add extract_subpop script (#7387)

.dockstore.yml

@@ -104,7 +104,7 @@ workflows:
          - ah_var_store
    - name: GvsValidateVatTable
      subclass: WDL
-     primaryDescriptorPath: /scripts/variantstore/wdl/GvsValidateVAT.wdl
+     primaryDescriptorPath: /scripts/variantstore/variant_annotations_table/GvsValidateVAT.wdl
      filters:
        branches:
          - ah_var_store

scripts/variantstore/wdl/extract/Dockerfile

@@ -8,6 +8,7 @@ RUN apt-get update && apt-get -y upgrade && apt-get -y install bcftools
 # Add the application source code.
 ADD create_cohort_extract_data_table.py /app
 ADD create_variant_annotation_table.py /app
+ADD extract_subpop.py /app
 
 WORKDIR /app
 ENTRYPOINT ["/bin/bash"]

scripts/variantstore/wdl/extract/extract_subpop.py

@@ -0,0 +1,22 @@
+import csv
+import argparse
+
+def extract_subpopulation(input_path, output_path):
+  with open(input_path, newline='') as tsvin, open(output_path, 'w', newline='') as csvout:
+    tsvin = csv.reader(tsvin, delimiter='\t')
+    csvout = csv.writer(csvout, delimiter='\t')
+
+    for row in tsvin:
+      csvout.writerow([row[0], row[4]])
+
+
+if __name__ == '__main__':
+  parser = argparse.ArgumentParser(allow_abbrev=False, description='Extract subpopulation per sample data out of a callset TSV')
+  parser.add_argument('--input_path',type=str, metavar='path', help='path to the original callset TSV', required=True)
+  parser.add_argument('--output_path',type=str, metavar='path', help='path for the output TSV', required=True)
+
+  args = parser.parse_args()
+
+  extract_subpopulation(args.input_path,
+                        args.output_path)
+