Drop Genotypes

lib/vcf_encoder.c

@@ -663,9 +663,11 @@ vcz_variant_encoder_write_format_fields(const vcz_variant_encoder_t *self,
 
     if (all_missing) {
         for (j = 0; j < num_samples + 1; j++) {
-            offset = append_string(buf, ".\t", 2, offset, buflen);
-            if (offset < 0) {
-                goto out;
+            if (num_samples > 0) {
+                offset = append_string(buf, ".\t", 2, offset, buflen);
+                if (offset < 0) {
+                    goto out;
+                }
             }
         }
     } else {

tests/test_bcftools_validation.py

@@ -49,6 +49,10 @@ def run_vcztools(args: str) -> str:
                 "view --no-version -e '(FMT/DP >= 8 | FMT/GQ>40) && POS > 100000'",
                 "sample.vcf.gz"
         ),
+        (
+                "view --no-version -G",
+                "sample.vcf.gz"
+        ),
     ]
 )
 # fmt: on

tests/test_vcf_writer.py

@@ -174,6 +174,18 @@ def test_write_vcf__samples(tmp_path, samples, expected_genotypes):
     assert variant.genotypes == expected_genotypes
 
 
+def test_write_vcf__no_samples(tmp_path):
+    original = pathlib.Path("tests/data/vcf") / "sample.vcf.gz"
+    vcz = vcz_path_cache(original)
+    output = tmp_path.joinpath("output.vcf")
+
+    write_vcf(vcz, output, drop_genotypes=True)
+
+    v = VCF(output)
+
+    assert v.samples == []
+
+
 @pytest.mark.parametrize(
     ("regions", "targets", "samples", "include", "expected_chrom_pos"),
     [

vcztools/cli.py

@@ -89,6 +89,13 @@ def query(path, list_samples, format):
     default=None,
     help="Samples to include.",
 )
+@click.option(
+    "-G",
+    "--drop-genotypes",
+    type=bool,
+    is_flag=True,
+    help="Drop genotypes.",
+)
 @click.option(
     "-t",
     "--targets",
@@ -110,6 +117,7 @@ def view(
     regions,
     targets,
     samples,
+    drop_genotypes,
     include,
     exclude,
 ):
@@ -121,6 +129,7 @@ def view(
         variant_regions=regions,
         variant_targets=targets,
         samples=samples,
+        drop_genotypes=drop_genotypes,
         include=include,
         exclude=exclude,
     )

vcztools/vcf_writer.py

@@ -85,6 +85,7 @@ def write_vcf(
     variant_regions=None,
     variant_targets=None,
     samples=None,
+    drop_genotypes: bool = False,
     include: Optional[str] = None,
     exclude: Optional[str] = None,
 ) -> None:
@@ -138,7 +139,12 @@ def write_vcf(
     root = zarr.open(vcz, mode="r")
 
     with open_file_like(output) as output:
-        if samples is None:
+        if samples and drop_genotypes:
+            raise ValueError("Cannot select samples and drop genotypes.")
+        elif drop_genotypes:
+            sample_ids = []
+            samples_selection = np.array([])
+        elif samples is None:
             sample_ids = root["sample_id"][:]
             samples_selection = None
         else:
@@ -300,7 +306,11 @@ def c_chunk_to_vcf(
     info_fields = {}
     num_samples = len(samples_selection) if samples_selection is not None else None
     for name, array in root.items():
-        if name.startswith("call_") and not name.startswith("call_genotype"):
+        if (
+            name.startswith("call_")
+            and not name.startswith("call_genotype")
+            and num_samples != 0
+        ):
             vcf_name = name[len("call_") :]
             format_fields[vcf_name] = get_vchunk_array(
                 array, v_chunk, v_mask_chunk, samples_selection
@@ -313,7 +323,7 @@ def c_chunk_to_vcf(
 
     gt = None
     gt_phased = None
-    if "call_genotype" in root:
+    if "call_genotype" in root and num_samples != 0:
         array = root["call_genotype"]
         gt = get_vchunk_array(array, v_chunk, v_mask_chunk, samples_selection)
         if "call_genotype_phased" in root:
@@ -380,7 +390,7 @@ def _generate_header(ds, original_header, sample_ids):
     info_fields = []
     format_fields = []
 
-    if "call_genotype" in ds:
+    if "call_genotype" in ds and len(sample_ids) > 0:
         # GT must be the first field if present, per the spec (section 1.6.2)
         format_fields.append("GT")
 
@@ -395,7 +405,8 @@ def _generate_header(ds, original_header, sample_ids):
             key = var[len("variant_") :]
             info_fields.append(key)
         elif (
-            var.startswith("call_")
+            len(sample_ids) > 0
+            and var.startswith("call_")
             and not var.endswith("_fill")
             and not var.endswith("_mask")
             and dims(arr)[0] == "variants"