Spaces:
Running
Running
File size: 4,052 Bytes
69ad385 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 |
#!/usr/bin/python3
# -*- coding: utf-8 -*-
import argparse
from glob import glob
import json
import os
from pathlib import Path
import random
import sys
pwd = os.path.abspath(os.path.dirname(__file__))
sys.path.append(os.path.join(pwd, "../../"))
import pandas as pd
from scipy.io import wavfile
from tqdm import tqdm
def get_args():
parser = argparse.ArgumentParser()
parser.add_argument("--file_dir", default="./", type=str)
parser.add_argument("--task", default="default", type=str)
parser.add_argument("--filename_patterns", 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 get_dataset(args):
filename_patterns = args.filename_patterns
filename_patterns = filename_patterns.split(" ")
print(filename_patterns)
file_dir = Path(args.file_dir)
file_dir.mkdir(exist_ok=True)
global_label_map = {
"bell": "bell",
"white_noise": "white_noise",
"low_white_noise": "white_noise",
"high_white_noise": "noise",
"music": "music",
"mute": "mute",
"noise": "noise",
"noise_mute": "noise_mute",
"voice": "voice",
"voicemail": "voicemail",
}
country_label_map = {
"bell": "voicemail",
"white_noise": "non_voicemail",
"low_white_noise": "non_voicemail",
"hight_white_noise": "non_voicemail",
"music": "non_voicemail",
"mute": "non_voicemail",
"noise": "non_voicemail",
"noise_mute": "non_voicemail",
"voice": "non_voicemail",
"voicemail": "voicemail",
"non_voicemail": "non_voicemail",
}
result = list()
for filename_pattern in filename_patterns:
filename_list = glob(filename_pattern)
for filename in tqdm(filename_list):
filename = Path(filename)
sample_rate, signal = wavfile.read(filename.as_posix())
if len(signal) < sample_rate * 2:
continue
folder = filename.parts[-2]
country = filename.parts[-4]
if folder not in global_label_map.keys():
continue
if folder not in country_label_map.keys():
continue
global_label = global_label_map[folder]
country_label = country_label_map[folder]
random1 = random.random()
random2 = random.random()
result.append({
"filename": filename,
"folder": folder,
"category": country,
"global_labels": global_label,
"country_labels": country_label,
"random1": random1,
"random2": random2,
"flag": "TRAIN" if random2 < 0.8 else "TEST",
})
df = pd.DataFrame(result)
pivot_table = pd.pivot_table(df, index=["global_labels"], values=["filename"], aggfunc="count")
print(pivot_table)
df = df.sort_values(by=["random1"], ascending=False)
df.to_excel(
file_dir / "dataset.xlsx",
index=False,
# encoding="utf_8_sig"
)
return
def split_dataset(args):
"""分割训练集, 测试集"""
file_dir = Path(args.file_dir)
file_dir.mkdir(exist_ok=True)
df = pd.read_excel(file_dir / "dataset.xlsx")
train = list()
test = list()
for i, row in df.iterrows():
flag = row["flag"]
if flag == "TRAIN":
train.append(row)
else:
test.append(row)
train = pd.DataFrame(train)
train.to_excel(
args.train_dataset,
index=False,
# encoding="utf_8_sig"
)
test = pd.DataFrame(test)
test.to_excel(
args.valid_dataset,
index=False,
# encoding="utf_8_sig"
)
return
def main():
args = get_args()
get_dataset(args)
split_dataset(args)
return
if __name__ == "__main__":
main()
|