feat(cli.vcf2parquet): format string is controlable

natir · Apr 27, 2023 · 8f178c7 · 8f178c7
1 parent 92e2261
commit 8f178c7
Show file tree

Hide file tree

Showing 2 changed files with 22 additions and 7 deletions.
diff --git a/src/variantplaner/cli.py b/src/variantplaner/cli.py
@@ -71,7 +71,20 @@ def main(threads: int = 1, verbose: int = 0) -> None:
     help="Path where the genotypes will be written in parquet",
     type=click.Path(dir_okay=False, writable=True, path_type=pathlib.Path),
 )
-def vcf2parquet(input_path: pathlib.Path, variants: pathlib.Path, genotypes: pathlib.Path) -> None:
+@click.option(
+    "-f",
+    "--format-string",
+    help="Value of FORMAT column, line not match with this are ignored",
+    type=str,
+    default="GT:AD:DP:GQ",
+    show_default=True,
+)
+def vcf2parquet(
+    input_path: pathlib.Path,
+    variants: pathlib.Path,
+    genotypes: pathlib.Path,
+    format_string: str = "GT:AD:DP:GQ",
+) -> None:
     """Convert a vcf in parquet."""
     logger = logging.getLogger("vcf2parquet")
 
@@ -89,7 +102,7 @@ def vcf2parquet(input_path: pathlib.Path, variants: pathlib.Path, genotypes: pat
     if genotypes:
         try:
             vcf_header = io.vcf.extract_header(input_path)
-            extract.genotypes(lf, io.vcf.format2expr(vcf_header, input_path), "GT:AD:DP:GQ").sink_parquet(genotypes)
+            extract.genotypes(lf, io.vcf.format2expr(vcf_header, input_path), format_string).sink_parquet(genotypes)
         except exception.NoGenotypeError:
             logger.exception("")
             sys.exit(2)

diff --git a/tests/test_cli_help.py b/tests/test_cli_help.py
@@ -123,11 +123,13 @@ def test_show_help_vcf2parquet() -> None:
   Convert a vcf in parquet.
 
 Options:
-  -i, --input-path FILE  Path to vcf input file  [required]
-  -v, --variants FILE    Path where the variants will be written in parquet
-                         [required]
-  -g, --genotypes FILE   Path where the genotypes will be written in parquet
-  --help                 Show this message and exit.
+  -i, --input-path FILE     Path to vcf input file  [required]
+  -v, --variants FILE       Path where the variants will be written in parquet
+                            [required]
+  -g, --genotypes FILE      Path where the genotypes will be written in parquet
+  -f, --format-string TEXT  Value of FORMAT column, line not match with this are
+                            ignored  [default: GT:AD:DP:GQ]
+  --help                    Show this message and exit.
 """
     )