Datasets:

Modalities:
Audio
Text
Formats:
csv
Libraries:
Datasets
pandas
License:
ru-vc-test / scripts /get_trials.py
nshmyrevgmail's picture
Add dataset files
cbc10c1
raw
history blame contribute delete
501 Bytes
#!/usr/bin/env python3
import sys
import csv
import random
random.seed(1234)
vids = {}
for line in csv.DictReader(open(sys.argv[1])):
fn = line['wav_filename'].split('/')[-1]
vid = "_".join(fn.split("_")[:-1])
if vid not in vids:
vids[vid] = "v_" + fn
keys = list(vids.keys())
size = len(keys)
for i in range(3000):
j = random.randint(0, size)
k = random.randint(0, size)
while k == j:
k = random.randint(0, size)
print (vids[keys[j]], vids[keys[k]])