Much simpler faster code for parsing fastq files.

[fastq-tools.git] / src / fastq-grep.c

diff --git a/src/fastq-grep.c b/src/fastq-grep.c
index 7c9e3bf60ec844a944719d9e87c03d93dfdf424b..4d1603f7ced45f3d34e28d00e60204689031d4a8 100644 (file)
--- a/src/fastq-grep.c
+++ b/src/fastq-grep.c
@@ -40,6 +40,7 @@ void print_help()
 "Options:\n"
 "  -i, --id                match the read id (by default, sequence is matched)\n"
 "  -v, --invert-match      select nonmatching entries\n"
+"  -m, --mismatches=FILE   output mismatching entries to the given file\n"
 "  -c, --count             output only the number of matching sequences\n"
 "  -h, --help              print this message\n"
 "  -V, --version           output version information and exit\n"
@@ -52,16 +53,16 @@ static int id_flag;
 
 
 
-void fastq_grep(FILE* fin, FILE* fout, pcre* re)
+void fastq_grep(FILE* fin, FILE* fout, FILE* mismatch_file, pcre* re)
 {
     int rc;
     int ovector[3];
     size_t count = 0;
 
-    fastq_t* fqf = fastq_open(fin);
-    seq_t* seq = fastq_alloc_seq();
+    fastq_t* fqf = fastq_create(fin);
+    seq_t* seq = seq_create();
 
-    while (fastq_next(fqf, seq)) {
+    while (fastq_read(fqf, seq)) {
 
         rc = pcre_exec(re,          /* pattern */
                        NULL,        /* extra data */
@@ -76,10 +77,13 @@ void fastq_grep(FILE* fin, FILE* fout, pcre* re)
             if (count_flag) count++;
             else            fastq_print(fout, seq);
         }
+        else if (mismatch_file) {
+            fastq_print(mismatch_file, seq);
+        }
     }
 
-    fastq_free_seq(seq);
-    fastq_close(fqf);
+    seq_free(seq);
+    fastq_free(fqf);
 
     if (count_flag) fprintf(fout, "%zu\n", count);
 }
@@ -106,11 +110,13 @@ int main(int argc, char* argv[])
     int opt;
     int opt_idx;
 
+    FILE* mismatch_file = NULL;
 
     static struct option long_options[] =
         { 
           {"id",           no_argument, &id_flag,     1},
           {"invert-match", no_argument, &invert_flag, 1},
+          {"mismatches",   required_argument, NULL, 'm'},
           {"count",        no_argument, &count_flag,  1},
           {"help",         no_argument, NULL, 'h'},
           {"version",      no_argument, NULL, 'V'},
@@ -118,9 +124,9 @@ int main(int argc, char* argv[])
         };
 
     while (1) {
-        opt = getopt_long(argc, argv, "ivchV", long_options, &opt_idx);
+        opt = getopt_long(argc, argv, "ivmchV", long_options, &opt_idx);
 
-        if( opt == -1 ) break;
+        if (opt == -1) break;
 
         switch (opt) {
             case 0:
@@ -137,6 +143,14 @@ int main(int argc, char* argv[])
                 invert_flag = 1;
                 break;
 
+            case 'm':
+                mismatch_file = fopen(optarg, "w");
+                if (mismatch_file == NULL) {
+                    fprintf(stderr, "No such file '%s'.\n", optarg);
+                    return 1;
+                }
+                break;
+
             case 'c':
                 count_flag = 1;
                 break;
@@ -174,7 +188,7 @@ int main(int argc, char* argv[])
 
 
     if (optind >= argc || (argc - optind == 1 && strcmp(argv[optind],"-") == 0)) {
-        fastq_grep(stdin, stdout, re);
+        fastq_grep(stdin, stdout, mismatch_file, re);
     }
     else {
         for (; optind < argc; optind++) {
@@ -184,13 +198,14 @@ int main(int argc, char* argv[])
                 continue;
             }
 
-            fastq_grep(fin, stdout, re);
+            fastq_grep(fin, stdout, mismatch_file, re);
 
             fclose(fin);
         }
     }
 
     pcre_free(re);
+    if (mismatch_file) fclose(mismatch_file);
 
     return 0;
 }