fix VCF exporter allele decoding and add validation test

Author: adilraza99

malariagen_data/anoph/to_vcf.py

@@ -119,12 +119,12 @@ def snp_calls_to_vcf(
                         alleles = allele_chunk[j]
                         ref = (
                             alleles[0].decode()
-                            if isinstance(alleles[0], bytes)
+                            if hasattr(alleles[0], "decode")
                             else str(alleles[0])
                         )
                         alt_alleles = []
                         for a in alleles[1:]:
-                            s = a.decode() if isinstance(a, bytes) else str(a)
+                            s = a.decode() if hasattr(a, "decode") else str(a)
                             if s:
                                 alt_alleles.append(s)
                         alt = ",".join(alt_alleles) if alt_alleles else "."

tests/anoph/test_vcf_exporter.py

@@ -110,6 +110,14 @@ def test_vcf_exporter(fixture, api: VcfExporter, tmp_path):
     ds_positions = sorted(ds["variant_position"].values.tolist())
     assert vcf_positions == ds_positions
 
+    # Allele values are clean strings, not byte-string representations.
+    for line in data_lines:
+        fields = line.split("\t")
+        ref, alt = fields[3], fields[4]
+        assert "b'" not in ref and "b'" not in alt, (
+            f"byte-string repr in REF/ALT: REF={ref!r} ALT={alt!r}"
+        )
+
 
 @parametrize_with_cases("fixture,api", cases=".")
 def test_vcf_exporter_overwrite(fixture, api: VcfExporter, tmp_path):