Exclude samples

tests/test_bcftools_validation.py

@@ -61,7 +61,11 @@ def run_vcztools(args: str) -> str:
         ("view --no-version -s NA00001", "sample.vcf.gz"),
         ("view --no-version -s NA00001,NA00003", "sample.vcf.gz"),
         ("view --no-version -s HG00096", "1kg_2020_chrM.vcf.gz"),
-        ("view --no-version -s '' --force-samples", "sample.vcf.gz")
+        ("view --no-version -s '' --force-samples", "sample.vcf.gz"),
+        ("view --no-version -s ^NA00001", "sample.vcf.gz"),
+        ("view --no-version -s ^NA00003,NA00002", "sample.vcf.gz"),
+        ("view --no-version -s ^NA00003,NA00002,NA00003", "sample.vcf.gz"),
+        ("view --no-version -S ^tests/data/txt/samples.txt", "sample.vcf.gz"),
     ]
 )
 # fmt: on

tests/test_vcf_writer.py

@@ -150,6 +150,9 @@ def test_write_vcf__regions(tmp_path, regions, targets,
         ("NA00001", [[0, 0, True]]),
         ("NA00001,NA00003", [[0, 0, True], [0, 1, False]]),
         ("NA00003,NA00001", [[0, 1, False], [0, 0, True]]),
+        ("^NA00002", [[0, 0, True], [0, 1, False]]),
+        ("^NA00003,NA00002", [[0, 0, True]]),
+        ("^NA00003,NA00002,NA00003", [[0, 0, True]]),
     ]
 )
 def test_write_vcf__samples(tmp_path, samples, expected_genotypes):
@@ -161,7 +164,12 @@ def test_write_vcf__samples(tmp_path, samples, expected_genotypes):
 
     v = VCF(output)
 
-    assert v.samples == samples.split(",")
+    if samples.startswith("^"):
+        # There are three samples in the sample VCF.
+        assert len(set(v.samples)) + len(set(samples[1:].split(","))) == 3
+        assert len(set(v.samples) & set(samples[1:].split(","))) == 0
+    else:
+        assert v.samples == samples.split(",")
 
     variant = next(v)
 

vcztools/cli.py

@@ -149,8 +149,13 @@ def view(
         raise ValueError(f"Output file extension must be .vcf, got: .{split[-1]}")
 
     if samples_file:
+        exclude_samples_file = samples_file.startswith("^")
+        samples_file = samples_file.lstrip("^")
         with open(samples_file) as file:
             samples = samples or ""
+            assert samples == "" or samples_file.startswith("^") == exclude_samples_file
+            if exclude_samples_file and not samples.startswith("^"):
+                samples = "^" + samples
             samples += ",".join(line.strip() for line in file.readlines())
 
     # TODO: use no_update when fixing https://github.com/sgkit-dev/vcztools/issues/75

vcztools/vcf_writer.py

@@ -153,10 +153,18 @@ def write_vcf(
             samples_selection = None
         else:
             all_samples = root["sample_id"][:]
+            exclude_samples = samples.startswith("^")
+            samples = samples.lstrip("^")
             sample_ids = np.array(samples.split(","))
             if np.all(sample_ids == np.array("")):
                 sample_ids = np.empty((0,))
+
             samples_selection = search(all_samples, sample_ids)
+            if exclude_samples:
+                samples_selection = np.setdiff1d(
+                    np.arange(all_samples.size), samples_selection
+                )
+                sample_ids = all_samples[samples_selection]
 
         if not no_header and vcf_header is None:
             if "vcf_header" in root.attrs: