test_data/splits.py

import csv

CSV_FILE = "info_simplified.csv"
TEST_FILE = "test.csv"
VAL_FILE = "validation.csv"
TRAIN_FILE = "train.csv"

test_file = []
val_file = []
train_file = []

with open(CSV_FILE) as fd:
    reader = csv.DictReader(fd)
    for row in reader:
        if row["split"] == "test":
            test_file.append(row)
        elif row["split"] == "validation":
            val_file.append(row)
        elif row["split"] == "train":
            train_file.append(row)

with open(TEST_FILE, "w+") as fd:
    fieldnames = test_file[0].keys()

    writer = csv.DictWriter(fd, fieldnames=fieldnames)
    writer.writeheader()
    writer.writerows(test_file)

with open(VAL_FILE, "w+") as fd:
    fieldnames = val_file[0].keys()

    writer = csv.DictWriter(fd, fieldnames=fieldnames)
    writer.writeheader()
    writer.writerows(val_file)

with open(TRAIN_FILE, "w+") as fd:
    fieldnames = train_file[0].keys()

    writer = csv.DictWriter(fd, fieldnames=fieldnames)
    writer.writeheader()
    writer.writerows(train_file)