from datasets import load_dataset
from tokenizers import ByteLevelBPETokenizer
from pythainlp.tokenize import word_tokenize

# load dataset
dataset = load_dataset("oscar", "unshuffled_deduplicated_th", split="train")

# Instantiate tokenizer
tokenizer = ByteLevelBPETokenizer()


def th_tokenize(text):
    result = " ".join(word_tokenize(text, engine="newmm", keep_whitespace=False))
    return result


def batch_iterator(batch_size=1000):
    for i in range(0, len(dataset), batch_size):
        yield [th_tokenize(text) for text in dataset[i : i + batch_size]["text"]]


# Customized training
tokenizer.train_from_iterator(
    batch_iterator(),
    vocab_size=50265,
    min_frequency=2,
    special_tokens=["<s>", "<pad>", "</s>", "<unk>", "<mask>",],
)

# Save files to disk
tokenizer.save(f"./tokenizer.json")