[FIX] not accepting BAM with empty header

CHANGELOG.md

@@ -100,6 +100,7 @@ If possible, provide tooling that performs the changes, e.g. a shell-script.
 * Requesting the alignment without also requesting the sequence for BAM files containing empty CIGAR strings does now
   not result in erroneous parsing ([\#2418](https://github.com/seqan/seqan3/pull/2418)).
 * BAM files with 64 references are now parsed correctly ([\#2423](https://github.com/seqan/seqan3/pull/2423)).
+* BAM files not containing a plain text header are now accepted ([\#2536](https://github.com/seqan/seqan3/pull/2536)).
 * Writing `gz`-compressed output no longer results in `bgzf`-compressed output. This change may have following effects
   ([\#2458](https://github.com/seqan/seqan3/pull/2458)):
   * A noticeable slowdown when writing `gz`-compressed content since, in contrast to `bgzf`, `gz` does not feature

include/seqan3/io/sam_file/format_bam.hpp

@@ -336,7 +336,9 @@ inline void format_bam::read_alignment_record(stream_type & stream,
         int32_t tmp32{};
         read_field(stream_view, tmp32);
 
-        if (tmp32 > 0) // header text is present
+        bool const header_present{tmp32 > 0};
+
+        if (header_present) // header text is present
             read_header(stream_view | views::take_exactly_or_throw(tmp32), header, ref_seqs);
 
         int32_t n_ref;
@@ -352,6 +354,15 @@ inline void format_bam::read_alignment_record(stream_type & stream,
 
             read_field(stream_view, tmp32); // l_ref (length of reference sequence)
 
+            // If there was no header text, we parse reference sequences block as header information
+            if (!header_present)
+            {
+                header.ref_ids().push_back(string_buffer);
+                header.ref_id_info.emplace_back(tmp32, "");
+                header.ref_dict[(header.ref_ids())[(header.ref_ids()).size() - 1]] = (header.ref_ids()).size() - 1;
+                continue;
+            }
+
             auto id_it = header.ref_dict.find(string_buffer);
 
             // sanity checks of reference information to existing header object:

test/unit/io/sam_file/format_bam_test.cpp

@@ -666,3 +666,23 @@ TEST_F(bam_format, issue2417)
 
     EXPECT_EQ(num_records, 1u);
 }
+
+// https://github.com/seqan/seqan3/issues/1201
+TEST_F(bam_format, issue1201)
+{
+    // A BAM file with an l_text of 0
+    std::string const input{
+        // read1	117	ref	1	0	*	=	1	0	ACGTA	IIIII
+        '\x42', '\x41', '\x4d', '\x01', '\x00', '\x00', '\x00', '\x00', '\x01', '\x00', '\x00', '\x00', '\x04', '\x00',
+        '\x00', '\x00', '\x72', '\x65', '\x66', '\x00', '\x70', '\x07', '\x00', '\x00', '\x2e', '\x00', '\x00', '\x00',
+        '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x06', '\x00', '\x49', '\x12', '\x00', '\x00',
+        '\x75', '\x00', '\x05', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00', '\x00',
+        '\x00', '\x00', '\x00', '\x00', '\x72', '\x65', '\x61', '\x64', '\x31', '\x00', '\x12', '\x48', '\x10', '\x28',
+        '\x28', '\x28', '\x28', '\x28'
+    };
+
+    std::istringstream stream{input};
+    seqan3::sam_file_input fin{stream, seqan3::format_bam{}};
+
+    fin.header().format_version;
+}