // --end-bonus
else if (c == 336) opt.flag |= MM_F_HARD_MLEVEL; // --hard-mask-level else if (c == 318)
else if (c == 337) opt.max_sw_mat = mm_parse_num(o.arg); // --cap-sw-mat opt.flag |= MM_F_INDEPEND_SEG; // --no-pairing
else if (c == 338) opt.max_qlen = mm_parse_num(o.arg); // --max-qlen else if (c == 320)
else if (c == 340) junc_bed = o.arg; // --junc-bed ipt.flag |= MM_I_NO_SEQ; // --idx-no-seq
else if (c == 341) opt.junc_bonus = atoi(o.arg); // --junc-bonus else if (c == 321)
else if (c == 342) opt.flag |= MM_F_SAM_HIT_ONLY; // --sam-hit-only opt.anchor_ext_shift = atoi(o.arg); // --end-seed-pen
else if (c == 343) opt.chain_gap_scale = atof(o.arg); // --chain-gap-scale else if (c == 322)
else if (c == 351) opt.chain_skip_scale = atof(o.arg); // --chain-skip-scale opt.flag |= MM_F_FOR_ONLY; // --for-only
else if (c == 344) alt_list = o.arg; // --alt else if (c == 323)
else if (c == 345) opt.alt_drop = atof(o.arg); // --alt-drop opt.flag |= MM_F_REV_ONLY; // --rev-only
else if (c == 346) opt.mask_len = mm_parse_num(o.arg); // --mask-len else if (c == 327)
else if (c == 348) opt.flag |= MM_F_QSTRAND | MM_F_NO_INV; // --qstrand opt.max_clip_ratio = atof(o.arg); // --max-clip-ratio
else if (c == 349) opt.cap_kalloc = mm_parse_num(o.arg); // --cap-kalloc else if (c == 328)
else if (c == 350) opt.q_occ_frac = atof(o.arg); // --q-occ-frac opt.min_mid_occ = atoi(o.arg); // --min-occ-floor
else if (c == 352) mm_dbg_flag |= MM_DBG_PRINT_CHAIN; // --print-chains else if (c == 329)
else if (c == 353) opt.flag |= MM_F_NO_HASH_NAME; // --no-hash-name opt.flag |= MM_F_OUT_MD; // --MD
else if (c == 354) opt.flag |= MM_F_SECONDARY_SEQ; // --secondary-seq else if (c == 331)
else if (c == 330) { opt.sc_ambi = atoi(o.arg); // --score-N
else if (c == 332)
opt.flag |= MM_F_EQX; // --eqx
else if (c == 333)
opt.flag |= MM_F_PAF_NO_HIT; // --paf-no-hit
else if (c == 334)
opt.split_prefix = o.arg; // --split-prefix
else if (c == 335)
opt.flag |= MM_F_NO_END_FLT; // --no-end-flt
else if (c == 336)
opt.flag |= MM_F_HARD_MLEVEL; // --hard-mask-level
else if (c == 337)
opt.max_sw_mat = mm_parse_num(o.arg); // --cap-sw-mat
else if (c == 338)
opt.max_qlen = mm_parse_num(o.arg); // --max-qlen
else if (c == 340)
junc_bed = o.arg; // --junc-bed
else if (c == 341)
opt.junc_bonus = atoi(o.arg); // --junc-bonus
else if (c == 342)
opt.flag |= MM_F_SAM_HIT_ONLY; // --sam-hit-only
else if (c == 343)
opt.chain_gap_scale = atof(o.arg); // --chain-gap-scale
else if (c == 351)
opt.chain_skip_scale = atof(o.arg); // --chain-skip-scale
else if (c == 344)
alt_list = o.arg; // --alt
else if (c == 345)
opt.alt_drop = atof(o.arg); // --alt-drop
else if (c == 346)
opt.mask_len = mm_parse_num(o.arg); // --mask-len
else if (c == 348)
opt.flag |= MM_F_QSTRAND | MM_F_NO_INV; // --qstrand
else if (c == 349)
opt.cap_kalloc = mm_parse_num(o.arg); // --cap-kalloc
else if (c == 350)
opt.q_occ_frac = atof(o.arg); // --q-occ-frac
else if (c == 352)
mm_dbg_flag |= MM_DBG_PRINT_CHAIN; // --print-chains
else if (c == 353)
opt.flag |= MM_F_NO_HASH_NAME; // --no-hash-name
else if (c == 354)
opt.flag |= MM_F_SECONDARY_SEQ; // --secondary-seq
else if (c == 330)
{
fprintf(stderr, "[WARNING] \033[1;31m --lj-min-ratio has been deprecated.\033[0m\n"); fprintf(stderr, "[WARNING] \033[1;31m --lj-min-ratio has been deprecated.\033[0m\n");
} else if (c == 314) { // --frag }
else if (c == 314)
{ // --frag
yes_or_no(&opt, MM_F_FRAG_MODE, o.longidx, o.arg, 1); yes_or_no(&opt, MM_F_FRAG_MODE, o.longidx, o.arg, 1);
} else if (c == 315) { // --secondary }
else if (c == 315)
{ // --secondary
yes_or_no(&opt, MM_F_NO_PRINT_2ND, o.longidx, o.arg, 0); yes_or_no(&opt, MM_F_NO_PRINT_2ND, o.longidx, o.arg, 0);
} else if (c == 316) { // --cs }
else if (c == 316)
{ // --cs
opt.flag |= MM_F_OUT_CS | MM_F_CIGAR; opt.flag |= MM_F_OUT_CS | MM_F_CIGAR;
if (o.arg == 0 || strcmp(o.arg, "short") == 0) { if (o.arg == 0 || strcmp(o.arg, "short") == 0)
{
opt.flag &= ~MM_F_OUT_CS_LONG; opt.flag &= ~MM_F_OUT_CS_LONG;
} else if (strcmp(o.arg, "long") == 0) { }
else if (strcmp(o.arg, "long") == 0)
{
opt.flag |= MM_F_OUT_CS_LONG; opt.flag |= MM_F_OUT_CS_LONG;
} else if (strcmp(o.arg, "none") == 0) { }
else if (strcmp(o.arg, "none") == 0)
{
opt.flag &= ~MM_F_OUT_CS; opt.flag &= ~MM_F_OUT_CS;
} else if (mm_verbose >= 2) { }
else if (mm_verbose >= 2)
{
fprintf(stderr, "[WARNING]\033[1;31m --cs only takes 'short' or 'long'. Invalid values are assumed to be 'short'.\033[0m\n"); fprintf(stderr, "[WARNING]\033[1;31m --cs only takes 'short' or 'long'. Invalid values are assumed to be 'short'.\033[0m\n");
} }
} else if (c == 319) { // --splice-flank }
else if (c == 319)
{ // --splice-flank
yes_or_no(&opt, MM_F_SPLICE_FLANK, o.longidx, o.arg, 1); yes_or_no(&opt, MM_F_SPLICE_FLANK, o.longidx, o.arg, 1);
} else if (c == 324) { // --heap-sort }
else if (c == 324)
{ // --heap-sort
yes_or_no(&opt, MM_F_HEAP_SORT, o.longidx, o.arg, 1); yes_or_no(&opt, MM_F_HEAP_SORT, o.longidx, o.arg, 1);
} else if (c == 326) { // --dual }
else if (c == 326)
{ // --dual
yes_or_no(&opt, MM_F_NO_DUAL, o.longidx, o.arg, 0); yes_or_no(&opt, MM_F_NO_DUAL, o.longidx, o.arg, 0);
} else if (c == 347) { // --rmq }
if (o.arg) yes_or_no(&opt, MM_F_RMQ, o.longidx, o.arg, 1); else if (c == 347)
else opt.flag |= MM_F_RMQ; { // --rmq
} else if (c == 'S') { if (o.arg)
yes_or_no(&opt, MM_F_RMQ, o.longidx, o.arg, 1);
else
opt.flag |= MM_F_RMQ;
}
else if (c == 'S')
{
opt.flag |= MM_F_OUT_CS | MM_F_CIGAR | MM_F_OUT_CS_LONG; opt.flag |= MM_F_OUT_CS | MM_F_CIGAR | MM_F_OUT_CS_LONG;
if (mm_verbose >= 2) if (mm_verbose >= 2)
fprintf(stderr, "[WARNING]\033[1;31m option -S is deprecated and may be removed in future. Please use --cs=long instead.\033[0m\n"); fprintf(stderr, "[WARNING]\033[1;31m option -S is deprecated and may be removed in future. Please use --cs=long instead.\033[0m\n");
} else if (c == 'V') { }
else if (c == 'V')
{
puts(MM_VERSION); puts(MM_VERSION);
return 0; return 0;
} else if (c == 'r') { }
else if (c == 'r')
{
opt.bw = (int)mm_parse_num2(o.arg, &s); opt.bw = (int)mm_parse_num2(o.arg, &s);
if (*s == ',') opt.bw_long = (int)mm_parse_num2(s + 1, &s); if (*s == ',')
} else if (c == 'U') { opt.bw_long = (int)mm_parse_num2(s + 1, &s);
}
else if (c == 'U')
{
opt.min_mid_occ = strtol(o.arg, &s, 10); opt.min_mid_occ = strtol(o.arg, &s, 10);
if (*s == ',') opt.max_mid_occ = strtol(s + 1, &s, 10); if (*s == ',')
} else if (c == 'f') { opt.max_mid_occ = strtol(s + 1, &s, 10);
}
else if (c == 'f')
{
double x; double x;
char *p; char *p;
x = strtod(o.arg, &p); x = strtod(o.arg, &p);
if (x < 1.0) opt.mid_occ_frac = x, opt.mid_occ = 0; if (x < 1.0)
else opt.mid_occ = (int)(x + .499); opt.mid_occ_frac = x, opt.mid_occ = 0;
if (*p == ',') opt.max_occ = (int)(strtod(p+1, &p) + .499); else
} else if (c == 'u') { opt.mid_occ = (int)(x + .499);
if (*o.arg == 'b') opt.flag |= MM_F_SPLICE_FOR|MM_F_SPLICE_REV; // both strands if (*p == ',')
else if (*o.arg == 'f') opt.flag |= MM_F_SPLICE_FOR, opt.flag &= ~MM_F_SPLICE_REV; // match GT-AG opt.max_occ = (int)(strtod(p + 1, &p) + .499);
else if (*o.arg == 'r') opt.flag |= MM_F_SPLICE_REV, opt.flag &= ~MM_F_SPLICE_FOR; // match CT-AC (reverse complement of GT-AG) }
else if (*o.arg == 'n') opt.flag &= ~(MM_F_SPLICE_FOR|MM_F_SPLICE_REV); // don't try to match the GT-AG signal else if (c == 'u')
else { {
if (*o.arg == 'b')
opt.flag |= MM_F_SPLICE_FOR | MM_F_SPLICE_REV; // both strands
else if (*o.arg == 'f')
opt.flag |= MM_F_SPLICE_FOR, opt.flag &= ~MM_F_SPLICE_REV; // match GT-AG
else if (*o.arg == 'r')
opt.flag |= MM_F_SPLICE_REV, opt.flag &= ~MM_F_SPLICE_FOR; // match CT-AC (reverse complement of GT-AG)
else if (*o.arg == 'n')
opt.flag &= ~(MM_F_SPLICE_FOR | MM_F_SPLICE_REV); // don't try to match the GT-AG signal
else
{
fprintf(stderr, "[ERROR]\033[1;31m unrecognized cDNA direction\033[0m\n"); fprintf(stderr, "[ERROR]\033[1;31m unrecognized cDNA direction\033[0m\n");
return 1; return 1;
} }
} else if (c == 'z') { }
else if (c == 'z')
{
opt.zdrop = opt.zdrop_inv = strtol(o.arg, &s, 10); opt.zdrop = opt.zdrop_inv = strtol(o.arg, &s, 10);
if (*s == ',') opt.zdrop_inv = strtol(s + 1, &s, 10); if (*s == ',')
} else if (c == 'O') { opt.zdrop_inv = strtol(s + 1, &s, 10);
}
else if (c == 'O')
{
opt.q = opt.q2 = strtol(o.arg, &s, 10); opt.q = opt.q2 = strtol(o.arg, &s, 10);
if (*s == ',') opt.q2 = strtol(s + 1, &s, 10); if (*s == ',')
} else if (c == 'E') { opt.q2 = strtol(s + 1, &s, 10);
}
else if (c == 'E')
{
opt.e = opt.e2 = strtol(o.arg, &s, 10); opt.e = opt.e2 = strtol(o.arg, &s, 10);
if (*s == ',') opt.e2 = strtol(s + 1, &s, 10); if (*s == ',')
opt.e2 = strtol(s + 1, &s, 10);
} }
} }
if ((opt.flag & MM_F_SPLICE) && (opt.flag & MM_F_FRAG_MODE)) { if ((opt.flag & MM_F_SPLICE) && (opt.flag & MM_F_FRAG_MODE))
{
fprintf(stderr, "[ERROR]\033[1;31m --splice and --frag should not be specified at the same time.\033[0m\n"); fprintf(stderr, "[ERROR]\033[1;31m --splice and --frag should not be specified at the same time.\033[0m\n");
return 1; return 1;
} }
if (!fnw && !(opt.flag&MM_F_CIGAR)) if (!fnw && !(opt.flag & MM_F_CIGAR))
ipt.flag |= MM_I_NO_SEQ; ipt.flag |= MM_I_NO_SEQ;
if (mm_check_opt(&ipt, &opt) < 0) if (mm_check_opt(&ipt, &opt) < 0)
return 1; return 1;
if (opt.best_n == 0) { if (opt.best_n == 0)
{
fprintf(stderr, "[WARNING]\033[1;31m changed '-N 0' to '-N %d --secondary=no'.\033[0m\n", old_best_n); fprintf(stderr, "[WARNING]\033[1;31m changed '-N 0' to '-N %d --secondary=no'.\033[0m\n", old_best_n);
opt.best_n = old_best_n, opt.flag |= MM_F_NO_PRINT_2ND; opt.best_n = old_best_n, opt.flag |= MM_F_NO_PRINT_2ND;
} }
if (argc == o.ind || fp_help == stdout) { if (argc == o.ind || fp_help == stdout)
{
fprintf(fp_help, "Usage: minimap2 [options] <target.fa>|<target.idx> [query.fa] [...]\n"); fprintf(fp_help, "Usage: minimap2 [options] <target.fa>|<target.idx> [query.fa] [...]\n");
fprintf(fp_help, "Options:\n"); fprintf(fp_help, "Options:\n");
fprintf(fp_help, " Indexing:\n"); fprintf(fp_help, " Indexing:\n");
@ -339,7 +551,7 @@ int main(int argc, char *argv[])
fprintf(fp_help, " -r NUM[,NUM] chaining/alignment bandwidth and long-join bandwidth [%d,%d]\n", opt.bw, opt.bw_long); fprintf(fp_help, " -r NUM[,NUM] chaining/alignment bandwidth and long-join bandwidth [%d,%d]\n", opt.bw, opt.bw_long);
fprintf(fp_help, " -n INT minimal number of minimizers on a chain [%d]\n", opt.min_cnt); fprintf(fp_help, " -n INT minimal number of minimizers on a chain [%d]\n", opt.min_cnt);
fprintf(fp_help, " -m INT minimal chaining score (matching bases minus log gap penalty) [%d]\n", opt.min_chain_score); fprintf(fp_help, " -m INT minimal chaining score (matching bases minus log gap penalty) [%d]\n", opt.min_chain_score);
// fprintf(fp_help, " -T INT SDUST threshold; 0 to disable SDUST [%d]\n", opt.sdust_thres); // TODO: this option is never used; might be buggy // fprintf(fp_help, " -T INT SDUST threshold; 0 to disable SDUST [%d]\n", opt.sdust_thres); // TODO: this option is never used; might be buggy
fprintf(fp_help, " -X skip self and dual mappings (for the all-vs-all mode)\n"); fprintf(fp_help, " -X skip self and dual mappings (for the all-vs-all mode)\n");
fprintf(fp_help, " -p FLOAT min secondary-to-primary score ratio [%g]\n", opt.pri_ratio); fprintf(fp_help, " -p FLOAT min secondary-to-primary score ratio [%g]\n", opt.pri_ratio);
fprintf(fp_help, " -N INT retain at most INT secondary alignments [%d]\n", opt.best_n); fprintf(fp_help, " -N INT retain at most INT secondary alignments [%d]\n", opt.best_n);
@ -364,7 +576,7 @@ int main(int argc, char *argv[])
fprintf(fp_help, " -Y use soft clipping for supplementary alignments\n"); fprintf(fp_help, " -Y use soft clipping for supplementary alignments\n");
fprintf(fp_help, " -t INT number of threads [%d]\n", n_threads); fprintf(fp_help, " -t INT number of threads [%d]\n", n_threads);
fprintf(fp_help, " -K NUM minibatch size for mapping [500M]\n"); fprintf(fp_help, " -K NUM minibatch size for mapping [500M]\n");
// fprintf(fp_help, " -v INT verbose level [%d]\n", mm_verbose); // fprintf(fp_help, " -v INT verbose level [%d]\n", mm_verbose);
fprintf(fp_help, " --version show version number\n"); fprintf(fp_help, " --version show version number\n");
fprintf(fp_help, " Preset:\n"); fprintf(fp_help, " Preset:\n");
fprintf(fp_help, " -x STR preset (always applied before other options; see minimap2.1 for details) []\n"); fprintf(fp_help, " -x STR preset (always applied before other options; see minimap2.1 for details) []\n");
@ -375,45 +587,55 @@ int main(int argc, char *argv[])
fprintf(fp_help, " - splice/splice:hq - long-read/Pacbio-CCS spliced alignment\n"); fprintf(fp_help, " - splice/splice:hq - long-read/Pacbio-CCS spliced alignment\n");
fprintf(fp_help, " - sr - genomic short-read mapping\n"); fprintf(fp_help, " - sr - genomic short-read mapping\n");
fprintf(fp_help, "\nSee `man ./minimap2.1' for detailed description of these and other advanced command-line options.\n"); fprintf(fp_help, "\nSee `man ./minimap2.1' for detailed description of these and other advanced command-line options.\n");
return fp_help == stdout? 0 : 1; return fp_help == stdout ? 0 : 1;
} }
if ((opt.flag & MM_F_SR) && argc - o.ind > 3) { if ((opt.flag & MM_F_SR) && argc - o.ind > 3)
{
fprintf(stderr, "[ERROR] incorrect input: in the sr mode, please specify no more than two query files.\n"); fprintf(stderr, "[ERROR] incorrect input: in the sr mode, please specify no more than two query files.\n");
return 1; return 1;
} }
idx_rdr = mm_idx_reader_open(argv[o.ind], &ipt, fnw); idx_rdr = mm_idx_reader_open(argv[o.ind], &ipt, fnw);
if (idx_rdr == 0) { if (idx_rdr == 0)
{
fprintf(stderr, "[ERROR] failed to open file '%s': %s\n", argv[o.ind], strerror(errno)); fprintf(stderr, "[ERROR] failed to open file '%s': %s\n", argv[o.ind], strerror(errno));
return 1; return 1;
} }
if (!idx_rdr->is_idx && fnw == 0 && argc - o.ind < 2) { if (!idx_rdr->is_idx && fnw == 0 && argc - o.ind < 2)
{
fprintf(stderr, "[ERROR] missing input: please specify a query file to map or option -d to keep the index\n"); fprintf(stderr, "[ERROR] missing input: please specify a query file to map or option -d to keep the index\n");
mm_idx_reader_close(idx_rdr); mm_idx_reader_close(idx_rdr);
return 1; return 1;
} }
if (opt.best_n == 0 && (opt.flag&MM_F_CIGAR) && mm_verbose >= 2) if (opt.best_n == 0 && (opt.flag & MM_F_CIGAR) && mm_verbose >= 2)
fprintf(stderr, "[WARNING]\033[1;31m `-N 0' reduces alignment accuracy. Please use --secondary=no to suppress secondary alignments.\033[0m\n"); fprintf(stderr, "[WARNING]\033[1;31m `-N 0' reduces alignment accuracy. Please use --secondary=no to suppress secondary alignments.\033[0m\n");
while ((mi = mm_idx_reader_read(idx_rdr, n_threads)) != 0) { while ((mi = mm_idx_reader_read(idx_rdr, n_threads)) != 0)
{
int ret; int ret;
if ((opt.flag & MM_F_CIGAR) && (mi->flag & MM_I_NO_SEQ)) { if ((opt.flag & MM_F_CIGAR) && (mi->flag & MM_I_NO_SEQ))
{
fprintf(stderr, "[ERROR] the prebuilt index doesn't contain sequences.\n"); fprintf(stderr, "[ERROR] the prebuilt index doesn't contain sequences.\n");
mm_idx_destroy(mi); mm_idx_destroy(mi);
mm_idx_reader_close(idx_rdr); mm_idx_reader_close(idx_rdr);
return 1; return 1;
} }
if ((opt.flag & MM_F_OUT_SAM) && idx_rdr->n_parts == 1) { if ((opt.flag & MM_F_OUT_SAM) && idx_rdr->n_parts == 1)
if (mm_idx_reader_eof(idx_rdr)) { {
if (mm_idx_reader_eof(idx_rdr))
{
if (opt.split_prefix == 0) if (opt.split_prefix == 0)
ret = mm_write_sam_hdr(mi, rg, MM_VERSION, argc, argv); ret = mm_write_sam_hdr(mi, rg, MM_VERSION, argc, argv);
else else
ret = mm_write_sam_hdr(0, rg, MM_VERSION, argc, argv); ret = mm_write_sam_hdr(0, rg, MM_VERSION, argc, argv);
} else { }
else
{
ret = mm_write_sam_hdr(0, rg, MM_VERSION, argc, argv); ret = mm_write_sam_hdr(0, rg, MM_VERSION, argc, argv);
if (opt.split_prefix == 0 && mm_verbose >= 2) if (opt.split_prefix == 0 && mm_verbose >= 2)
fprintf(stderr, "[WARNING]\033[1;31m For a multi-part index, no @SQ lines will be outputted. Please use --split-prefix.\033[0m\n"); fprintf(stderr, "[WARNING]\033[1;31m For a multi-part index, no @SQ lines will be outputted. Please use --split-prefix.\033[0m\n");
} }
if (ret != 0) { if (ret != 0)
{
mm_idx_destroy(mi); mm_idx_destroy(mi);
mm_idx_reader_close(idx_rdr); mm_idx_reader_close(idx_rdr);
return 1; return 1;
@ -422,25 +644,37 @@ int main(int argc, char *argv[])
if (mm_verbose >= 3) if (mm_verbose >= 3)
fprintf(stderr, "[M::%s::%.3f*%.2f] loaded/built the index for %d target sequence(s)\n", fprintf(stderr, "[M::%s::%.3f*%.2f] loaded/built the index for %d target sequence(s)\n",
__func__, realtime() - mm_realtime0, cputime() / (realtime() - mm_realtime0), mi->n_seq); __func__, realtime() - mm_realtime0, cputime() / (realtime() - mm_realtime0), mi->n_seq);
if (argc != o.ind + 1) mm_mapopt_update(&opt, mi); if (argc != o.ind + 1)
if (mm_verbose >= 3) mm_idx_stat(mi); mm_mapopt_update(&opt, mi);
if (junc_bed) mm_idx_bed_read(mi, junc_bed, 1); if (mm_verbose >= 3)
if (alt_list) mm_idx_alt_read(mi, alt_list); mm_idx_stat(mi);
if (argc - (o.ind + 1) == 0) { if (junc_bed)
mm_idx_bed_read(mi, junc_bed, 1);
if (alt_list)
mm_idx_alt_read(mi, alt_list);
if (argc - (o.ind + 1) == 0)
{
mm_idx_destroy(mi); mm_idx_destroy(mi);
continue; // no query files continue; // no query files
} }
ret = 0; ret = 0;
if (!(opt.flag & MM_F_FRAG_MODE)) { if (!(opt.flag & MM_F_FRAG_MODE))
for (i = o.ind + 1; i < argc; ++i) { {
// 如果是read / read overlap操作，那么把第一个输入的read.fq当做索引，后续的read比对到该索引
for (i = o.ind + 1; i < argc; ++i)
{
ret = mm_map_file(mi, argv[i], &opt, n_threads); ret = mm_map_file(mi, argv[i], &opt, n_threads);
if (ret < 0) break; if (ret < 0)
break;
} }
} else { }
ret = mm_map_file_frag(mi, argc - (o.ind + 1), (const char**)&argv[o.ind + 1], &opt, n_threads); else
{
ret = mm_map_file_frag(mi, argc - (o.ind + 1), (const char **)&argv[o.ind + 1], &opt, n_threads);
} }
mm_idx_destroy(mi); mm_idx_destroy(mi);
if (ret < 0) { if (ret < 0)
{
fprintf(stderr, "ERROR: failed to map the query file\n"); fprintf(stderr, "ERROR: failed to map the query file\n");
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
@ -449,19 +683,42 @@ int main(int argc, char *argv[])
mm_idx_reader_close(idx_rdr); mm_idx_reader_close(idx_rdr);
if (opt.split_prefix) if (opt.split_prefix)
mm_split_merge(argc - (o.ind + 1), (const char**)&argv[o.ind + 1], &opt, n_parts); mm_split_merge(argc - (o.ind + 1), (const char **)&argv[o.ind + 1], &opt, n_parts);
if (fflush(stdout) == EOF) { if (fflush(stdout) == EOF)
{
perror("[ERROR] failed to write the results"); perror("[ERROR] failed to write the results");
exit(EXIT_FAILURE); exit(EXIT_FAILURE);
} }
if (mm_verbose >= 3) { if (mm_verbose >= 3)
{
fprintf(stderr, "[M::%s] Version: %s\n", __func__, MM_VERSION); fprintf(stderr, "[M::%s] Version: %s\n", __func__, MM_VERSION);
fprintf(stderr, "[M::%s] CMD:", __func__); fprintf(stderr, "[M::%s] CMD:", __func__);
for (i = 0; i < argc; ++i) for (i = 0; i < argc; ++i)
fprintf(stderr, " %s", argv[i]); fprintf(stderr, " %s", argv[i]);
fprintf(stderr, "\n[M::%s] Real time: %.3f sec; CPU: %.3f sec; Peak RSS: %.3f GB\n", __func__, realtime() - mm_realtime0, cputime(), peakrss() / 1024.0 / 1024.0 / 1024.0); fprintf(stderr, "\n[M::%s] Real time: %.3f sec; CPU: %.3f sec; Peak RSS: %.3f GB\n", __func__, realtime() - mm_realtime0, cputime(), peakrss() / 1024.0 / 1024.0 / 1024.0);
} }
#ifdef ANALYSIS_PERF
fprintf(stderr, "\n");
fprintf(stderr, "time_mm_idx_reader_read: %f s\n", time_mm_idx_reader_read / 1000.0);
fprintf(stderr, "time_mm_map_file_frag: %f s\n", time_mm_map_file_frag / 1000.0);
fprintf(stderr, "time_map_work_for_block_1: %f s\n", time_map_work_for_block_1 / 1000.0 / n_threads);
fprintf(stderr, "time_map_work_for_block_2: %f s\n", time_map_work_for_block_2 / 1000.0 / n_threads);
fprintf(stderr, "time_map_work_for_block_3: %f s\n", time_map_work_for_block_3 / 1000.0 / n_threads);
fprintf(stderr, "time_mm_map_frag_b1: %f s\n", time_mm_map_frag_b1 / 1000.0 / n_threads);
fprintf(stderr, "time_mm_map_frag_b2: %f s\n", time_mm_map_frag_b2 / 1000.0 / n_threads);
fprintf(stderr, "time_mm_map_frag_b3: %f s\n", time_mm_map_frag_b3 / 1000.0 / n_threads);
fprintf(stderr, "time_mm_map_frag_b4: %f s\n", time_mm_map_frag_b4 / 1000.0 / n_threads);
fprintf(stderr, "time_mm_map_frag_b5: %f s\n", time_mm_map_frag_b5 / 1000.0 / n_threads);
fprintf(stderr, "time_mm_map_frag_b6: %f s\n", time_mm_map_frag_b6 / 1000.0 / n_threads);
fprintf(stderr, "time_collect_seed_hits_heap: %f s\n", time_collect_seed_hits_heap / 1000.0 / n_threads);
fprintf(stderr, "time_collect_seed_hits: %f s\n", time_collect_seed_hits / 1000.0 / n_threads);
fprintf(stderr, "time_mg_lchain_dp: %f s\n", time_mg_lchain_dp / 1000.0 / n_threads);
fprintf(stderr, "time_mg_chain_backtrack: %f s\n", time_mg_chain_backtrack / 1000.0 / n_threads);
fprintf(stderr, "\n");
#endif
return 0; return 0;
} }

649
map.c
View File

File diff suppressed because it is too large Load Diff

3
minimap.h
View File

@ -7,6 +7,9 @@
#define MM_VERSION "2.26-r1175" #define MM_VERSION "2.26-r1175"
// 用来开关调试性能分析，运行时间等信息
#define ANALYSIS_PERF 1
#define MM_F_NO_DIAG (0x001LL) // no exact diagonal hit #define MM_F_NO_DIAG (0x001LL) // no exact diagonal hit
#define MM_F_NO_DUAL (0x002LL) // skip pairs where query name is lexicographically larger than target name #define MM_F_NO_DUAL (0x002LL) // skip pairs where query name is lexicographically larger than target name
#define MM_F_CIGAR (0x004LL) #define MM_F_CIGAR (0x004LL)

2
sketch.c
View File

@ -106,7 +106,7 @@ void mm_sketch(void *km, const char *str, int len, int w, int k, uint32_t rid, i
kmer[0] = (kmer[0] << 2 | c) & mask; // forward k-mer kmer[0] = (kmer[0] << 2 | c) & mask; // forward k-mer
kmer[1] = (kmer[1] >> 2) | (3ULL^c) << shift1; // reverse k-mer kmer[1] = (kmer[1] >> 2) | (3ULL^c) << shift1; // reverse k-mer
if (kmer[0] == kmer[1]) continue; // skip "symmetric k-mers" as we don't know it strand if (kmer[0] == kmer[1]) continue; // skip "symmetric k-mers" as we don't know it strand
z = kmer[0] < kmer[1]? 0 : 1; // strand z = kmer[0] < kmer[1]? 0 : 1; // strand // kmer的strand到底是什么意思？为什么通过比较就能确定正反？
++l; ++l;
if (l >= k && kmer_span < 256) { if (l >= k && kmer_span < 256) {
info.x = hash64(kmer[z], mask) << 8 | kmer_span; info.x = hash64(kmer[z], mask) << 8 | kmer_span;