Monolingual khakas data:

• located - khakas_texts/data

import pandas as pd

name = 'khakaschiry'
path = f'khakas_texts/data/{name}.csv'
df = pd.read_csv(path)
texts = df['text'].values.tolist()

print(texts[0])

name	source	# texts	# sentences	# words	type of text	comment
khakaschiry	https://www.khakaschiry.ru	12013	271183	2914305	news article	perfect texts in the Khakas language
vk_ah_tashyl	https://vk.com/club31631018	828	4884	45917	vk posts	about 10-20 percent of the texts may be in Russian
vk_khakas_chiry	https://vk.com/khakas_chiry	2045	36793	401597	vk posts	about 10-20 percent of the texts may be in Russian
vk_khakas_radio	https://vk.com/club183612544	5138	36532	348552	vk posts	about 10-20 percent of the texts may be in Russian
vk_ust_chul	https://vk.com/club16506982	1267	6217	50187	vk posts	about 10-20 percent of the texts may be in Russian

Parallel corpus:

• located - khakas_texts_with_russian_translation/data
• kjh - khakas
• ru - russian
• source - https://khakas.altaica.ru/texts

import pandas as pd

mode = 'train'
path = f'khakas_texts_with_russian_translation/data/{mode}.csv'
df = pd.read_csv(path)
pairs = df[['kjh', 'ru']].values.tolist()
print(pairs[0])

	# sentences
train	79836
val	1000
test	1000

	train	val	test
# kjh words	589470	11734	11554
# ru words	642068	12893	12528