anno chunks bed and unresolved sv fix

can now count chunks with bed file restriction
Fix some SVs being incorrectly marked as unresolved

truvari/annotations/chunks.py

@@ -19,6 +19,8 @@ def parse_args(args):
                                      formatter_class=argparse.RawDescriptionHelpFormatter)
     parser.add_argument("input", type=str,
                         help="Input VCF")
+    parser.add_argument("-b", "--bed", type=str, default=None,
+                        help="Bed file of variants to chunk")
     parser.add_argument("-o", "--output", type=str, default="/dev/stdout",
                         help="Output name")
     parser.add_argument("-c", "--chunksize", type=int, default=500,
@@ -55,6 +57,9 @@ def chunks_main(args):
     m.params.sizemax = args.sizemax
     m.params.chunksize = args.chunksize
     m.params.refdist = args.chunksize
+    if args.bed:
+        regions = truvari.build_region_tree(v, includebed=args.bed)
+        v = truvari.region_filter(v, regions)
     c = truvari.chunker(m, ('base', v))
 
     with open(args.output, 'w') as fout:

truvari/matching.py

@@ -158,8 +158,8 @@ def filter_call(self, entry, base=False):
                 f"Cannot compare multi-allelic records. Please split\nline {str(entry)}")
 
         # Don't compare BNDs, ever
-        if entry.alleles_variant_types[1] == 'BND':
-            return True
+        #if entry.alleles_variant_types[1] == 'BND':
+        # return True
 
         if self.params.passonly and truvari.entry_is_filtered(entry):
             return True
@@ -312,7 +312,7 @@ def chunker(matcher, *files):
             continue
 
         # check symbolic, resolve if needed/possible
-        if matcher.params.pctseq != 0 and entry.alts[0].startswith('<'):
+        if matcher.params.pctseq != 0 and entry.alleles_variant_types[-1] == 'BND' or entry.alts[0].startswith('<'):
             was_resolved = resolve_sv(entry,
                                       matcher.reference,
                                       matcher.params.dup_to_ins)
@@ -354,6 +354,10 @@ def resolve_sv(entry, ref, dup_to_ins=False):
     if ref is None or entry.alts[0] in ['<CNV>', '<INS>'] or entry.start > ref.get_reference_length(entry.chrom):
         return False
 
+    if entry.alleles_variant_types[1] == 'BND' and "SVTYPE" in entry.info \
+            and truvari.entry_variant_type(entry) == truvari.SV.DEL:
+        entry.alts = ['<DEL>']
+
     seq = ref.fetch(entry.chrom, entry.start, entry.stop)
     if entry.alts[0] == '<DEL>':
         entry.ref = seq