examples/vm_sound_classification/step_2_make_vocabulary.py

#!/usr/bin/python3
# -*- coding: utf-8 -*-
import argparse
import os
from pathlib import Path
import sys

pwd = os.path.abspath(os.path.dirname(__file__))
sys.path.append(os.path.join(pwd, "../../"))

import pandas as pd

from toolbox.torch.utils.data.vocabulary import Vocabulary


def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument("--vocabulary_dir", default="vocabulary", type=str)

    parser.add_argument("--train_dataset", default="train.xlsx", type=str)
    parser.add_argument("--valid_dataset", default="valid.xlsx", type=str)

    args = parser.parse_args()
    return args


def main():
    args = get_args()

    train_dataset = pd.read_excel(args.train_dataset)
    valid_dataset = pd.read_excel(args.valid_dataset)

    vocabulary = Vocabulary()

    # train
    for i, row in train_dataset.iterrows():
        label = row["labels"]
        vocabulary.add_token_to_namespace(label, namespace="labels")

    # valid
    for i, row in valid_dataset.iterrows():
        label = row["labels"]
        vocabulary.add_token_to_namespace(label, namespace="labels")

    vocabulary.save_to_files(args.vocabulary_dir)

    return


if __name__ == "__main__":
    main()