-
Notifications
You must be signed in to change notification settings - Fork 2
/
dataset_creator.py
270 lines (260 loc) · 11.5 KB
/
dataset_creator.py
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
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
from optimizer import ensembler
import collections
import os
import logging
import logging.handlers
import logging.config
import numpy as np
import re
import sys
logging_dict = {
'version': 1,
'disable_existing_loggers': False,
'formatters': {
'date': {
'format': '%(asctime)s [%(name)-15s] %(levelname)-7s: %(message)s',
},
'no_date': {
'format': '[%(name)-15s] %(levelname)-7s: %(message)s',
},
},
'handlers': {
'file_handler': {
'class': 'logging.handlers.TimedRotatingFileHandler',
'filename': 'logs/optimizer.log',
'when': 'H',
'interval': 1,
'backupCount': 360,
'formatter': 'date',
'encoding': 'utf8'
},
'console': {
'class': 'logging.StreamHandler',
'level': 'DEBUG',
'formatter': 'no_date'
}
},
'loggers': {
'': {
'handlers': ['file_handler', 'console'],
'level': 'DEBUG',
'propagate': True
}
}
}
class DatasetCreator():
keys_single_letter = {
'C': 0, 'C#': 1, 'Db': 1, 'D': 2,
'D#': 3, 'Eb': 3, 'E': 4, 'F': 5,
'F#': 6, 'Gb': 6, 'G': 7, 'G#': 8,
'Ab': 8, 'A': 9, 'A#': 10, 'Bb': 10,
'B': 11,
'c': 12, 'c#': 13, 'db': 13, 'd': 14,
'd#': 15, 'eb': 15, 'e': 16, 'f': 17,
'f#': 18, 'gb': 18, 'g': 19, 'g#': 20,
'ab': 20, 'a': 21, 'a#': 22, 'bb': 22,
'b': 23
}
keys_duple = {
('c', 'major'): 0, ('c#', 'major'): 1,
('db', 'major'): 1, ('d', 'major'): 2,
('d#', 'major'): 3, ('eb', 'major'): 3,
('e', 'major'): 4, ('f', 'major'): 5,
('f#', 'major'): 6, ('gb', 'major'): 6,
('g', 'major'): 7, ('g#', 'major'): 8,
('ab', 'major'): 8, ('a', 'major'): 9,
('a#', 'major'): 10, ('bb', 'major'): 10,
('b', 'major'): 11,
('c', 'minor'): 12, ('c#', 'minor'): 13,
('db', 'minor'): 13, ('d', 'minor'): 14,
('d#', 'minor'): 15, ('eb', 'minor'): 15,
('e', 'minor'): 16, ('f', 'minor'): 17,
('f#', 'minor'): 18, ('gb', 'minor'): 18,
('g', 'minor'): 19, ('g#', 'minor'): 20,
('ab', 'minor'): 20, ('a', 'minor'): 21,
('a#', 'minor'): 22, ('bb', 'minor'): 22,
('b', 'minor'): 23
}
def __init__(self, dataset_folder):
self.logger = logging.getLogger('DatasetCreator')
self.logger.info('DatasetCreator() <- dataset_folder={}'.format(dataset_folder))
self.dataset_okay = False
self.has_features = False
self.ensemble = None
# Check critical stuff:
# dataset dir
if not os.path.isdir(dataset_folder):
self.logger.error('The dataset provided is not a folder')
exit()
self.dataset_folder = dataset_folder[:-1] if dataset_folder.endswith('/') else dataset_folder
if not any(c.isalpha() for c in self.dataset_folder):
self.logger.warning('The dataset folder, {}, does not contain any parsable name.'.format(self.dataset_folder))
self.name = 'unknown_dataset'
self.name = self.dataset_folder.rsplit('/')[-1]
self.logger.info('Dataset name: {}'.format(self.name))
self.feature_folder = os.path.join(dataset_folder, 'features')
# features folder
if not os.path.isdir(self.feature_folder):
self.logger.error('The dataset folder must contain a "features" folder')
exit()
self.files = []
self.files_no_extension = []
# files there and right formats
for f in os.listdir(self.feature_folder):
tokens = f.rsplit('.', 1)
if tokens[1] not in ['mid', 'csv', 'wav']:
self.logger.error('Unknown file type {}. Expecting .wav, .csv, or .mid'.format(f))
exit()
no_extension = tokens[0]
self.files.append(f)
self.files_no_extension.append(no_extension)
if not self.files:
self.logger.error('The features folder seems to be empty.')
exit()
# Now, not so critical stuff
self.annotation_folder = os.path.join(dataset_folder, 'annotations')
if not os.path.isdir(self.annotation_folder):
self.logger.warning('The dataset folder does not contain annotations, assuming this was expected')
self.annotation_folder = None
# And not critical at all
self.splits_folder = os.path.join(dataset_folder, 'splits')
if not os.path.isdir(self.splits_folder):
self.logger.info('No data splits provided. Working on the entire dataset')
self.splits_folder = None
self.features = collections.OrderedDict()
self.annotations = collections.OrderedDict()
self.splits = {}
if self.splits_folder:
for f in os.listdir(self.splits_folder):
tokens = f.rsplit('.', 1)
filepath = os.path.join(self.splits_folder, f)
if len(tokens) < 2 or tokens[1] not in ['txt']:
self.logger.warning('Expected {} to be a .txt file.'.format(filepath))
no_extension = tokens[0]
with open(filepath, newline='') as fd:
split = fd.readlines()
split = [x.strip() for x in split]
self.splits[no_extension] = split
self.logger.info('Found the following splits: {}'.format(self.splits))
# If we are here, everything should be okay
self.dataset_okay = True
def parse_label(self, annotation):
annotation = annotation.strip().split()
label = -1
if len(annotation) == 1:
key = annotation[0]
if key in DatasetCreator.keys_single_letter:
label = DatasetCreator.keys_single_letter[key]
elif len(annotation) == 2:
key = (annotation[0].lower(), annotation[1].lower())
if key in DatasetCreator.keys_duple:
label = DatasetCreator.keys_duple[key]
if label == -1:
self.logger.error('Did not recognize the format of label {}'.format(annotation))
return label
def compute_features(self, key_profiles, key_transitions, mixed_profiles=False):
self.logger.info('compute_features() <- key_profiles={}, key_transitions={}, mixed_profiles={}'.format(key_profiles, key_transitions, mixed_profiles))
if not self.dataset_okay:
return
self.key_profiles = key_profiles
self.key_transitions = key_transitions
self.mixed_profiles = mixed_profiles
ens = ensembler.Ensembler(key_profiles, key_transitions)
for i in range(len(self.files)):
feature_filepath = os.path.join(self.feature_folder, self.files[i])
features = ens.evaluate(feature_filepath, mixed_profiles)
features = [f for l in features for f in l]
feature_array = np.array(features)
self.features[self.files_no_extension[i]] = feature_array
if self.annotation_folder:
annotation_name = '{}.key'.format(self.files_no_extension[i])
annotation_filepath = os.path.join(self.annotation_folder, annotation_name)
if not os.path.exists(annotation_filepath):
self.logger.error('Could not find file {}'.format(annotation_filepath))
return
with open(annotation_filepath, newline='') as fd:
annotation_data = fd.readlines()[0]
label = self.parse_label(annotation_data)
self.annotations[self.files_no_extension[i]] = label
self.has_features = True
def write(self, output_dir='./datasets'):
self.logger.info('write() <- output_dir={}, dataset_name={}'.format(output_dir, self.name))
if not self.has_features:
return
if not os.path.exists(output_dir):
self.logger.warning('Output folder {} does not exist. Trying to create it'.format(output_dir))
os.makedirs(output_dir)
self.logger.warning('Success.')
ensemble_filename = '{}_ensemble.txt'.format(self.name)
ensemble_filepath = os.path.join(output_dir, ensemble_filename)
self.logger.info('writing {}'.format(ensemble_filepath))
with open(ensemble_filepath, 'w') as fd:
fd.write('{}\n'.format(', '.join(self.key_profiles)))
fd.write('{}\n'.format(', '.join(self.key_transitions)))
fd.write('mixed_profiles={}\n'.format(self.mixed_profiles))
features_filename = '{}_features.pkl'.format(self.name)
feature_filepath = os.path.join(output_dir, features_filename)
feature_array = list(self.features.values())
self.logger.info('writing {}'.format(feature_filepath))
np.array(feature_array).dump(feature_filepath)
if self.annotations:
annotation_filename = '{}_annotations.pkl'.format(self.name)
annotation_filepath = os.path.join(output_dir, annotation_filename)
annotation_array = list(self.annotations.values())
self.logger.info('writing {}'.format(annotation_filepath))
np.array(annotation_array).dump(annotation_filepath)
for k,v in self.splits.items():
split_features = []
split_annotations = []
for f in v:
if f not in self.features:
self.logger.error('Could not find file {} in {} split. This split will be ignored'.format(f, k))
split_features = []
split_annotations = []
break
split_features.append(self.features[f])
if self.annotations:
if f not in self.annotations:
self.logger.error('Could not find annotation {} in {} split. This split will be ignored'.format(f, k))
split_features = []
split_annotations = []
break
split_annotations.append(self.annotations[f])
if not split_features:
continue
features_filename = '{}-{}_features.pkl'.format(self.name, k)
feature_filepath = os.path.join(output_dir, features_filename)
self.logger.info('writing {}'.format(feature_filepath))
np.array(split_features).dump(feature_filepath)
if split_annotations:
annotation_filename = '{}-{}_annotations.pkl'.format(self.name, k)
annotation_filepath = os.path.join(output_dir, annotation_filename)
self.logger.info('writing {}'.format(annotation_filepath))
np.array(split_annotations).dump(annotation_filepath)
if __name__ == '__main__':
if not os.path.exists('logs'):
os.makedirs('logs')
logging.config.dictConfig(logging_dict)
logger = logging.getLogger('dataset_creator')
if len(sys.argv) != 2 or not os.path.isdir(sys.argv[1]):
logger.error('You need to provide a dataset')
exit()
dataset = sys.argv[1]
key_profiles = [
'aarden_essen',
'temperley',
'krumhansl_kessler',
'bellman_budge',
'albrecht_shanahan1',
'albrecht_shanahan2',
'sapp',
'simple_natural_minor',
'simple_harmonic_minor',
'simple_melodic_minor',]
key_transitions = [
'ktg_exponential5',
'ktg_exponential10',
'ktg_exponential15',]
dc = DatasetCreator(dataset)
dc.compute_features(key_profiles, key_transitions)
dc.write()