imvladikon commited on
Commit
395ac9f
1 Parent(s): f3c3607

Create to_generate.py

Browse files
Files changed (1) hide show
  1. to_generate.py +131 -0
to_generate.py ADDED
@@ -0,0 +1,131 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env python3
2
+ # -*- coding: utf-8 -*-
3
+ import glob
4
+ import json
5
+ import os
6
+ from functools import partial
7
+ from pathlib import Path
8
+
9
+ import datasets
10
+
11
+ VERSION = datasets.Version("0.0.1")
12
+
13
+ SAMPLE_RATE = 16000
14
+
15
+ URLS = {'introduction_psychology': 'data/introduction_psychology.zip',
16
+ 'a_descriptive_statistics': 'data/a_descriptive_statistics.zip',
17
+ 'inherited_how_does_the_internet_work': 'data/inherited_how_does_the_internet_work.zip',
18
+ 'data_structures': 'data/data_structures.zip',
19
+ 'computational_thinking': 'data/computational_thinking.zip',
20
+ 'physics_intro': 'data/physics_intro.zip',
21
+ 'data_intro': 'data/data_intro.zip',
22
+ 'introduction_to_the_philosophy_of_education': 'data/introduction_to_the_philosophy_of_education.zip',
23
+ 'yad_vashem': 'data/yad_vashem.zip',
24
+ 'algorithms': 'data/algorithms.zip',
25
+ 'blue-and-white_tv': 'data/blue-and-white_tv.zip',
26
+ 'covid_mindfulness': 'data/covid_mindfulness.zip',
27
+ 'preparation_for_a_job_interview-_ways_to_success': 'data/preparation_for_a_job_interview-_ways_to_success.zip',
28
+ 'what_is_the_world_introduction_to_general_chemistry': 'data/what_is_the_world_introduction_to_general_chemistry.zip',
29
+ 'networking_create_a_network_of_professional_progress': 'data/networking_create_a_network_of_professional_progress.zip',
30
+ 'how_to_learn': 'data/how_to_learn.zip',
31
+ 'post_modern_education': 'data/post_modern_education.zip',
32
+ 'introduction_to_renewable_energy': 'data/introduction_to_renewable_energy.zip',
33
+ 'science_communication': 'data/science_communication.zip',
34
+ 'introduction_to_physics_-_mechanics': 'data/introduction_to_physics_-_mechanics.zip',
35
+ 'negotiations_different_cultures_and_what_between_them': 'data/negotiations_different_cultures_and_what_between_them.zip',
36
+ 'computation_models': 'data/computation_models.zip',
37
+ 'object_oriented_programming': 'data/object_oriented_programming.zip',
38
+ 'from_another_angle_math': 'data/from_another_angle_math.zip'}
39
+
40
+
41
+ class CampusHebrewSpeech(datasets.GeneratorBasedBuilder):
42
+ BUILDER_CONFIGS = [
43
+ datasets.BuilderConfig(name="campus_hebrew_speech",
44
+ version=VERSION,
45
+ description=f"Campus Hebrew Speech Recognition dataset")
46
+ ]
47
+
48
+ def _info(self):
49
+ return datasets.DatasetInfo(
50
+ description="Hebrew speech datasets",
51
+ features=datasets.Features(
52
+ {
53
+ "uid": datasets.Value("string"),
54
+ "file_id": datasets.Value("string"),
55
+ "audio": datasets.Audio(sampling_rate=SAMPLE_RATE),
56
+ "sentence": datasets.Value("string"),
57
+ "n_segment": datasets.Value("int32"),
58
+ "duration_ms": datasets.Value("float32"),
59
+ "language": datasets.Value("string"),
60
+ "sample_rate": datasets.Value("int32"),
61
+ "course": datasets.Value("string"),
62
+ "sentence_length": datasets.Value("int32"),
63
+ "n_tokens": datasets.Value("int32"),
64
+ }
65
+ ),
66
+ supervised_keys=("audio", "sentence"),
67
+ homepage="https://huggingface.co/datasets/imvladikon/hebrew_speech_campus",
68
+ citation="TODO",
69
+ )
70
+
71
+ def _split_generators(self, dl_manager):
72
+ # course_links = {
73
+ # course: dl_manager.download_and_extract(link) + "/" + course
74
+ # for course, link in URLS.items()
75
+ # }
76
+ course_links = {
77
+ 'from_another_angle-_mathematics_teaching_practices': 'data/from_another_angle-_mathematics_teaching_practices',
78
+ 'introduction_to_renewable_energy': 'data/introduction_to_renewable_energy',
79
+ 'negotiations_different_cultures_and_what_between_them': 'data/negotiations_different_cultures_and_what_between_them',
80
+ 'introduction_to_physics_-_mechanics': 'data/introduction_to_physics_-_mechanics',
81
+ 'networking-_create_a_network_of_professional_progress': 'data/networking-_create_a_network_of_professional_progress',
82
+ 'post_modern_education': 'data/post_modern_education',
83
+ 'physics_intro': 'data/physics_intro',
84
+ 'algorithms': 'data/algorithms',
85
+ 'blue-and-white_tv': 'data/blue-and-white_tv',
86
+ 'what_is_the_world-_introduction_to_general_chemistry': 'data/what_is_the_world-_introduction_to_general_chemistry',
87
+ 'computational_thinking': 'data/computational_thinking',
88
+ 'preparation_for_a_job_interview-_ways_to_success': 'data/preparation_for_a_job_interview-_ways_to_success',
89
+ 'a_descriptive_statistics': 'data/a_descriptive_statistics',
90
+ 'yad_vashem': 'data/yad_vashem',
91
+ 'introduction_to_the_philosophy_of_education': 'data/introduction_to_the_philosophy_of_education',
92
+ 'covid_mindfulness': 'data/covid_mindfulness',
93
+ 'data_structures': 'data/data_structures',
94
+ 'computation_models': 'data/computation_models',
95
+ 'introduction_psychology': 'data/introduction_psychology',
96
+ 'how_to_learn': 'data/how_to_learn',
97
+ 'inherited_-_how_does_the_internet_work': 'data/inherited_-_how_does_the_internet_work',
98
+ 'object_oriented_programming': 'data/object_oriented_programming',
99
+ 'science_communication': 'data/science_communication',
100
+ 'data_intro': 'data/data_intro'}
101
+
102
+ return [
103
+ datasets.SplitGenerator(
104
+ name=datasets.Split.TRAIN,
105
+ gen_kwargs={
106
+ "course_links": course_links,
107
+ "split": "train"},
108
+ )
109
+ ]
110
+
111
+ def _generate_examples(self, course_links, split):
112
+ idx = 0
113
+ for course, root_path in course_links.items():
114
+ root_path = "/content/hebrew_speech_campus/" + root_path
115
+ for metadata_file in Path(root_path).glob("*.json"):
116
+ audio_file = Path(metadata_file).stem + ".wav"
117
+ metadata = json.load(open(metadata_file, encoding="utf-8"))
118
+ yield idx, {
119
+ "uid": metadata["file"].split("_")[0],
120
+ "file_id": Path(metadata["file"]).stem,
121
+ "audio": os.path.join(root_path, audio_file),
122
+ "sentence": metadata["text"],
123
+ "n_segment": metadata["n_segment"],
124
+ "duration_ms": 1000 * metadata["duration"],
125
+ "language": metadata["language"],
126
+ "sample_rate": SAMPLE_RATE,
127
+ "course": course,
128
+ "sentence_length": len(metadata["text"]),
129
+ "n_tokens": metadata["text"].count(" ") + 1,
130
+ }
131
+ idx += 1