phdkhanh2507 commited on
Commit
781d61a
·
verified ·
1 Parent(s): 09f9688

Delete vietnamese-speaker-video.py

Browse files
Files changed (1) hide show
  1. vietnamese-speaker-video.py +0 -131
vietnamese-speaker-video.py DELETED
@@ -1,131 +0,0 @@
1
- # Copyright 2023 Thinh T. Duong
2
- import os
3
- import datasets
4
- from huggingface_hub import HfFileSystem
5
- from typing import List, Tuple
6
-
7
-
8
- logger = datasets.logging.get_logger(__name__)
9
- fs = HfFileSystem()
10
-
11
-
12
- _CITATION = """
13
-
14
- """
15
- _DESCRIPTION = """
16
- This dataset contain videos of Vietnamese speakers.
17
- """
18
- _HOMEPAGE = "https://github.com/tanthinhdt/vietnamese-av-asr"
19
- _REPO_PATH = "datasets/phdkhanh2507/vietnamese-speaker-video"
20
- _REPO_URL = f"https://huggingface.co/{_REPO_PATH}/resolve/main"
21
- _URLS = {
22
- "meta": f"{_REPO_URL}/metadata/" + "{channel}.parquet",
23
- "video": f"{_REPO_URL}/video/" + "{channel}.zip",
24
- }
25
- _CONFIGS = ["all"]
26
- if fs.exists(_REPO_PATH + "/metadata"):
27
- _CONFIGS.extend([
28
- os.path.basename(file_name)[:-8]
29
- for file_name in fs.listdir(_REPO_PATH + "/metadata", detail=False)
30
- if file_name.endswith(".parquet")
31
- ])
32
-
33
-
34
- class VietnameseSpeakerVideoConfig(datasets.BuilderConfig):
35
- """Vietnamese Speaker Video configuration."""
36
-
37
- def __init__(self, name, **kwargs) -> None:
38
- """
39
- :param name: Name of subset.
40
- :param kwargs: Arguments.
41
- """
42
- super().__init__(
43
- name=name,
44
- version=datasets.Version("1.0.0"),
45
- description=_DESCRIPTION,
46
- **kwargs,
47
- )
48
-
49
-
50
- class VietnameseSpeakerVideo(datasets.GeneratorBasedBuilder):
51
- """Vietnamese Speaker Video dataset."""
52
-
53
- BUILDER_CONFIGS = [VietnameseSpeakerVideoConfig(name) for name in _CONFIGS]
54
- DEFAULT_CONFIG_NAME = "all"
55
-
56
- def _info(self) -> datasets.DatasetInfo:
57
- features = datasets.Features({
58
- "id": datasets.Value("string"),
59
- "channel": datasets.Value("string"),
60
- "video": datasets.Value("string"),
61
- "fps": datasets.Value("int8"),
62
- "sampling_rate": datasets.Value("int64"),
63
- })
64
-
65
- return datasets.DatasetInfo(
66
- description=_DESCRIPTION,
67
- features=features,
68
- homepage=_HOMEPAGE,
69
- citation=_CITATION,
70
- )
71
-
72
- def _split_generators(
73
- self, dl_manager: datasets.DownloadManager
74
- ) -> List[datasets.SplitGenerator]:
75
- """
76
- Get splits.
77
- :param dl_manager: Download manager.
78
- :return: Splits.
79
- """
80
- config_names = _CONFIGS[1:] if self.config.name == "all" else [self.config.name]
81
-
82
- metadata_paths = dl_manager.download(
83
- [_URLS["meta"].format(channel=channel) for channel in config_names]
84
- )
85
- video_dirs = dl_manager.download_and_extract(
86
- [_URLS["video"].format(channel=channel) for channel in config_names]
87
- )
88
-
89
- video_dict = {
90
- channel: video_dir for channel, video_dir in zip(config_names, video_dirs)
91
- }
92
-
93
- return [
94
- datasets.SplitGenerator(
95
- name=datasets.Split.TRAIN,
96
- gen_kwargs={
97
- "metadata_paths": metadata_paths,
98
- "video_dict": video_dict,
99
- },
100
- ),
101
- ]
102
-
103
- def _generate_examples(
104
- self, metadata_paths: List[str],
105
- video_dict: dict,
106
- ) -> Tuple[int, dict]:
107
- """
108
- Generate examples from metadata.
109
- :param metadata_paths: Paths to metadata.
110
- :param video_dict: Paths to directory containing videos.
111
- :yield: Example.
112
- """
113
- dataset = datasets.load_dataset(
114
- "parquet",
115
- data_files=metadata_paths,
116
- split="train",
117
- use_auth_token=True,
118
- )
119
- for i, sample in enumerate(dataset):
120
- channel = sample["channel"]
121
- video_path = os.path.join(
122
- video_dict[channel], channel, sample["id"] + ".avi"
123
- )
124
-
125
- yield i, {
126
- "id": sample["id"],
127
- "channel": channel,
128
- "video": video_path,
129
- "fps": sample["fps"],
130
- "sampling_rate": sample["sampling_rate"],
131
- }