csv2cases.py

import sys
import os
import json
from csv import DictReader

files_to_parse = sys.argv[1:]
for f in files_to_parse:
    print("Reading cases from %s" % f)
    json_filename = "%s.json" % os.path.splitext(f)[0]
    cases = []
    with open(f) as csvfile:
        reader = DictReader(csvfile)
        for row in reader:
            case_dict = dict(row)
            description = case_dict['Comment/History'].replace("\n", " ").strip()
            patient = dict()
            patient['age'] = int(case_dict['Age'])
            patient['sex'] = case_dict['Gender'].upper()
            slide_list = [case_dict['Image ID']]
            case = {'slides': slide_list, 'description': description, 'patient': patient}
            print("  Found case: %s" % case)
            cases.append(case)
        cases_json = json.dumps({'cases': cases}, indent=2)
        with open(json_filename, "w") as json_file:
            print("Writing JSON to %s" % json_filename)
            json_file.write(cases_json)