Support excluding samples with files

sgkit-dev · Sep 11, 2024 · acdeb15 · acdeb15
1 parent 6708f47
commit acdeb15
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 0 deletions.
diff --git a/tests/test_bcftools_validation.py b/tests/test_bcftools_validation.py
@@ -65,6 +65,7 @@ def run_vcztools(args: str) -> str:
         ("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

diff --git a/vcztools/cli.py b/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