Merge branch 'comment-header' into 'dev'

Rework reheadering to always output valid samtag

See merge request epi2melabs/fastcat!17

CHANGELOG.md

@@ -4,6 +4,10 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [v0.9.0]
+### Fixed
+- Ensure reheadered fastq is indeed formatted as a valid SAM tag(s).
+
 ## [v0.8.0]
 ### Added
 - Option to bamstats to add 'sample_name' column equivalent to fastcat.

src/common.c

@@ -67,6 +67,16 @@ char *substring(char *string, int position, int length) {
    return ptr;
 }
 
+int replace_char(char *str, char orig, char rep) {
+    char *ix = str;
+    int n = 0;
+    while((ix = strchr(ix, orig)) != NULL) {
+        *ix++ = rep;
+        n++;
+    }
+    return n;
+}
+
 const double qprobs[100] = {
     1.00000000e+00, 7.94328235e-01, 6.30957344e-01, 5.01187234e-01,
     3.98107171e-01, 3.16227766e-01, 2.51188643e-01, 1.99526231e-01,
@@ -125,4 +135,4 @@ float mean_qual_from_bam(u_int8_t* qual, size_t len) {
     }
     qsum /= len;
     return -10 * log10(qsum);
-}
+}

src/common.h

@@ -44,6 +44,17 @@ void *xalloc(size_t num, size_t size, char* msg);
  */
 char *substring(char *string, int position, int length);
 
+/** Globally replace a char in a char*
+ * 
+ * @param str char* source string
+ * @param orig original character
+ * @param rep replacement
+ * @returns number of times replacement made
+ *
+ */
+int replace_char(char *str, char orig, char rep);
+
+
 // https://en.wikipedia.org/wiki/Kahan_summation_algorithm
 void kahan_sum(double* sum, double term, double* c);
 

src/fastcat/writer.c

@@ -2,6 +2,7 @@
 #include <sys/types.h>
 
 #include "writer.h"
+#include "common.h"
 #include "../fastqcomments.h"
 
 char* strip_path(char* input) {
@@ -73,12 +74,28 @@ void _write_read(writer writer, kseq_t* seq, read_meta meta, void* handle) {
     static const char* wcomment_fmt = "@%s %s\n%s\n+\n%s\n";
     static const char* nocomment_fmt = "@%s\n%s\n+\n%s\n";
 
-    if (writer->reheader && meta->valid) {
-        (*write)(
-            handle, reheader_fmt,
-            seq->name.s, meta->runid, meta->barcode, meta->barcode_alias,
-            meta->flow_cell_id, meta->start_time, meta->read_number, meta->channel, 
-            seq->seq.s, seq->qual.s);
+    if (writer->reheader) {
+        // If we have all items, write individual tags, else just a generic comment tag.
+        // We could write out the explicit tags we have but a) thats effort b) any
+        // files from a device will have all these fields c) doesn't clearly indicate
+        // something went wrong d) allows us to pass on arbitrary information (so
+        // is probably how we should have done this in the first place).
+        if (meta->valid) {
+            (*write)(
+                handle, reheader_fmt,
+                seq->name.s, meta->runid, meta->barcode, meta->barcode_alias,
+                meta->flow_cell_id, meta->start_time, meta->read_number, meta->channel, 
+                seq->seq.s, seq->qual.s);
+        }
+        else {
+            char* buf = xalloc(seq->comment.l + 6, sizeof(char), "Temporary buffer");
+            strncpy(buf, "CO:Z:", 5);
+            strncat(buf, seq->comment.s, seq->comment.l);
+            // remove tabs to space (because multi-tags are tab delimited), this is fine because the existing comments should only contain space as delimiter (not information).
+            replace_char(buf, '\t', ' ');
+            (*write)(handle, wcomment_fmt, seq->name.s, buf, seq->seq.s, seq->qual.s);
+            free(buf);
+        }
     }
     else if (seq->comment.l > 0) {
         (*write)(handle, wcomment_fmt, seq->name.s, seq->comment.s, seq->seq.s, seq->qual.s);

src/version.h

@@ -1,2 +1,2 @@
 
-const char *argp_program_version = "0.8.0";
+const char *argp_program_version = "0.9.0";