my-python-scripts/make_individual.py at main · sage-khan/my-python-scripts · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
import os
import pandas as pd
import librosa


train_dir = "train"
os.makedirs(train_dir, exist_ok=True)

test_dir = "test"
os.makedirs(test_dir, exist_ok=True)
data_dir = "/home/metanet/ProgramFiles/kaldi/kaldi/egs/ASR-for-urdu/s5/wav/"


def get_audio_length(full_path_string):
    y, sr = librosa.load(full_path_string)  # librosa.ex(audio_filepath[:-4]))
    end_time = librosa.get_duration(y=y, sr=sr)
    return end_time


def get_utt_id(full_path_string, number):
    _, folder_name, _ = full_path_string.split("/")
    utt_id = folder_name + f"_{number}"
    return utt_id


def get_rec_id(full_path_string):
    _, folder_name, filename = full_path_string.split("/")
    rec_id = folder_name.split("_")[0] + "_" + filename[:-4] + "_wav"

    return rec_id


def get_segments(full_path_string):
    _, folder_name, filename = full_path_string.split("/")
    rec_id = folder_name.split("_")[0] + "_" + filename
    return rec_id


def make_utt2spk(dataframe):
    utt2spk = [f"{utt} {utt}" for utt in dataframe.utt_id]
    sample_num = len(utt2spk)
    train_num = int(sample_num * 0.9)
    train_data, test_data = split_data(utt2spk)

    test_utt2spk = os.path.join(test_dir, "utt2spk")
    with open(test_utt2spk, "w") as u2s:
        u2s.write("\n".join(test_data))
    train_utt2spk = os.path.join(train_dir, "utt2spk")
    with open(train_utt2spk, "w") as u2s:
        u2s.write("\n".join(train_data))


def make_wavscp(dataframe):
    utt_scp = [f"{rec} {data_dir}{file_path}" for rec, file_path in zip(dataframe.rec_id, dataframe.wav_filename)]
    sample_num = len(utt_scp)
    train_num = int(sample_num * 0.9)
    train_data, test_data = split_data(utt_scp)

    test_utt_scp = os.path.join(test_dir, "wav.scp")
    with open(test_utt_scp, "w") as scp:
        scp.write("\n".join(test_data))

    train_utt_scp = os.path.join(train_dir, "wav.scp")
    with open(train_utt_scp, "w") as scp:
        scp.write("\n".join(train_data))


def make_segments(df):
    segments = [f"{utt} {rec_id} 0 {end}" for utt, rec_id, end in zip(df.utt_id, df.rec_id, df.end)]
    sample_num = len(segments)
    train_num = int(sample_num * 0.9)
    train_data, test_data = split_data(segments)

    test_segments = os.path.join(test_dir, "segments")
    with open(test_segments, "w") as seg:
        seg.write("\n".join(test_data))

    train_segments = os.path.join(train_dir, "segments")
    with open(train_segments, "w") as seg:
        seg.write("\n".join(train_data))


def make_text(dataframe):
    utt_text = [f"{utt} {text}" for utt, text in zip(dataframe.utt_id, dataframe.transcript)]
    sample_num = len(utt_text)
    train_num = int(sample_num * 0.9)
    train_data, test_data = split_data(utt_text)

    test_text = os.path.join(test_dir, "text")
    with open(test_text, "w") as tf:
        tf.write("\n".join(test_data))
    train_text = os.path.join(train_dir, "text")
    with open(train_text, "w") as tf:
        tf.write("\n".join(train_data))


def split_data(data):
    train = []
    test = []
    for i in range(len(data)):
        if i % 10 == 0:
            test.append(data[i])
        else:
            train.append(data[i])
    return train, test

def main(full_source):

    if not os.path.isfile(full_source):
        print("No such file exist: {}".format(full_source))
        return
    # open tsv file
    print("Reading {} ...".format(full_source))
    dataframe = pd.read_csv(full_source)
    header = dataframe.columns # ["wav_filename", "transcript"]
    print(header)

    dataframe_copied = dataframe.copy()
    dataframe_copied["_id_"] = list(range(len(dataframe_copied)))
    dataframe_copied["utt_id"] = dataframe_copied.apply(lambda x: get_utt_id(x['wav_filename'], x['_id_']), axis=1)
    # dataframe_copied["spk_id"] = dataframe_copied["wav_filename"].apply(get_utt_id)


    dataframe_copied["rec_id"] = dataframe_copied["wav_filename"].apply(get_rec_id)
    # dataframe_copied["start"] = dataframe_copied["wav_filename"].apply(get_utt_id)
    dataframe_copied["end"] = dataframe_copied["wav_filename"].apply(get_audio_length)

    make_text(dataframe_copied)
    make_utt2spk(dataframe_copied)
    make_wavscp(dataframe_copied)
    make_segments(dataframe_copied)
    pass


if __name__ == '__main__':
    source_full_data = 'whole_data_urdu_16k.csv'
    main(source_full_data)