[134fd7]: / clinical_ts / timeseries_utils.py

Download this file

829 lines (686 with data), 35.6 kB

  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
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
# AUTOGENERATED! DO NOT EDIT! File to edit: nbs/A_timeseries_utils.ipynb (unless otherwise specified).
__all__ = ['butter_filter', 'butter_filter_frequency_response', 'apply_butter_filter', 'save_dataset', 'load_dataset',
'dataset_add_chunk_col', 'dataset_add_length_col', 'dataset_add_labels_col', 'dataset_add_mean_col',
'dataset_add_median_col', 'dataset_add_std_col', 'dataset_add_iqr_col', 'dataset_get_stats',
'npys_to_memmap_batched', 'npys_to_memmap', 'reformat_as_memmap', 'TimeseriesDatasetCrops', 'RandomCrop',
'CenterCrop', 'GaussianNoise', 'Rescale', 'ToTensor', 'Normalize', 'NormalizeBatch', 'ButterFilter',
'ChannelFilter', 'Transform', 'TupleTransform', 'aggregate_predictions']
# Cell
import numpy as np
import pandas as pd
import torch
import torch.utils.data
from torch import nn
from pathlib import Path
from scipy.stats import iqr
try:
import pickle5 as pickle
except ImportError as e:
import pickle
#Note: due to issues with the numpy rng for multiprocessing (https://github.com/pytorch/pytorch/issues/5059) that could be fixed by a custom worker_init_fn we use random throught for convenience
import random
#Note: multiprocessing issues with python lists and dicts (https://github.com/pytorch/pytorch/issues/13246) and pandas dfs (https://github.com/pytorch/pytorch/issues/5902)
import multiprocessing as mp
from skimage import transform
import warnings
warnings.filterwarnings("ignore", category=UserWarning)
from scipy.signal import butter, sosfilt, sosfiltfilt, sosfreqz
from tqdm.auto import tqdm
# Cell
#https://stackoverflow.com/questions/12093594/how-to-implement-band-pass-butterworth-filter-with-scipy-signal-butter
def butter_filter(lowcut=10, highcut=20, fs=50, order=5, btype='band'):
'''returns butterworth filter with given specifications'''
nyq = 0.5 * fs
low = lowcut / nyq
high = highcut / nyq
sos = butter(order, [low, high] if btype=="band" else (low if btype=="low" else high), analog=False, btype=btype, output='sos')
return sos
def butter_filter_frequency_response(filter):
'''returns frequency response of a given filter (result of call of butter_filter)'''
w, h = sosfreqz(filter)
#gain vs. freq(Hz)
#plt.plot((fs * 0.5 / np.pi) * w, abs(h))
return w,h
def apply_butter_filter(data, filter, forwardbackward=True):
'''pass filter from call of butter_filter to data (assuming time axis at dimension 0)'''
if(forwardbackward):
return sosfiltfilt(filter, data, axis=0)
else:
data = sosfilt(filter, data, axis=0)
# Cell
def save_dataset(df,lbl_itos,mean,std,target_root,filename_postfix="",protocol=4):
target_root = Path(target_root)
df.to_pickle(target_root/("df"+filename_postfix+".pkl"), protocol=protocol)
if(isinstance(lbl_itos,dict)):#dict as pickle
outfile = open(target_root/("lbl_itos"+filename_postfix+".pkl"), "wb")
pickle.dump(lbl_itos, outfile, protocol=protocol)
outfile.close()
else:#array
np.save(target_root/("lbl_itos"+filename_postfix+".npy"),lbl_itos)
np.save(target_root/("mean"+filename_postfix+".npy"),mean)
np.save(target_root/("std"+filename_postfix+".npy"),std)
def load_dataset(target_root,filename_postfix="",df_mapped=True):
target_root = Path(target_root)
# if(df_mapped):
# df = pd.read_pickle(target_root/("df_memmap"+filename_postfix+".pkl"))
# else:
# df = pd.read_pickle(target_root/("df"+filename_postfix+".pkl")
### due to pickle 5 protocol error
if(df_mapped):
df = pickle.load(open(target_root/("df_memmap"+filename_postfix+".pkl"), "rb"))
else:
df = pickle.load(open(target_root/("df"+filename_postfix+".pkl"), "rb"))
if((target_root/("lbl_itos"+filename_postfix+".pkl")).exists()):#dict as pickle
infile = open(target_root/("lbl_itos"+filename_postfix+".pkl"), "rb")
lbl_itos=pickle.load(infile)
infile.close()
else:#array
lbl_itos = np.load(target_root/("lbl_itos"+filename_postfix+".npy"))
mean = np.load(target_root/("mean"+filename_postfix+".npy"))
std = np.load(target_root/("std"+filename_postfix+".npy"))
return df, lbl_itos, mean, std
# Cell
def dataset_add_chunk_col(df, col="data"):
'''add a chunk column to the dataset df'''
df["chunk"]=df.groupby(col).cumcount()
def dataset_add_length_col(df, col="data", data_folder=None):
'''add a length column to the dataset df'''
df[col+"_length"]=df[col].apply(lambda x: len(np.load(x if data_folder is None else data_folder/x, allow_pickle=True)))
def dataset_add_labels_col(df, col="label", data_folder=None):
'''add a column with unique labels in column col'''
df[col+"_labels"]=df[col].apply(lambda x: list(np.unique(np.load(x if data_folder is None else data_folder/x, allow_pickle=True))))
def dataset_add_mean_col(df, col="data", axis=(0), data_folder=None):
'''adds a column with mean'''
df[col+"_mean"]=df[col].apply(lambda x: np.mean(np.load(x if data_folder is None else data_folder/x, allow_pickle=True),axis=axis))
def dataset_add_median_col(df, col="data", axis=(0), data_folder=None):
'''adds a column with median'''
df[col+"_median"]=df[col].apply(lambda x: np.median(np.load(x if data_folder is None else data_folder/x, allow_pickle=True),axis=axis))
def dataset_add_std_col(df, col="data", axis=(0), data_folder=None):
'''adds a column with mean'''
df[col+"_std"]=df[col].apply(lambda x: np.std(np.load(x if data_folder is None else data_folder/x, allow_pickle=True),axis=axis))
def dataset_add_iqr_col(df, col="data", axis=(0), data_folder=None):
'''adds a column with mean'''
df[col+"_iqr"]=df[col].apply(lambda x: iqr(np.load(x if data_folder is None else data_folder/x, allow_pickle=True),axis=axis))
def dataset_get_stats(df, col="data", simple=True):
'''creates (weighted) means and stds from mean, std and length cols of the df'''
if(simple):
return df[col+"_mean"].mean(), df[col+"_std"].mean()
else:
#https://notmatthancock.github.io/2017/03/23/simple-batch-stat-updates.html
#or https://gist.github.com/thomasbrandon/ad5b1218fc573c10ea4e1f0c63658469
def combine_two_means_vars(x1,x2):
(mean1,var1,n1) = x1
(mean2,var2,n2) = x2
mean = mean1*n1/(n1+n2)+ mean2*n2/(n1+n2)
var = var1*n1/(n1+n2)+ var2*n2/(n1+n2)+n1*n2/(n1+n2)/(n1+n2)*np.power(mean1-mean2,2)
return (mean, var, (n1+n2))
def combine_all_means_vars(means,vars,lengths):
inputs = list(zip(means,vars,lengths))
result = inputs[0]
for inputs2 in inputs[1:]:
result= combine_two_means_vars(result,inputs2)
return result
means = list(df[col+"_mean"])
vars = np.power(list(df[col+"_std"]),2)
lengths = list(df[col+"_length"])
mean,var,length = combine_all_means_vars(means,vars,lengths)
return mean, np.sqrt(var)
# Cell
def npys_to_memmap_batched(npys, target_filename, max_len=0, delete_npys=True, batch_length=900000):
memmap = None
start = np.array([0])#start_idx in current memmap file (always already the next start- delete last token in the end)
length = []#length of segment
filenames= []#memmap files
file_idx=[]#corresponding memmap file for sample
shape=[]#shapes of all memmap files
data = []
data_lengths=[]
dtype = None
for idx,npy in tqdm(list(enumerate(npys))):
data.append(np.load(npy, allow_pickle=True))
data_lengths.append(len(data[-1]))
if(idx==len(npys)-1 or np.sum(data_lengths)>batch_length):#flush
data = np.concatenate(data)
if(memmap is None or (max_len>0 and start[-1]>max_len)):#new memmap file has to be created
if(max_len>0):
filenames.append(target_filename.parent/(target_filename.stem+"_"+str(len(filenames))+".npy"))
else:
filenames.append(target_filename)
shape.append([np.sum(data_lengths)]+[l for l in data.shape[1:]])#insert present shape
if(memmap is not None):#an existing memmap exceeded max_len
del memmap
#create new memmap
start[-1] = 0
start = np.concatenate([start,np.cumsum(data_lengths)])
length = np.concatenate([length,data_lengths])
memmap = np.memmap(filenames[-1], dtype=data.dtype, mode='w+', shape=data.shape)
else:
#append to existing memmap
start = np.concatenate([start,start[-1]+np.cumsum(data_lengths)])
length = np.concatenate([length,data_lengths])
shape[-1] = [start[-1]]+[l for l in data.shape[1:]]
memmap = np.memmap(filenames[-1], dtype=data.dtype, mode='r+', shape=tuple(shape[-1]))
#store mapping memmap_id to memmap_file_id
file_idx=np.concatenate([file_idx,[(len(filenames)-1)]*len(data_lengths)])
#insert the actual data
memmap[start[-len(data_lengths)-1]:start[-len(data_lengths)-1]+len(data)]=data[:]
memmap.flush()
dtype = data.dtype
data = []#reset data storage
data_lengths = []
start= start[:-1]#remove the last element
#cleanup
for npy in npys:
if(delete_npys is True):
npy.unlink()
del memmap
#convert everything to relative paths
filenames= [f.name for f in filenames]
#save metadata
np.savez(target_filename.parent/(target_filename.stem+"_meta.npz"),start=start,length=length,shape=shape,file_idx=file_idx,dtype=dtype,filenames=filenames)
def npys_to_memmap(npys, target_filename, max_len=0, delete_npys=True):
memmap = None
start = []#start_idx in current memmap file
length = []#length of segment
filenames= []#memmap files
file_idx=[]#corresponding memmap file for sample
shape=[]
for idx,npy in tqdm(list(enumerate(npys))):
data = np.load(npy, allow_pickle=True)
if(memmap is None or (max_len>0 and start[-1]+length[-1]>max_len)):
if(max_len>0):
filenames.append(target_filename.parent/(target_filename.stem+"_"+str(len(filenames)+".npy")))
else:
filenames.append(target_filename)
if(memmap is not None):#an existing memmap exceeded max_len
shape.append([start[-1]+length[-1]]+[l for l in data.shape[1:]])
del memmap
#create new memmap
start.append(0)
length.append(data.shape[0])
memmap = np.memmap(filenames[-1], dtype=data.dtype, mode='w+', shape=data.shape)
else:
#append to existing memmap
start.append(start[-1]+length[-1])
length.append(data.shape[0])
memmap = np.memmap(filenames[-1], dtype=data.dtype, mode='r+', shape=tuple([start[-1]+length[-1]]+[l for l in data.shape[1:]]))
#store mapping memmap_id to memmap_file_id
file_idx.append(len(filenames)-1)
#insert the actual data
memmap[start[-1]:start[-1]+length[-1]]=data[:]
memmap.flush()
if(delete_npys is True):
npy.unlink()
del memmap
#append final shape if necessary
if(len(shape)<len(filenames)):
shape.append([start[-1]+length[-1]]+[l for l in data.shape[1:]])
#convert everything to relative paths
filenames= [f.name for f in filenames]
#save metadata
np.savez(target_filename.parent/(target_filename.stem+"_meta.npz"),start=start,length=length,shape=shape,file_idx=file_idx,dtype=data.dtype,filenames=filenames)
def reformat_as_memmap(df, target_filename, data_folder=None, annotation=False, max_len=0, delete_npys=True,col_data="data",col_label="label", batch_length=0):
npys_data = []
npys_label = []
for id,row in df.iterrows():
npys_data.append(data_folder/row[col_data] if data_folder is not None else row[col_data])
if(annotation):
npys_label.append(data_folder/row[col_label] if data_folder is not None else row[col_label])
if(batch_length==0):
npys_to_memmap(npys_data, target_filename, max_len=max_len, delete_npys=delete_npys)
else:
npys_to_memmap_batched(npys_data, target_filename, max_len=max_len, delete_npys=delete_npys,batch_length=batch_length)
if(annotation):
if(batch_length==0):
npys_to_memmap(npys_label, target_filename.parent/(target_filename.stem+"_label.npy"), max_len=max_len, delete_npys=delete_npys)
else:
npys_to_memmap_batched(npys_label, target_filename.parent/(target_filename.stem+"_label.npy"), max_len=max_len, delete_npys=delete_npys, batch_length=batch_length)
#replace data(filename) by integer
df_mapped = df.copy()
df_mapped["data_original"]=df_mapped.data
df_mapped["data"]=np.arange(len(df_mapped))
df_mapped.to_pickle(target_filename.parent/("df_"+target_filename.stem+".pkl"))
return df_mapped
# Cell
class TimeseriesDatasetCrops(torch.utils.data.Dataset):
"""timeseries dataset with partial crops."""
def __init__(self, df, output_size, chunk_length, min_chunk_length, memmap_filename=None, npy_data=None, random_crop=True, data_folder=None, num_classes=2, copies=0, col_lbl="label", stride=None, start_idx=0, annotation=False, transforms=None, sample_items_per_record=1):
"""
accepts three kinds of input:
1) filenames pointing to aligned numpy arrays [timesteps,channels,...] for data and either integer labels or filename pointing to numpy arrays[timesteps,...] e.g. for annotations
2) memmap_filename to memmap file (same argument that was passed to reformat_as_memmap) for data [concatenated,...] and labels- label column in df corresponds to index in this memmap
3) npy_data [samples,ts,...] (either path or np.array directly- also supporting variable length input) - label column in df corresponds to sampleid
transforms: list of callables (transformations) or (preferred) single instance e.g. from torchvision.transforms.Compose (applied in the specified order i.e. leftmost element first)
col_lbl = None: return dummy label 0 (e.g. for unsupervised pretraining)
"""
assert not((memmap_filename is not None) and (npy_data is not None))
# require integer entries if using memmap or npy
assert (memmap_filename is None and npy_data is None) or df.data.dtype==np.int64
self.timeseries_df_data = np.array(df["data"])
if(self.timeseries_df_data.dtype not in [np.int16, np.int32, np.int64]):
assert(memmap_filename is None and npy_data is None) #only for filenames in mode files
self.timeseries_df_data = np.array(df["data"].astype(str)).astype(np.string_)
if(col_lbl is None):# use dummy labels
self.timeseries_df_label = np.zeros(len(df))
else: # use actual labels
if(isinstance(df[col_lbl].iloc[0],list) or isinstance(df[col_lbl].iloc[0],np.ndarray)):#stack arrays/lists for proper batching
self.timeseries_df_label = np.stack(df[col_lbl])
else: # single integers/floats
self.timeseries_df_label = np.array(df[col_lbl])
if(self.timeseries_df_label.dtype not in [np.int16, np.int32, np.int64, np.float32, np.float64]): #everything else cannot be batched anyway mp.Manager().list(self.timeseries_df_label)
assert(annotation and memmap_filename is None and npy_data is None)#only for filenames in mode files
self.timeseries_df_label = np.array(df[col_lbl].apply(lambda x:str(x))).astype(np.string_)
self.output_size = output_size
self.data_folder = data_folder
self.transforms = transforms
if(isinstance(self.transforms,list) or isinstance(self.transforms,np.ndarray)):
print("Warning: the use of list as arguments for transforms is dicouraged")
self.annotation = annotation
self.col_lbl = col_lbl
self.c = num_classes
self.mode="files"
if(memmap_filename is not None):
self.memmap_meta_filename = memmap_filename.parent/(memmap_filename.stem+"_meta.npz")
self.mode="memmap"
memmap_meta = np.load(self.memmap_meta_filename, allow_pickle=True)
self.memmap_start = memmap_meta["start"]
self.memmap_shape = memmap_meta["shape"]
self.memmap_length = memmap_meta["length"]
self.memmap_file_idx = memmap_meta["file_idx"]
self.memmap_dtype = np.dtype(str(memmap_meta["dtype"]))
self.memmap_filenames = np.array(memmap_meta["filenames"]).astype(np.string_)#save as byte to avoid issue with mp
if(annotation):
memmap_meta_label = np.load(self.memmap_meta_filename.parent/("_".join(self.memmap_meta_filename.stem.split("_")[:-1])+"_label_meta.npz"), allow_pickle=True)
self.memmap_shape_label = memmap_meta_label["shape"]
self.memmap_filenames_label = np.array(memmap_meta_label["filenames"]).astype(np.string_)
self.memmap_dtype_label = np.dtype(str(memmap_meta_label["dtype"]))
elif(npy_data is not None):
self.mode="npy"
if(isinstance(npy_data,np.ndarray) or isinstance(npy_data,list)):
self.npy_data = np.array(npy_data)
assert(annotation is False)
else:
self.npy_data = np.load(npy_data, allow_pickle=True)
if(annotation):
self.npy_data_label = np.load(npy_data.parent/(npy_data.stem+"_label.npy"), allow_pickle=True)
self.random_crop = random_crop
self.sample_items_per_record = sample_items_per_record
self.df_idx_mapping=[]
self.start_idx_mapping=[]
self.end_idx_mapping=[]
for df_idx,(id,row) in enumerate(df.iterrows()):
if(self.mode=="files"):
data_length = row["data_length"]
elif(self.mode=="memmap"):
data_length= self.memmap_length[row["data"]]
else: #npy
data_length = len(self.npy_data[row["data"]])
if(chunk_length == 0):#do not split
idx_start = [start_idx]
idx_end = [data_length]
else:
idx_start = list(range(start_idx,data_length,chunk_length if stride is None else stride))
idx_end = [min(l+chunk_length, data_length) for l in idx_start]
#remove final chunk(s) if too short
for i in range(len(idx_start)):
if(idx_end[i]-idx_start[i]< min_chunk_length):
del idx_start[i:]
del idx_end[i:]
break
#append to lists
for _ in range(copies+1):
for i_s,i_e in zip(idx_start,idx_end):
self.df_idx_mapping.append(df_idx)
self.start_idx_mapping.append(i_s)
self.end_idx_mapping.append(i_e)
#convert to np.array to avoid mp issues with python lists
self.df_idx_mapping = np.array(self.df_idx_mapping)
self.start_idx_mapping = np.array(self.start_idx_mapping)
self.end_idx_mapping = np.array(self.end_idx_mapping)
def __len__(self):
return len(self.df_idx_mapping)
@property
def is_empty(self):
return len(self.df_idx_mapping)==0
def __getitem__(self, idx):
lst=[]
for _ in range(self.sample_items_per_record):
#determine crop idxs
timesteps= self.get_sample_length(idx)
if(self.random_crop):#random crop
if(timesteps==self.output_size):
start_idx_rel = 0
else:
start_idx_rel = random.randint(0, timesteps - self.output_size -1)#np.random.randint(0, timesteps - self.output_size)
else:
start_idx_rel = (timesteps - self.output_size)//2
if(self.sample_items_per_record==1):
return self._getitem(idx,start_idx_rel)
else:
lst.append(self._getitem(idx,start_idx_rel))
return tuple(lst)
def _getitem(self, idx,start_idx_rel):
#low-level function that actually fetches the data
df_idx = self.df_idx_mapping[idx]
start_idx = self.start_idx_mapping[idx]
end_idx = self.end_idx_mapping[idx]
#determine crop idxs
timesteps= end_idx - start_idx
assert(timesteps>=self.output_size)
start_idx_crop = start_idx + start_idx_rel
end_idx_crop = start_idx_crop+self.output_size
#print(idx,start_idx,end_idx,start_idx_crop,end_idx_crop)
#load the actual data
if(self.mode=="files"):#from separate files
data_filename = str(self.timeseries_df_data[df_idx],encoding='utf-8') #todo: fix potential issues here
if self.data_folder is not None:
data_filename = self.data_folder/data_filename
data = np.load(data_filename, allow_pickle=True)[start_idx_crop:end_idx_crop] #data type has to be adjusted when saving to npy
ID = data_filename.stem
if(self.annotation is True):
label_filename = str(self.timeseries_df_label[df_idx],encoding='utf-8')
if self.data_folder is not None:
label_filename = self.data_folder/label_filename
label = np.load(label_filename, allow_pickle=True)[start_idx_crop:end_idx_crop] #data type has to be adjusted when saving to npy
else:
label = self.timeseries_df_label[df_idx] #input type has to be adjusted in the dataframe
elif(self.mode=="memmap"): #from one memmap file
memmap_idx = self.timeseries_df_data[df_idx] #grab the actual index (Note the df to create the ds might be a subset of the original df used to create the memmap)
memmap_file_idx = self.memmap_file_idx[memmap_idx]
idx_offset = self.memmap_start[memmap_idx]
#wi = torch.utils.data.get_worker_info()
#pid = 0 if wi is None else wi.id#os.getpid()
#print("idx",idx,"ID",ID,"idx_offset",idx_offset,"start_idx_crop",start_idx_crop,"df_idx", self.df_idx_mapping[idx],"pid",pid)
mem_filename = str(self.memmap_filenames[memmap_file_idx],encoding='utf-8')
mem_file = np.memmap(self.memmap_meta_filename.parent/mem_filename, self.memmap_dtype, mode='r', shape=tuple(self.memmap_shape[memmap_file_idx]))
data = np.copy(mem_file[idx_offset + start_idx_crop: idx_offset + end_idx_crop])
del mem_file
#print(mem_file[idx_offset + start_idx_crop: idx_offset + end_idx_crop])
if(self.annotation):
mem_filename_label = str(self.memmap_filenames_label[memmap_file_idx],encoding='utf-8')
mem_file_label = np.memmap(self.memmap_meta_filename.parent/mem_filename_label, self.memmap_dtype_label, mode='r', shape=tuple(self.memmap_shape_label[memmap_file_idx]))
label = np.copy(mem_file_label[idx_offset + start_idx_crop: idx_offset + end_idx_crop])
del mem_file_label
else:
label = self.timeseries_df_label[df_idx]
else:#single npy array
ID = self.timeseries_df_data[df_idx]
data = self.npy_data[ID][start_idx_crop:end_idx_crop]
if(self.annotation):
label = self.npy_data_label[ID][start_idx_crop:end_idx_crop]
else:
label = self.timeseries_df_label[df_idx]
sample = (data,label)
if(isinstance(self.transforms,list)):#transforms passed as list
for t in self.transforms:
sample = t(sample)
elif(self.transforms is not None):#single transform e.g. from torchvision.transforms.Compose
sample = self.transforms(sample)
return sample
def get_sampling_weights(self, class_weight_dict,length_weighting=False, timeseries_df_group_by_col=None):
'''
class_weight_dict: dictionary of class weights
length_weighting: weigh samples by length
timeseries_df_group_by_col: column of the pandas df used to create the object'''
assert(self.annotation is False)
assert(length_weighting is False or timeseries_df_group_by_col is None)
weights = np.zeros(len(self.df_idx_mapping),dtype=np.float32)
length_per_class = {}
length_per_group = {}
for iw,(i,s,e) in enumerate(zip(self.df_idx_mapping,self.start_idx_mapping,self.end_idx_mapping)):
label = self.timeseries_df_label[i]
weight = class_weight_dict[label]
if(length_weighting):
if label in length_per_class.keys():
length_per_class[label] += e-s
else:
length_per_class[label] = e-s
if(timeseries_df_group_by_col is not None):
group = timeseries_df_group_by_col[i]
if group in length_per_group.keys():
length_per_group[group] += e-s
else:
length_per_group[group] = e-s
weights[iw] = weight
if(length_weighting):#need second pass to properly take into account the total length per class
for iw,(i,s,e) in enumerate(zip(self.df_idx_mapping,self.start_idx_mapping,self.end_idx_mapping)):
label = self.timeseries_df_label[i]
weights[iw]= (e-s)/length_per_class[label]*weights[iw]
if(timeseries_df_group_by_col is not None):
for iw,(i,s,e) in enumerate(zip(self.df_idx_mapping,self.start_idx_mapping,self.end_idx_mapping)):
group = timeseries_df_group_by_col[i]
weights[iw]= (e-s)/length_per_group[group]*weights[iw]
weights = weights/np.min(weights)#normalize smallest weight to 1
return weights
def get_id_mapping(self):
return self.df_idx_mapping
def get_sample_id(self,idx):
return self.df_idx_mapping[idx]
def get_sample_length(self,idx):
return self.end_idx_mapping[idx]-self.start_idx_mapping[idx]
def get_sample_start(self,idx):
return self.start_idx_mapping[idx]
# Cell
class RandomCrop(object):
"""Crop randomly the image in a sample.
"""
def __init__(self, output_size,annotation=False):
self.output_size = output_size
self.annotation = annotation
def __call__(self, sample):
data, label = sample
timesteps= len(data)
assert(timesteps>=self.output_size)
if(timesteps==self.output_size):
start=0
else:
start = random.randint(0, timesteps - self.output_size-1) #np.random.randint(0, timesteps - self.output_size)
data = data[start: start + self.output_size]
if(self.annotation):
label = label[start: start + self.output_size]
return data, label
# Cell
class CenterCrop(object):
"""Center crop the image in a sample.
"""
def __init__(self, output_size, annotation=False):
self.output_size = output_size
self.annotation = annotation
def __call__(self, sample):
data, label = sample
timesteps= len(data)
start = (timesteps - self.output_size)//2
data = data[start: start + self.output_size]
if(self.annotation):
label = label[start: start + self.output_size]
return data, label
# Cell
class GaussianNoise(object):
"""Add gaussian noise to sample.
"""
def __init__(self, scale=0.1):
self.scale = scale
def __call__(self, sample):
if self.scale ==0:
return sample
else:
data, label = sample
data = data + np.reshape(np.array([random.gauss(0,self.scale) for _ in range(np.prod(data.shape))]),data.shape)#np.random.normal(scale=self.scale,size=data.shape).astype(np.float32)
return data, label
# Cell
class Rescale(object):
"""Rescale by factor.
"""
def __init__(self, scale=0.5,interpolation_order=3):
self.scale = scale
self.interpolation_order = interpolation_order
def __call__(self, sample):
if self.scale ==1:
return sample
else:
data, label = sample
timesteps_new = int(self.scale * len(data))
data = transform.resize(data,(timesteps_new,data.shape[1]),order=interpolation_order).astype(np.float32)
return data,label
# Cell
class ToTensor(object):
"""Convert ndarrays in sample to Tensors."""
def __init__(self, transpose_data=True, transpose_label=False):
#swap channel and time axis for direct application of pytorch's convs
self.transpose_data=transpose_data
self.transpose_label=transpose_label
def __call__(self, sample):
def _to_tensor(data,transpose=False):
if(isinstance(data,np.ndarray)):
if(transpose):#seq,[x,y,]ch
return torch.from_numpy(np.moveaxis(data,-1,0))
else:
return torch.from_numpy(data)
else:#default_collate will take care of it
return data
data, label = sample
if not isinstance(data,tuple):
data = _to_tensor(data,self.transpose_data)
else:
data = tuple(_to_tensor(x,self.transpose_data) for x in data)
if not isinstance(label,tuple):
label = _to_tensor(label,self.transpose_label)
else:
label = tuple(_to_tensor(x,self.transpose_label) for x in label)
return data,label #returning as a tuple (potentially of lists)
# Cell
class Normalize(object):
"""Normalize using given stats.
"""
def __init__(self, stats_mean, stats_std, input=True, channels=[]):
self.stats_mean=stats_mean.astype(np.float32) if stats_mean is not None else None
self.stats_std=stats_std.astype(np.float32)+1e-8 if stats_std is not None else None
self.input = input
if(len(channels)>0):
for i in range(len(stats_mean)):
if(not(i in channels)):
self.stats_mean[:,i]=0
self.stats_std[:,i]=1
def __call__(self, sample):
datax, labelx = sample
data = datax if self.input else labelx
#assuming channel last
if(self.stats_mean is not None):
data = data - self.stats_mean
if(self.stats_std is not None):
data = data/self.stats_std
if(self.input):
return (data, labelx)
else:
return (datax, data)
# Cell
class NormalizeBatch(object):
"""Normalize using batch statistics.
axis: tuple of integers of axis numbers to be normalized over (by default everything but the last)
"""
def __init__(self, input=True, channels=[],axis=None):
self.channels = channels
self.channels_keep = None
self.input = input
self.axis = axis
def __call__(self, sample):
datax, labelx = sample
data = datax if self.input else labelx
#assuming channel last
#batch_mean = np.mean(data,axis=tuple(range(0,len(data)-1)))
#batch_std = np.std(data,axis=tuple(range(0,len(data)-1)))+1e-8
batch_mean = np.mean(data,axis=self.axis if self.axis is not None else tuple(range(0,len(data.shape)-1)))
batch_std = np.std(data,axis=self.axis if self.axis is not None else tuple(range(0,len(data.shape)-1)))+1e-8
if(len(self.channels)>0):
if(self.channels_keep is None):
self.channels_keep = np.setdiff(range(data.shape[-1]),self.channels)
batch_mean[self.channels_keep]=0
batch_std[self.channels_keep]=1
data = (data - batch_mean)/batch_std
if(self.input):
return (data, labelx)
else:
return (datax, data)
# Cell
class ButterFilter(object):
"""Apply filter
"""
def __init__(self, lowcut=50, highcut=50, fs=100, order=5, btype='band', forwardbackward=True, input=True):
self.filter = butter_filter(lowcut,highcut,fs,order,btype)
self.input = input
self.forwardbackward = forwardbackward
def __call__(self, sample):
datax, labelx = sample
data = datax if self.input else labelx
if(self.forwardbackward):
data = sosfiltfilt(self.filter, data, axis=0)
else:
data = sosfilt(self.filter, data, axis=0)
if(self.input):
return (data, labelx)
else:
return (datax, data)
# Cell
class ChannelFilter(object):
"""Select certain channels.
"""
def __init__(self, channels=[0], input=True):
self.channels = channels
self.input = input
def __call__(self, sample):
data,label = sample
if(self.input):
return (data[...,self.channels], label)
else:
return (data, label[...,self.channels])
# Cell
class Transform(object):
"""Transforms data using a given function i.e. data_new = func(data) for input is True else label_new = func(label)
"""
def __init__(self, func, input=False):
self.func = func
self.input = input
def __call__(self, sample):
data,label = sample
if(self.input):
return (self.func(data), label)
else:
return (data, self.func(label))
# Cell
class TupleTransform(object):
"""Transforms data using a given function (operating on both data and label and return a tuple) i.e. data_new, label_new = func(data_old, label_old)
"""
def __init__(self, func, input=False):
self.func = func
def __call__(self, sample):
data,label = sample
return self.func(data,label)
# Cell
def aggregate_predictions(preds,targs=None,idmap=None,aggregate_fn = np.mean,verbose=False):
'''
aggregates potentially multiple predictions per sample (can also pass targs for convenience)
idmap: idmap as returned by TimeSeriesCropsDataset's get_id_mapping
preds: ordered predictions as returned by learn.get_preds()
aggregate_fn: function that is used to aggregate multiple predictions per sample (most commonly np.amax or np.mean)
'''
if(idmap is not None and len(idmap)!=len(np.unique(idmap))):
if(verbose):
print("aggregating predictions...")
preds_aggregated = []
targs_aggregated = []
for i in np.unique(idmap):
preds_local = preds[np.where(idmap==i)[0]]
preds_aggregated.append(aggregate_fn(preds_local,axis=0))
if targs is not None:
targs_local = targs[np.where(idmap==i)[0]]
assert(np.all(targs_local==targs_local[0])) #all labels have to agree
targs_aggregated.append(targs_local[0])
if(targs is None):
return np.array(preds_aggregated)
else:
return np.array(preds_aggregated),np.array(targs_aggregated)
else:
if(targs is None):
return preds
else:
return preds,targs