Convert GDC JSON with files details to a TSV with the columns: case_id, file_id, file_name

convert-gdc-files-json-to-tsv.py

import json
import pandas
import argparse

parser = argparse.ArgumentParser(description='Convert GDC JSON with files details to a TSV with the columns: case_id, file_id, file_name')
parser.add_argument('-j', '--json', metavar='JSON', type=str,
                    help='JSON file with file details from GDC')
parser.add_argument('-o', '--output', metavar='TSV',
                    default='file-dict.tsv',
                    help='Path to output file')

args = parser.parse_args()

print(args)

with open(args.json, 'r') as f:
    j = json.load(f)

case_ids = [d['cases'][0]['case_id'] for d in j]
file_ids = [d['file_id'] for d in j]
file_names = [d['file_name'] for d in j]

df = pandas.DataFrame({'case_id': case_ids, 'file_id': file_ids, 'file_name': file_names})
df.to_csv(args.output, sep ='\t', index = False)