albertvillanova HF staff commited on
Commit
05c785c
1 Parent(s): 8cccce5

Delete loading script

Browse files
Files changed (1) hide show
  1. BA_Datensatz_V2.py +0 -82
BA_Datensatz_V2.py DELETED
@@ -1,82 +0,0 @@
1
- import datasets
2
- import csv
3
- import json
4
- import os
5
- import csv
6
-
7
- # Defining Access
8
- _DATA_URL = "https://huggingface.co/datasets/LennyBijan/BA_Datensatz_V2/resolve/main/data"
9
-
10
-
11
- class BA_Datensatz_V2(datasets.GeneratorBasedBuilder):
12
-
13
- def _info(self):
14
- return datasets.DatasetInfo(
15
- description="German Dataset that focuses on juristical data",
16
- features=datasets.Features({
17
- "clip_id": datasets.Value("string"),
18
- "path": datasets.Value("string"),
19
- "audio": datasets.Audio(sampling_rate=16_000),
20
- "sentence": datasets.Value("string"),
21
- "split": datasets.Value("string")
22
- }),
23
- supervised_keys=None,
24
- )
25
-
26
- def _split_generators(self, dl_manager):
27
- dl_manager.download_config.ignore_url_params = True
28
-
29
- audio_path = {}
30
- local_extracted_archive = {}
31
- metadata_path = {}
32
- split_type = {"train": datasets.Split.TRAIN, "test": datasets.Split.TEST}
33
- for split in split_type:
34
- audio_path[split] = dl_manager.download(f"{_DATA_URL}/audio_{split}.tgz")
35
- local_extracted_archive[split] = dl_manager.extract(audio_path[split]) if not dl_manager.is_streaming else None
36
- metadata_path[split] = dl_manager.download_and_extract(f"{_DATA_URL}/metadata_{split}.csv.gz")
37
- path_to_clips = "BA_Datensatz_V2"
38
-
39
- return [
40
- datasets.SplitGenerator(
41
- name=split_type[split],
42
- gen_kwargs={
43
- "local_extracted_archive": local_extracted_archive[split],
44
- "audio_files": dl_manager.iter_archive(audio_path[split]),
45
- "metadata_path": dl_manager.download_and_extract(metadata_path[split]),
46
- "path_to_clips": path_to_clips,
47
- "split": split # pass the split name to _generate_examples
48
- },
49
- ) for split in split_type
50
- ]
51
-
52
- def _generate_examples(self, audio_files, metadata_path, path_to_clips, local_extracted_archive, split):
53
- metadata = {}
54
- # Open and read the metadata CSV file.
55
- with open(metadata_path, "r", encoding="utf-8") as f:
56
- reader = csv.reader(f)
57
- for row in reader:
58
- filename, sentence = row
59
- clip_id = filename.split('_')[0]
60
- # Append the split name dynamically to the path_to_clips.
61
- path = os.path.join(path_to_clips, split +"/wav/"+ clip_id, filename)
62
- metadata[path] = {
63
- "clip_id": clip_id,
64
- "sentence": sentence,
65
- "path": path,
66
- }
67
- id_ = 0
68
- for path, file_content in audio_files: # No need to adjust the path if no local extraction.
69
- if path in metadata:
70
- result = dict(metadata[path])
71
- path = os.path.join(local_extracted_archive, path) if local_extracted_archive else path
72
- audio_data = {"path": path, "bytes": file_content.read()}
73
- result["audio"] = audio_data
74
- result["path"] = path
75
- yield id_, result
76
- id_ += 1
77
- else:
78
- print(f"No metadata entry for {path}")
79
-
80
-
81
-
82
-