From 6e245391fc8edbc9bd529e0d6f04cac4742ac393 Mon Sep 17 00:00:00 2001
From: Thomas Krannich <krannich479@googlemail.com>
Date: Tue, 15 Mar 2022 15:17:39 +0100
Subject: [PATCH] initial change to see if SAM/BAM headers are modified
---
Makefile | 2 +-
src/popins2_merge_and_set_mate.h | 31 ++++++++++++++++++++++++++++---
2 files changed, 29 insertions(+), 4 deletions(-)
diff --git a/Makefile b/Makefile
index b5e8591..9f4ba8e 100644
--- a/Makefile
+++ b/Makefile
@@ -18,7 +18,7 @@ TOOLS=-DSAMTOOLS=\"$(SAMTOOLS)\" -DBWA=\"$(BWA)\" -DSICKLE=\"$(SICKLE)\" -DVELVE
# Date and version number from git
DATE := on $(shell git log --pretty=format:"%cd" --date=iso | cut -f 1,2 -d " " | head -n 1)
-VERSION := 0.0.4-snake-$(shell git log --pretty=format:"%h" --date=iso | head -n 1)
+VERSION := 0.0.5-snake-$(shell git log --pretty=format:"%h" --date=iso | head -n 1)
CXXFLAGS += -DDATE=\""$(DATE)"\" -DVERSION=\""$(VERSION)"\"
# Compiler flags
diff --git a/src/popins2_merge_and_set_mate.h b/src/popins2_merge_and_set_mate.h
index 588e8c4..7205ff2 100644
--- a/src/popins2_merge_and_set_mate.h
+++ b/src/popins2_merge_and_set_mate.h
@@ -101,6 +101,7 @@ inline void readRecordAndCorrectRIds(BamAlignmentRecord & record, BamFileIn & st
// ==========================================================================
inline void mergeHeaders(BamHeader & header, FormattedFileContext<BamFileOut, Owner<> >::Type & context, BamFileIn & stream1, BamFileIn & stream2){
+
StringSet<CharString> contigNames;
NameStoreCache<StringSet<CharString> > nameStoreCache;
String<int32_t> contigLengths;
@@ -109,11 +110,35 @@ inline void mergeHeaders(BamHeader & header, FormattedFileContext<BamFileOut, Ow
readHeader(header, stream1);
BamHeader header2;
readHeader(header2, stream2);
- for (unsigned i = 0; i < length(header2); ++i)
- {
- if (header2[i].type != BAM_HEADER_FIRST)
+
+ std::unordered_set<char*> program_ids;
+
+ for (unsigned i = 0; i < length(header2); ++i){
+
+ if (header2[i].type != BAM_HEADER_FIRST){
+
+ if (header2[i].type == BAM_HEADER_PROGRAM){
+ CharString PG_ID;
+ bool keyFound = getTagValue(PG_ID, "ID", header2[i]);
+
+ if ( program_ids.find(toCString(PG_ID)) != program_ids.end() ){ // if PG ID was seen already
+ BamHeaderRecord hrecord = header2[i];
+ CharString suf = ".mH"; // mH = "mergeHeader" to trace in SAM/BAM files
+ append(PG_ID, suf);
+ setTagValue("ID", PG_ID, hrecord);
+ appendValue(header, hrecord);
+ continue;
+ }
+ else{
+ program_ids.insert(toCString(PG_ID));
+ }
+
+ }
+
appendValue(header, header2[i]);
+ }
}
+
std::stable_sort(begin(header, Standard()), end(header, Standard()), BamHeaderRecordTypeLess());
// Fill sequence names into nameStoreCache.
--
GitLab