Add missing header tags as per the spec (PR pysam-developers#1238)

Add omitted tags to KNOWN_HEADER_FIELDS and VALID_HEADER_ORDER, so that AlignmentHeader.from_dict()/to_dict() will preserve all header field tags defined in the SAM specification. Addresses pysam-developers#1237. Exercise TP:linear in some header tests. Co-authored-by: John Marshall <jmarshall@hey.com>
indraniel · Oct 20, 2023 · cdc0ed1 · cdc0ed1
1 parent 98a37ad
commit cdc0ed1
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 8 deletions.
diff --git a/pysam/libcalignmentfile.pyx b/pysam/libcalignmentfile.pyx
@@ -111,27 +111,30 @@ VALID_HEADER_TYPES = {"HD" : Mapping,
 VALID_HEADERS = ("HD", "SQ", "RG", "PG", "CO")
 
 # default type conversions within SAM header records
-KNOWN_HEADER_FIELDS = {"HD" : {"VN" : str, "SO" : str, "GO" : str},
+KNOWN_HEADER_FIELDS = {"HD" : {"VN" : str, "SO" : str, "GO" : str,
+                               "SS" : str,},
                        "SQ" : {"SN" : str, "LN" : int, "AS" : str,
                                "M5" : str, "SP" : str, "UR" : str,
-                               "AH" : str,},
+                               "AH" : str, "TP" : str, "DS" : str,
+                               "AN" : str,},
                        "RG" : {"ID" : str, "CN" : str, "DS" : str,
                                "DT" : str, "FO" : str, "KS" : str,
                                "LB" : str, "PG" : str, "PI" : str,
                                "PL" : str, "PM" : str, "PU" : str,
-                               "SM" : str,},
+                               "SM" : str, "BC" : str,},
                        "PG" : {"ID" : str, "PN" : str, "CL" : str,
                                "PP" : str, "DS" : str, "VN" : str,},}
 
 # output order of fields within records. Ensure that CL is at
 # the end as parsing a CL will ignore any subsequent records.
-VALID_HEADER_ORDER = {"HD" : ("VN", "SO", "GO"),
+VALID_HEADER_ORDER = {"HD" : ("VN", "SO", "SS", "GO"),
                       "SQ" : ("SN", "LN", "AS", "M5",
-                               "UR", "SP", "AH"),
+                              "UR", "SP", "AH", "TP",
+                              "DS", "AN"),
                       "RG" : ("ID", "CN", "SM", "LB",
                               "PU", "PI", "DT", "DS",
                               "PL", "FO", "KS", "PG",
-                              "PM"),
+                              "PM", "BC"),
                       "PG" : ("PN", "ID", "VN", "PP",
                               "DS", "CL"),}
 

diff --git a/tests/AlignmentFileHeader_test.py b/tests/AlignmentFileHeader_test.py
@@ -24,7 +24,7 @@ class TestHeaderConstruction(unittest.TestCase):
 
     header_dict = odict(
         [('SQ', [odict([('LN', 1575), ('SN', 'chr1'), ('AH', 'chr1:5000000-5010000')]),
-                 odict([('LN', 1584), ('SN', 'chr2'), ('AH', '*')])]),
+                 odict([('LN', 1584), ('SN', 'chr2'), ('AH', '*'), ('TP', 'linear')])]),
          ('RG', [odict([('LB', 'SC_1'), ('ID', 'L1'), ('SM', 'NA12891'),
                         ('PU', 'SC_1_10'), ("CN", "name:with:colon")]),
                  odict([('LB', 'SC_2'), ('ID', 'L2'), ('SM', 'NA12891'),
@@ -37,7 +37,7 @@ class TestHeaderConstruction(unittest.TestCase):
 
     header_text = ("@HD\tVN:1.0\n"
                    "@SQ\tSN:chr1\tLN:1575\tAH:chr1:5000000-5010000\n"
-                   "@SQ\tSN:chr2\tLN:1584\tAH:*\n"
+                   "@SQ\tSN:chr2\tLN:1584\tAH:*\tTP:linear\n"
                    "@RG\tID:L1\tPU:SC_1_10\tLB:SC_1\tSM:NA12891\tCN:name:with:colon\n"
                    "@RG\tID:L2\tPU:SC_2_12\tLB:SC_2\tSM:NA12891\tCN:name:with:colon\n"
                    "@PG\tID:P1\tVN:1.0\n"