8 #include "sam_header.h"
13 KHASH_SET_INIT_STR(rg)
15 // When counting records instead of printing them,
16 // data passed to the bam_fetch callback is encapsulated in this struct.
22 typedef khash_t(rg) *rghash_t;
24 static rghash_t g_rghash = 0;
25 static int g_min_mapQ = 0, g_flag_on = 0, g_flag_off = 0;
26 static char *g_library, *g_rg;
27 static int g_sol2sanger_tbl[128];
30 void *bed_read(const char *fn);
31 void bed_destroy(void *_h);
32 int bed_overlap(const void *_h, const char *chr, int beg, int end);
34 static void sol2sanger(bam1_t *b)
37 uint8_t *qual = bam1_qual(b);
38 if (g_sol2sanger_tbl[30] == 0) {
39 for (l = 0; l != 128; ++l) {
40 g_sol2sanger_tbl[l] = (int)(10.0 * log(1.0 + pow(10.0, (l - 64 + 33) / 10.0)) / log(10.0) + .499);
41 if (g_sol2sanger_tbl[l] >= 93) g_sol2sanger_tbl[l] = 93;
44 for (l = 0; l < b->core.l_qseq; ++l) {
47 qual[l] = g_sol2sanger_tbl[q];
51 static inline int __g_skip_aln(const bam_header_t *h, const bam1_t *b)
53 if (b->core.qual < g_min_mapQ || ((b->core.flag & g_flag_on) != g_flag_on) || (b->core.flag & g_flag_off))
55 if (g_bed && b->core.tid >= 0 && !bed_overlap(g_bed, h->target_name[b->core.tid], b->core.pos, bam_calend(&b->core, bam1_cigar(b))))
57 if (g_rg || g_rghash) {
58 uint8_t *s = bam_aux_get(b, "RG");
60 if (g_rg) return (strcmp(g_rg, (char*)(s + 1)) == 0)? 0 : 1;
62 khint_t k = kh_get(rg, g_rghash, (char*)(s + 1));
63 return (k != kh_end(g_rghash))? 0 : 1;
68 const char *p = bam_get_library((bam_header_t*)h, b);
69 return (p && strcmp(p, g_library) == 0)? 0 : 1;
74 static char *drop_rg(char *hdtxt, rghash_t h, int *len)
76 char *p = hdtxt, *q, *r, *s;
78 memset(&str, 0, sizeof(kstring_t));
82 if (q == 0) q = p + strlen(p);
83 if (q - p < 3) break; // the line is too short; then stop
84 if (strncmp(p, "@RG\t", 4) == 0) {
87 if ((r = strstr(p, "\tID:")) != 0) {
89 for (s = r; *s != '\0' && *s != '\n' && *s != '\t'; ++s);
93 if (k != kh_end(h)) toprint = 1;
97 kputsn(p, q - p, &str); kputc('\n', &str);
105 // callback function for bam_fetch() that prints nonskipped records
106 static int view_func(const bam1_t *b, void *data)
108 if (!__g_skip_aln(((samfile_t*)data)->header, b))
109 samwrite((samfile_t*)data, b);
113 // callback function for bam_fetch() that counts nonskipped records
114 static int count_func(const bam1_t *b, void *data)
116 if (!__g_skip_aln(((count_func_data_t*)data)->header, b)) {
117 (*((count_func_data_t*)data)->count)++;
122 static int usage(int is_long_help);
124 int main_samview(int argc, char *argv[])
126 int c, is_header = 0, is_header_only = 0, is_bamin = 1, ret = 0, compress_level = -1, is_bamout = 0, slx2sngr = 0, is_count = 0;
127 int of_type = BAM_OFDEC, is_long_help = 0;
129 samfile_t *in = 0, *out = 0;
130 char in_mode[5], out_mode[5], *fn_out = 0, *fn_list = 0, *fn_ref = 0, *fn_rg = 0;
132 /* parse command-line options */
133 strcpy(in_mode, "r"); strcpy(out_mode, "w");
134 while ((c = getopt(argc, argv, "Sbct:h1Ho:q:f:F:ul:r:xX?T:CR:L:")) >= 0) {
136 case 'c': is_count = 1; break;
137 case 'C': slx2sngr = 1; break;
138 case 'S': is_bamin = 0; break;
139 case 'b': is_bamout = 1; break;
140 case 't': fn_list = strdup(optarg); is_bamin = 0; break;
141 case 'h': is_header = 1; break;
142 case 'H': is_header_only = 1; break;
143 case 'o': fn_out = strdup(optarg); break;
144 case 'f': g_flag_on = strtol(optarg, 0, 0); break;
145 case 'F': g_flag_off = strtol(optarg, 0, 0); break;
146 case 'q': g_min_mapQ = atoi(optarg); break;
147 case 'u': compress_level = 0; break;
148 case '1': compress_level = 1; break;
149 case 'l': g_library = strdup(optarg); break;
150 case 'L': g_bed = bed_read(optarg); break;
151 case 'r': g_rg = strdup(optarg); break;
152 case 'R': fn_rg = strdup(optarg); break;
153 case 'x': of_type = BAM_OFHEX; break;
154 case 'X': of_type = BAM_OFSTR; break;
155 case '?': is_long_help = 1; break;
156 case 'T': fn_ref = strdup(optarg); is_bamin = 0; break;
157 default: return usage(is_long_help);
160 if (compress_level >= 0) is_bamout = 1;
161 if (is_header_only) is_header = 1;
162 if (is_bamout) strcat(out_mode, "b");
164 if (of_type == BAM_OFHEX) strcat(out_mode, "x");
165 else if (of_type == BAM_OFSTR) strcat(out_mode, "X");
167 if (is_bamin) strcat(in_mode, "b");
168 if (is_header) strcat(out_mode, "h");
169 if (compress_level >= 0) {
171 tmp[0] = compress_level + '0'; tmp[1] = '\0';
172 strcat(out_mode, tmp);
174 if (argc == optind) return usage(is_long_help); // potential memory leak...
176 // read the list of read groups
181 g_rghash = kh_init(rg);
182 fp_rg = fopen(fn_rg, "r");
183 while (!feof(fp_rg) && fscanf(fp_rg, "%s", buf) > 0) // this is not a good style, but bear me...
184 kh_put(rg, g_rghash, strdup(buf), &ret); // we'd better check duplicates...
188 // generate the fn_list if necessary
189 if (fn_list == 0 && fn_ref) fn_list = samfaipath(fn_ref);
190 // open file handlers
191 if ((in = samopen(argv[optind], in_mode, fn_list)) == 0) {
192 fprintf(pysamerr, "[main_samview] fail to open \"%s\" for reading.\n", argv[optind]);
196 if (in->header == 0) {
197 fprintf(pysamerr, "[main_samview] fail to read the header from \"%s\".\n", argv[optind]);
201 if (g_rghash) { // FIXME: I do not know what "bam_header_t::n_text" is for...
204 tmp = drop_rg(in->header->text, g_rghash, &l);
205 free(in->header->text);
206 in->header->text = tmp;
207 in->header->l_text = l;
209 if (!is_count && (out = samopen(fn_out? fn_out : "-", out_mode, in->header)) == 0) {
210 fprintf(pysamerr, "[main_samview] fail to open \"%s\" for writing.\n", fn_out? fn_out : "standard output");
214 if (is_header_only) goto view_end; // no need to print alignments
216 if (argc == optind + 1) { // convert/print the entire file
217 bam1_t *b = bam_init1();
219 while ((r = samread(in, b)) >= 0) { // read one alignment from `in'
220 if (!__g_skip_aln(in->header, b)) {
221 if (slx2sngr) sol2sanger(b);
222 if (!is_count) samwrite(out, b); // write the alignment to `out'
227 fprintf(pysamerr, "[main_samview] truncated file.\n");
231 } else { // retrieve alignments in specified regions
233 bam_index_t *idx = 0;
234 if (is_bamin) idx = bam_index_load(argv[optind]); // load BAM index
235 if (idx == 0) { // index is unavailable
236 fprintf(pysamerr, "[main_samview] random alignment retrieval only works for indexed BAM files.\n");
240 for (i = optind + 1; i < argc; ++i) {
241 int tid, beg, end, result;
242 bam_parse_region(in->header, argv[i], &tid, &beg, &end); // parse a region in the format like `chr2:100-200'
243 if (tid < 0) { // reference name is not found
244 fprintf(pysamerr, "[main_samview] region \"%s\" specifies an unknown reference name. Continue anyway.\n", argv[i]);
249 count_func_data_t count_data = { in->header, &count };
250 result = bam_fetch(in->x.bam, idx, tid, beg, end, &count_data, count_func);
252 result = bam_fetch(in->x.bam, idx, tid, beg, end, out, view_func);
254 fprintf(pysamerr, "[main_samview] retrieval of region \"%s\" failed due to truncated file or corrupt BAM index file\n", argv[i]);
259 bam_index_destroy(idx); // destroy the BAM index
263 if (is_count && ret == 0) {
264 printf("%d\n", count);
266 // close files, free and return
267 free(fn_list); free(fn_ref); free(fn_out); free(g_library); free(g_rg); free(fn_rg);
268 if (g_bed) bed_destroy(g_bed);
271 for (k = 0; k < kh_end(g_rghash); ++k)
272 if (kh_exist(g_rghash, k)) free((char*)kh_key(g_rghash, k));
273 kh_destroy(rg, g_rghash);
281 static int usage(int is_long_help)
283 fprintf(pysamerr, "\n");
284 fprintf(pysamerr, "Usage: samtools view [options] <in.bam>|<in.sam> [region1 [...]]\n\n");
285 fprintf(pysamerr, "Options: -b output BAM\n");
286 fprintf(pysamerr, " -h print header for the SAM output\n");
287 fprintf(pysamerr, " -H print header only (no alignments)\n");
288 fprintf(pysamerr, " -S input is SAM\n");
289 fprintf(pysamerr, " -u uncompressed BAM output (force -b)\n");
290 fprintf(pysamerr, " -1 fast compression (force -b)\n");
291 fprintf(pysamerr, " -x output FLAG in HEX (samtools-C specific)\n");
292 fprintf(pysamerr, " -X output FLAG in string (samtools-C specific)\n");
293 fprintf(pysamerr, " -c print only the count of matching records\n");
294 fprintf(pysamerr, " -L FILE output alignments overlapping the input BED FILE [null]\n");
295 fprintf(pysamerr, " -t FILE list of reference names and lengths (force -S) [null]\n");
296 fprintf(pysamerr, " -T FILE reference sequence file (force -S) [null]\n");
297 fprintf(pysamerr, " -o FILE output file name [stdout]\n");
298 fprintf(pysamerr, " -R FILE list of read groups to be outputted [null]\n");
299 fprintf(pysamerr, " -f INT required flag, 0 for unset [0]\n");
300 fprintf(pysamerr, " -F INT filtering flag, 0 for unset [0]\n");
301 fprintf(pysamerr, " -q INT minimum mapping quality [0]\n");
302 fprintf(pysamerr, " -l STR only output reads in library STR [null]\n");
303 fprintf(pysamerr, " -r STR only output reads in read group STR [null]\n");
304 fprintf(pysamerr, " -? longer help\n");
305 fprintf(pysamerr, "\n");
307 fprintf(pysamerr, "Notes:\n\
309 1. By default, this command assumes the file on the command line is in\n\
310 the BAM format and it prints the alignments in SAM. If `-t' is\n\
311 applied, the input file is assumed to be in the SAM format. The\n\
312 file supplied with `-t' is SPACE/TAB delimited with the first two\n\
313 fields of each line consisting of the reference name and the\n\
314 corresponding sequence length. The `.fai' file generated by `faidx'\n\
315 can be used here. This file may be empty if reads are unaligned.\n\
317 2. SAM->BAM conversion: `samtools view -bT ref.fa in.sam.gz'.\n\
319 3. BAM->SAM conversion: `samtools view in.bam'.\n\
321 4. A region should be presented in one of the following formats:\n\
322 `chr1', `chr2:1,000' and `chr3:1000-2,000'. When a region is\n\
323 specified, the input alignment file must be an indexed BAM file.\n\
325 5. Option `-u' is preferred over `-b' when the output is piped to\n\
326 another samtools command.\n\
328 6. In a string FLAG, each character represents one bit with\n\
329 p=0x1 (paired), P=0x2 (properly paired), u=0x4 (unmapped),\n\
330 U=0x8 (mate unmapped), r=0x10 (reverse), R=0x20 (mate reverse)\n\
331 1=0x40 (first), 2=0x80 (second), s=0x100 (not primary), \n\
332 f=0x200 (failure) and d=0x400 (duplicate). Note that `-x' and\n\
333 `-X' are samtools-C specific. Picard and older samtools do not\n\
334 support HEX or string flags.\n\
339 int main_import(int argc, char *argv[])
344 fprintf(pysamerr, "Usage: bamtk import <in.ref_list> <in.sam> <out.bam>\n");
348 argv2 = calloc(6, sizeof(char*));
349 argv2[0] = "import", argv2[1] = "-o", argv2[2] = argv[3], argv2[3] = "-bt", argv2[4] = argv[1], argv2[5] = argv[2];
350 ret = main_samview(argc2, argv2);