Add check for phased calls

Since it is not clear how we should handle phased calls, we check if the
variants are phased and error out if this is the case.

tests/test_evaluate.py

@@ -133,3 +133,27 @@ def test_truncated_positive_no_call(NA12878_positive_truncated):
     """ Variants which are missing from the positive vcf do not count towards
     alleles_no_call """
     assert NA12878_positive_truncated['alleles_no_call'] == 0
+
+
+def test_phased_positive():
+    """ Test error message when the positive vcf contains phased variants """
+    filename = 'tests/cases/gatk.vcf.gz'
+    phased = 'tests/cases/dummy_phased_blank.vcf.gz'
+    call = VCF(filename, gts012=True)
+    phased = VCF(phased, gts012=True)
+    with pytest.raises(NotImplementedError,
+                       match='Phased variants are not supported'):
+        site_concordancy(call, phased, call_samples=['NA12878'],
+                         positive_samples=['BLANK'])
+
+
+def test_phased_call():
+    """ Test error message when the call vcf contains phased variants """
+    filename = 'tests/cases/gatk.vcf.gz'
+    phased = 'tests/cases/dummy_phased_blank.vcf.gz'
+    call = VCF(phased, gts012=True)
+    positive = VCF(filename, gts012=True)
+    with pytest.raises(NotImplementedError,
+                       match='Phased variants are not supported'):
+        site_concordancy(call, positive, call_samples=['BLANK'],
+                         positive_samples=['NA12878'])

vtools/evaluate.py

@@ -165,6 +165,10 @@ def site_concordancy(call_vcf: VCF,
             pos = pos_record.genotypes[p_s]
             cal = call_record.genotypes[c_s]
 
+            # If the genotypes are phased
+            if pos[2] or cal[2]:
+                raise NotImplementedError('Phased variants are not supported')
+
             # Parse the genotypes and add the results into d
             parse_variants(pos, cal, d)