Transform a json document to a ljson document (one json per line)

json_to_ljson.py

import json
import os
from glob import glob

input_doc = "TODO"  # can be a json doc or a folder containing only jsons

output_ljson = "{}_output.ljson".format('.'.join(input_doc.split('.')[:-1]))


def transform_input_to_dumpable(input_doc):
    # Read input
    with open(input_doc) as f:
        content = json.load(f)

    # Transform json object into a json array in case we have only one element
    if isinstance(content, dict):
        content = [content]

    return content


# Parse json content and write in the output_file
with open(output_ljson, 'w') as f:
    input_docs = [input_doc]  # by default take the file
    # If input is directory
    if os.path.isdir(input_doc):
        input_docs = glob("{}/*".format(input_doc))

    # Parse every json file
    for doc in input_docs:
        content = transform_input_to_dumpable(doc)
        for elem in content:
            f.write(json.dumps(elem) + "\n")