Allow "Source" and "Version" in INFO header (#106)

* Allow source and version in INFO header * Bump version --------- Co-authored-by: Daniel Nilsson <[email protected]>
Clinical-Genomics · Oct 24, 2024 · fca33bd · fca33bd
1 parent 9b69d59
commit fca33bd
Show file tree

Hide file tree

Showing 3 changed files with 44 additions and 1 deletion.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -8,6 +8,11 @@ Please add a new candidate release at the top after changing the latest one. Fee
 
 Try to use the following format:
 
+## [unreleased]
+### Fixed
+- The optional fields Source and Version are allowed in the VCF header([#106](https://github.com/Clinical-Genomics/genmod/pull/106))
+
+
 ## [3.9]
 - Fixed wrong models when chromosome X was named `chrX` and not `X` ([#135](https://github.com/Clinical-Genomics/genmod/pull/135))
 - Added GitHub Actions workflows for automatic publishing to PyPI on release, and keep a changelog reminder ([#136](https://github.com/Clinical-Genomics/genmod/pull/136))

diff --git a/genmod/vcf_tools/header_parser.py b/genmod/vcf_tools/header_parser.py
@@ -56,6 +56,8 @@ def __init__(self):
             Number=(?P<number>-?\d+|\.|[AGR]),
             Type=(?P<type>Integer|Float|Flag|Character|String),
             Description="(?P<desc>[^"]*)"
+            (?:,Source="(?P<source>[^"]+)")?
+            (?:,Version="(?P<version>[^"]+)")?
             >''', re.VERBOSE)
         self.filter_pattern = re.compile(r'''\#\#FILTER=<
             ID=(?P<id>[^,]+),
@@ -99,7 +101,8 @@ def parse_meta_data(self, line):
 
             matches = [
                 match.group('id'), match.group('number'), 
-                match.group('type'), match.group('desc')
+                match.group('type'), match.group('desc'),
+                match.group('source'), match.group('version')
             ]
 
             # extra_info is a dictionary to check the metadata about the INFO values:

diff --git a/tests/vcf_tools/test_header_parser.py b/tests/vcf_tools/test_header_parser.py
@@ -12,6 +12,41 @@ def test_parse_info():
     ## THEN assert it is added to the parser
     assert 'MQ' in head.info_dict
 
+def test_parse_info_with_source():
+    ## GIVEN a header object
+    head = HeaderParser()
+    assert 'MQ' not in head.info_dict
+    info_line = '##INFO=<ID=MQ,Number=1,Type=Float,Description="RMS Mapping Quality",Source="bwa">'
+
+    ## WHEN parsing a correct info line
+    head.parse_meta_data(info_line)
+
+    ## THEN assert it is added to the parser
+    assert 'MQ' in head.info_dict
+
+def test_parse_info_with_version():
+    ## GIVEN a header object
+    head = HeaderParser()
+    assert 'MQ' not in head.info_dict
+    info_line = '##INFO=<ID=MQ,Number=1,Type=Float,Description="RMS Mapping Quality",Version="1.0.0">'
+
+    ## WHEN parsing a correct info line
+    head.parse_meta_data(info_line)
+
+    ## THEN assert it is added to the parser
+    assert 'MQ' in head.info_dict
+
+def test_parse_info_with_source_and_version():
+    ## GIVEN a header object
+    head = HeaderParser()
+    assert 'MQ' not in head.info_dict
+    info_line = '##INFO=<ID=MQ,Number=1,Type=Float,Description="RMS Mapping Quality",Source="bwa",Version="1.0.0">'
+
+    ## WHEN parsing a correct info line
+    head.parse_meta_data(info_line)
+
+    ## THEN assert it is added to the parser
+    assert 'MQ' in head.info_dict
 
 def test_parse_contig():
     ## GIVEN a header object