Create a random sample of records from a JSON Lines file

sample_jsonl.py

import json
import math


with open("data/repos.jsonl") as f:
    records = [json.loads(x) for x in f]

count = len(records)
sample_count = math.ceil(.01*count)
sample_records = random.sample(records, sample_count)

with open("data/repos_sample_0.01.jsonl", "w") as f:
    for record in sample_records:
        f.write(json.dumps(record) + "\n")

print("Sampled {} records from {} original records.".format(
    sample_count,
    count
))