Convert JSON to CSV in Python

json_to_csv

#!/usr/bin/python3

import csv
import io
import json
import pdb
import sys
from collections import OrderedDict


def serialize_cell(cell):
    if isinstance(cell, dict) or isinstance(cell, list):
        return json.dumps(cell)
    else:
        return cell


def post_process(row, prefix=()):
    new_row = OrderedDict()

    for key, value in row.items():
        if isinstance(value, dict):
            subdict = post_process(value, prefix + (key,))
            for subkey, subval in subdict.items():
                new_row[subkey] = subval
        else:
            new_key = ".".join(prefix + (key,))
            new_row[new_key] = serialize_cell(value)

    return new_row


def get_master_keyset(rows):
    aggregate = set()

    for row in rows:
        aggregate = aggregate | set(row)

    return tuple(sorted(aggregate))


def json_string_to_csv(raw):
    as_json = json.loads(raw, object_pairs_hook=OrderedDict)

    key_sets = {tuple(row) for row in as_json}
    if len(key_sets) > 1:
        raise RuntimeError("Inconsistent keys across rows")
    (key_set,) = key_sets

    processed = [post_process(row) for row in as_json]
    master_keyset = get_master_keyset(processed)

    def orderer(key):
        prefix = key.split(".")[0]
        return (key not in key_set, key_set.index(prefix), key)

    sorted_keyset = tuple(sorted(master_keyset, key=orderer))

    buffer = io.StringIO()

    dict_writer = csv.DictWriter(buffer, sorted_keyset)
    dict_writer.writeheader()
    dict_writer.writerows(processed)

    return buffer.getvalue()


def main():
    raw_as_json = sys.stdin.read()
    raw_as_csv = json_string_to_csv(raw_as_json)
    print(raw_as_csv)


if __name__ == "__main__":
    main()