-
Notifications
You must be signed in to change notification settings - Fork 3
/
data_input.py
772 lines (594 loc) · 26.5 KB
/
data_input.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
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
# Copyright 2017 The Chiron Authors. All Rights Reserved.
#
#This Source Code Form is subject to the terms of the Mozilla Public
#License, v. 2.0. If a copy of the MPL was not distributed with this
#file, You can obtain one at http://mozilla.org/MPL/2.0/.
#
"""
This file is a revision of Chiron(V0.3)'s chiron_input.py file.
(https://github.com/haotianteng/Chiron)
We make the following changes from the original one:
1. different padding approach for generating fixed-length signal vector.
2. revision of read_raw() that returns varibles including segment length, gold labels and signal vector of fixed length.
"""
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function
import collections, os, sys, tempfile
import h5py
import numpy as np
from statsmodels import robust
from six.moves import range, zip
import tensorflow as tf
from chiron.utils import progress
#from chiron.chiron_input import read_raw_data_sets
from utility import *
import matplotlib
matplotlib.use('Agg') # this need for the linux env
import matplotlib.pyplot as plt
##################Previous implementation ################
raw_labels = collections.namedtuple('raw_labels', ['start', 'length', 'base'])
class Flags(object):
def __init__(self):
self.max_segments_number = None
self.MAXLEN = 1e9 # original run 1e4 Maximum Length of the holder in biglist. 1e5 by default
FLAGS = Flags()
class biglist(object):
"""
biglist class, read into memory if reads number < MAXLEN, otherwise read into a hdf5 file.
"""
def __init__(self,
data_handle,
dtype='float32',
length=0,
cache=False,
max_len=1e5):
self.handle = data_handle
self.dtype = dtype
self.holder = list()
self.length = length
self.max_len = max_len
self.cache = cache # Mark if the list has been saved into hdf5 or not
@property
def shape(self):
return self.handle.shape
def append(self, item):
self.holder.append(item)
self.check_save()
def __add__(self, add_list):
self.holder += add_list
self.check_save()
return self
def __len__(self):
return self.length + len(self.holder)
def resize(self, size, axis=0):
self.save_rest()
if self.cache:
self.handle.resize(size, axis=axis)
self.length = len(self.handle)
else:
self.holder = self.holder[:size]
def save_rest(self):
if self.cache:
if len(self.holder) != 0:
self.save()
def check_save(self):
if len(self.holder) > self.max_len:
self.save()
self.cache = True
def save(self):
if type(self.holder[0]) is list:
max_sub_len = max([len(sub_a) for sub_a in self.holder])
shape = self.handle.shape
for item in self.holder:
item.extend([0] * (max(shape[1], max_sub_len) - len(item)))
if max_sub_len > shape[1]:
self.handle.resize(max_sub_len, axis=1)
self.handle.resize(self.length + len(self.holder), axis=0)
self.handle[self.length:] = self.holder
self.length += len(self.holder)
del self.holder[:]
self.holder = list()
else:
self.handle.resize(self.length + len(self.holder), axis=0)
self.handle[self.length:] = self.holder
self.length += len(self.holder)
del self.holder[:]
self.holder = list()
def __getitem__(self, val):
if self.cache:
if len(self.holder) != 0:
self.save()
return self.handle[val]
else:
return self.holder[val]
class DataSet(object):
def __init__(self, event, event_length, label, label_length, label_vec, label_segs, for_eval=False,):
"""Custruct a DataSet."""
if for_eval == False:
assert len(event) == len(label) and len(event_length) == len(label_length) and len(event) == len(event_length), "Sequence length for event \
and label does not of event and label should be same, \
event:%d , label:%d" % (len(event), len(label))
self._event = event
self._event_length = event_length
self._label = label
self._label_vec = label_vec # new added
self._label_segs = label_segs # new added
self._label_length = label_length
self._reads_n = len(event)
self._epochs_completed = 0
self._index_in_epoch = 0
self._for_eval = for_eval
self._perm = np.arange(self._reads_n)
@property
def event(self):
return self._event
@property
def label(self):
return self._label
@property
def label_vec(self):
return self._label_vec
@property
def label_segs(self):
return self._label_segs
@property
def event_length(self):
return self._event_length
@property
def label_length(self):
return self._label_length
@property
def reads_n(self):
return self._reads_n
@property
def index_in_epoch(self):
return self._index_in_epoch
@property
def epochs_completed(self):
return self._epochs_completed
@property
def for_eval(self):
return self._for_eval
@property
def perm(self):
return self._perm
def read_into_memory(self, index):
event = np.asarray(list(zip([self._event[i] for i in index],
[self._event_length[i] for i in index])))
if not self.for_eval:
label = np.asarray(list(zip([self._label[i] for i in index],
[self._label_length[i] for i in index])))
label_vec = np.asarray(list(zip([self._label_vec[i] for i in index],
[self._label_segs[i] for i in index])))
else:
label = []
label_vec = []
return event, label, label_vec
def next_batch(self, batch_size, shuffle=True, sig_norm=False):
"""Return next batch in batch_size from the data set.
Input Args:
batch_size:A scalar indicate the batch size.
shuffle: boolean, indicate if the data should be shuffled after each epoch.
sig_norm: If the signal need to be normalized, if sig_norm set
to True when read the data, then the redundant sig_norm is not required.
Output Args:
inputX,sequence_length,label_batch: tuple of (indx,vals,shape)
"""
if self.epochs_completed>=1 and self.for_eval:
print("Warning, evaluation dataset already finish one iteration.")
start = self._index_in_epoch
# Shuffle for the first epoch
if self._epochs_completed == 0 and start == 0:
if shuffle:
np.random.shuffle(self._perm)
# Go to the next epoch
if start + batch_size > self.reads_n:
# Finished epoch
self._epochs_completed += 1
# Get the rest samples in this epoch
rest_reads_n = self.reads_n - start
event_rest_part, label_rest_part, label_vec_rest_part = self.read_into_memory(self._perm[start:self._reads_n])
start = 0
if self._for_eval:
event_batch = event_rest_part
label_batch = label_rest_part
label_vec_batch = label_vec_rest_part
self._index_in_epoch = 0
end = 0
# Shuffle the data
else:
if shuffle:
np.random.shuffle(self._perm)
# Start next epoch
self._index_in_epoch = batch_size - rest_reads_n
end = self._index_in_epoch
event_new_part, label_new_part, label_vec_new_part = self.read_into_memory(self._perm[start:end])
if event_rest_part.shape[0] == 0:
event_batch = event_new_part
label_batch = label_new_part
label_vec_batch = label_vec_new_part
elif event_new_part.shape[0] == 0:
event_batch = event_rest_part
label_batch = label_rest_part
label_vec_batch = label_vec_rest_part
else:
event_batch = np.concatenate((event_rest_part, event_new_part), axis=0)
label_batch = np.concatenate((label_rest_part, label_new_part), axis=0)
label_vec_batch = np.concatenate((label_vec_rest_part, label_vec_new_part), axis=0)
else:
self._index_in_epoch += batch_size
end = self._index_in_epoch
event_batch, label_batch, label_vec_batch = self.read_into_memory(self._perm[start:end])
# errors happens here
if(len(label_vec_batch) > 0):
label_segs = [ x for x in label_vec_batch[:,1] ]
label_raw = [ x for x in label_batch[:,0]]
else:
label_segs, label_raw = None, None
if not self._for_eval:
label_batch = batch2sparse(label_batch)
seq_length = event_batch[:, 1].astype(np.int32)
# conditional added
if(len(label_vec_batch) > 0):
return np.vstack(event_batch[:, 0]).astype(np.float32), seq_length, label_batch, \
np.vstack(label_vec_batch[:,0]).astype(np.int32), label_segs, label_raw
else:
return np.vstack(event_batch[:, 0]).astype(np.float32), seq_length, label_batch, \
None, None, None
def read_data_for_eval(file_path,
start_index=0,
step=20,
seg_length=300,
sig_norm="median",
reverse = False):
"""
Input Args:
file_path: file path to a signal file.
start_index: the index of the signal start to read.
step: sliding step size.
seg_length: length of segments.
sig_norm: if the signal need to be normalized.
reverse: if the signal need to be reversed.
"""
if not file_path.endswith('.signal'):
raise ValueError('A .signal file is required.')
else:
event = list()
event_len = list()
label = list()
label_len = list()
f_signal = read_signal(file_path)
if reverse:
f_signal = f_signal[::-1]
f_signal = f_signal[start_index:]
sig_len = len(f_signal)
for indx in range(0, sig_len, step):
segment_sig = f_signal[indx:indx + seg_length]
segment_len = len(segment_sig)
padding(segment_sig, seg_length)
event.append(segment_sig)
event_len.append(segment_len)
#evaluation = DataSet(event=event, event_length=event_len, label=label, label_length=label_len, for_eval=True)
## take care about the label_vec, label_segs
evaluation = DataSet(event=event, event_length=event_len, label=label, label_length=label_len, label_vec = list(), label_segs= list(), for_eval=True)
return evaluation
def read_cache_dataset(h5py_file_path):
"""Notice: Return a data reader for a h5py_file, call this function multiple
time for parallel reading, this will give you N dependent dataset reader,
each reader read independently from the h5py file."""
hdf5_record = h5py.File(h5py_file_path, "r")
event_h = hdf5_record['event/record']
event_length_h = hdf5_record['event/length']
label_h = hdf5_record['label/record']
label_vec_h = hdf5_record['label/record_vec']
label_seg_h = hdf5_record['label/record_seg']
label_length_h = hdf5_record['label/length']
event_len = len(event_h)
label_len = len(label_h)
assert len(event_h) == len(event_length_h)
assert len(label_h) == len(label_length_h)
event = biglist(data_handle=event_h, length=event_len, cache=True)
event_length = biglist(data_handle=event_length_h, length=event_len,
cache=True)
label = biglist(data_handle=label_h, length=label_len, cache=True)
label_length = biglist(data_handle=label_length_h, length=label_len,
cache=True)
label_vec = biglist(data_handle=label_vec_h, length=event_len, cache=True)
label_seg = biglist(data_handle=label_seg_h, length=label_len, cache=True)
return DataSet(event=event, event_length=event_length, label=label,
label_length=label_length, label_vec = label_vec, label_segs= label_seg)
## Read from raw data
def read_raw_data_sets(data_dir, h5py_file_path=None, seq_length=300, k_mer=1, max_segments_num=FLAGS.max_segments_number):
# make temp record
if h5py_file_path is None:
h5py_file_path = tempfile.mkdtemp() + '/temp_record.hdf5'
else:
try:
os.remove(os.path.abspath(h5py_file_path))
except:
pass
if not os.path.isdir(os.path.dirname(os.path.abspath(h5py_file_path))):
os.mkdir(os.path.dirname(os.path.abspath(h5py_file_path)))
with h5py.File(h5py_file_path, "a") as hdf5_record:
event_h = hdf5_record.create_dataset('event/record', dtype='float32', shape=(0, seq_length), maxshape=(None, seq_length))
event_length_h = hdf5_record.create_dataset('event/length', dtype='int32', shape=(0,), maxshape=(None,), chunks=True)
label_h = hdf5_record.create_dataset('label/record', dtype='int32',shape=(0, 0), maxshape=(None, seq_length))
label_length_h = hdf5_record.create_dataset('label/length',dtype='int32', shape=(0,), maxshape=(None,))
label_vec_h = hdf5_record.create_dataset('label/record_vec', dtype='int32',shape=(0, seq_length), maxshape=(None, seq_length))
label_segs_h = hdf5_record.create_dataset('label/record_seg', dtype='int32',shape=(0, 0), maxshape=(None, seq_length))
event = biglist(data_handle=event_h, max_len=FLAGS.MAXLEN)
event_length = biglist(data_handle=event_length_h, max_len=FLAGS.MAXLEN)
label = biglist(data_handle=label_h, max_len=FLAGS.MAXLEN)
label_length = biglist(data_handle=label_length_h, max_len=FLAGS.MAXLEN)
label_vec = biglist(data_handle=label_vec_h, max_len=FLAGS.MAXLEN)
label_segs = biglist(data_handle=label_segs_h, max_len=FLAGS.MAXLEN)
count = 0
file_count = 0
for name in os.listdir(data_dir):
if name.endswith(".signal"):
file_pre = os.path.splitext(name)[0]
f_signal = read_signal(data_dir + name)
if len(f_signal) == 0:
continue
try:
f_label = read_label(data_dir + file_pre + '.label',
skip_start=10,
window_n=int((k_mer - 1) / 2))
except:
sys.stdout.write("Read the label %s fail.Skipped." % (name))
continue
# read_raw singals, it will be segmented into (event, label)
tmp_event, tmp_event_length, tmp_label, tmp_label_length, tmp_label_vec, tmp_label_segs = read_raw(f_signal, f_label, seq_length)
event += tmp_event
event_length += tmp_event_length
label += tmp_label
label_length += tmp_label_length
label_vec += tmp_label_vec
label_segs += tmp_label_segs
del tmp_event
del tmp_event_length
del tmp_label
del tmp_label_length
del tmp_label_vec
del tmp_label_segs
count = len(event)
if file_count % 10 == 0:
if max_segments_num is not None:
sys.stdout.write("%d/%d events read. \n" % (count, max_segments_num))
if len(event) > max_segments_num:
event.resize(max_segments_num)
label.resize(max_segments_num)
label_vec.resize(max_segments_num)
label_segs.resize(max_segments_num)
event_length.resize(max_segments_num)
label_length.resize(max_segments_num)
break
else:
sys.stdout.write("%d lines read. \n" % (count))
file_count += 1
if event.cache:
train = read_cache_dataset(h5py_file_path)
else:
train = DataSet(event=event, event_length=event_length, label=label,
label_length=label_length, label_vec=label_vec, label_segs= label_segs)
return train
#data normalization applied here
def read_signal(file_path, normalize="median"):
f_h = open(file_path, 'r')
signal = list()
for line in f_h:
signal += [float(x) for x in line.split()]
signal = np.asarray(signal)
if len(signal) == 0:
return signal.tolist()
if normalize == "mean":
signal = (signal - np.mean(signal)) / np.float(np.std(signal))
elif normalize == "median":
signal = (signal - np.median(signal)) / np.float(robust.mad(signal))
return signal.tolist()
def read_label(file_path, skip_start=10, window_n=0):
f_h = open(file_path, 'r')
start = list()
length = list()
base = list()
all_base = list()
if skip_start < window_n:
skip_start = window_n
for line in f_h:
record = line.split()
all_base.append(base2ind(record[2]))
f_h.seek(0, 0) # Back to the start
file_len = len(all_base)
for count, line in enumerate(f_h):
record = line.split()
if count < skip_start or count > (file_len - skip_start - 1):
continue
start.append(int(record[0]))
length.append(int(record[1]) - int(record[0]))
k_mer = 0
for i in range(window_n * 2 + 1):
k_mer = k_mer * 4 + all_base[count + i - window_n]
base.append(k_mer)
return raw_labels(start=start, length=length, base=base)
# joint loading, can revise here
def read_raw(raw_signal, raw_label, max_seq_length):
label_val = list()
label_length = list()
label_val_vec = list()
label_segs = list()
event_val = list()
event_length = list()
current_length = 0
current_label = []
current_event = []
current_label_vec = []
current_segs = []
for indx, segment_length in enumerate(raw_label.length):
current_start = raw_label.start[indx]
current_base = raw_label.base[indx]
if current_length + segment_length < max_seq_length:
current_event += raw_signal[current_start:current_start + segment_length]
current_label.append(current_base)
current_label_vec += [current_base] * segment_length
current_length += segment_length
current_segs += [segment_length]
else:
if current_length > (max_seq_length / 2) and len(current_label) > 5:
current_event += raw_signal[current_start: current_start + (max_seq_length - current_length)]
# padding(current_event, max_seq_length, raw_signal[current_start : current_start + max_seq_length])
padding(current_label_vec, max_seq_length, [ current_base ] * (max_seq_length - current_length))
# 20190515 add the incomplete information
current_label.append(current_base)
current_segs += [ max_seq_length - current_length ]
# update the length last
current_length += max_seq_length - current_length
event_val.append(current_event)
event_length.append(current_length)
# add to the list
label_val.append(current_label)
label_length.append(len(current_label))
label_val_vec.append(current_label_vec)
label_segs.append(current_segs)
# Begin a new event-label, resetting
current_event = raw_signal[current_start:current_start + segment_length]
current_length = segment_length
current_label = [current_base]
current_label_vec = [current_base] * segment_length
current_segs = [segment_length]
if segment_length == 0:
print("----val-%s set==0---" %(current_label))
#print(len(event_val))
#print("-"*30)
#print(label_length)
#print("="*30)
return event_val, event_length, label_val, label_length, label_val_vec, label_segs
# padding the rest of part with vector
def padding(x, L, padding_list=None):
"""Padding the vector x to length L"""
len_x = len(x)
assert len_x <= L, "Length of vector x is larger than the padding length"
zero_n = L - len_x
if padding_list is None:
x.extend([0] * zero_n)
elif len(padding_list) < zero_n:
x.extend(padding_list + [0] * (zero_n - len(padding_list)))
else:
x.extend(padding_list[0:zero_n])
return None
def batch2sparse(label_batch):
"""Transfer a batch of label to a sparse tensor
"""
values = []
indices = []
for batch_i, label_list in enumerate(label_batch[:, 0]):
for indx, label in enumerate(label_list):
if indx >= label_batch[batch_i, 1]:
break
indices.append([batch_i, indx])
values.append(label)
shape = [len(label_batch), max(label_batch[:, 1])]
return indices, values, shape
################################################
# loading data
################################################
def loading_data(fileConfigSet, cacheFile, len_encoding=True):
if os.path.exists(cacheFile):
with h5py.File(cacheFile, "r") as hf:
X = hf["X_data"][:]
seq_len = hf["seq_len"][:]
label = [hf["Y_ctc/index"][:], hf["Y_ctc/value"][:], hf["Y_ctc/shape"]]
label_vec = hf["Y_vec"][:]
label_seg = [ hf["Y_seg/"+str(i)][:] for i in range(len(X)) ]
# checking the h5py loading process
label_raw = [ hf["label_raw/"+str(i)][:] for i in range(len(X))]
else:
print("Now caching the data ... ")
ds = read_raw_data_sets(fileConfigSet[0], fileConfigSet[1], 300, 1)
# call the chiron function for loading the data
X, seq_len, label, label_vec, label_seg, label_raw = ds.next_batch(ds._reads_n)
with h5py.File(cacheFile, "w") as hf:
hf.create_dataset("X_data", data=X)
hf.create_dataset("seq_len", data=seq_len)
# this used for sparse matrix
hf.create_dataset("Y_vec", data=label_vec)
hf.create_dataset("Y_ctc/index", data=label[0])
hf.create_dataset("Y_ctc/value", data=label[1])
hf.create_dataset("Y_ctc/shape", data=label[2])
for i in range(len(label_raw)):
hf.create_dataset("Y_seg/"+str(i), data=label_seg[i])
# be careful about the label-0 conflict issues
hf.create_dataset("label_raw/"+str(i), data=np.array(label_raw[i], dtype=int))
print("[%d] segments Data loading Done!" %(X.shape[0]))
if len_encoding == False:
return X, seq_len, label, label_vec, label_seg, label_raw
########## 2019/06/15 added #########
# print("@ (*) 8-label (1,1) split transformation AAA to A1A2A1 ... ")
label_vec_new = []
for i in range(len(label_raw)):
gold_raw = validRawLabel(label_raw[i]) # remove the end-tails of 0s
# transform to new labels
newLabels = label2D_Transform_11(gold_raw)
newLabels_noRecovery = newLabels #label2D_Transform_norecovery_11(gold_raw)
assert(len(gold_raw) == len(newLabels_noRecovery))
newLabelVec = []
# generate labels for the signals
for j in range(len(gold_raw)):
newLabelVec.extend([ newLabels_noRecovery[j] ] * label_seg[i][j])
# padding,take care about this padding part!!!, why the length is different
assert(len(newLabelVec) == len(label_vec[i]))
label_vec_new.append(newLabelVec)
label_new = np.array(label_vec_new)
return X, seq_len, label, label_vec, label_seg, label_raw, label_new
def test():
### Input Test ###
Data_dir = '/data/workspace/nanopore/data/chiron_data/train/'
#train = read_tfrecord(Data_dir,"train.tfrecords",seq_length=1000)
eval = read_raw_data_sets(Data_dir)
for i in range(5):
inputX, sequence_length, label, label_vec, label_seg = eval.next_batch(1)
bks = np.cumsum(label_seg[0])
print(inputX.shape)
print(label[2])
print(label_vec.shape)
print("-"*10)
print(label_seg[0])
bks = np.cumsum([0] + label_seg[0])
print(bks)
"""
print(len(label_seg[0]))
print(label_vec)
print(label[1])
print(len(label[1]))
print(sequence_length)
"""
doPlot=True
if doPlot:
fig=plt.figure(figsize=(50,10))
plt.plot(np.array(range(inputX.shape[1]), dtype=np.int16), \
inputX[0,:])
plt.xticks(bks[:-1], ind2base(label[1]), color="brown",fontsize=26)
for bk in bks:
plt.axvline(bk, linestyle="-.", color="red")
plt.savefig("/dropbox/currency"+str(i)+".png")
plt.close("all")
# filtering out the outliner signals in the training.
def signalFiltering(X, threshold=10):
Idx = []
count_filtered = 0
for i in range(X.shape[0]):
tag = True
for j in range(X.shape[1]):
if np.abs(X[i,j]) > threshold:
#print(X[i,j])
count_filtered += 1
tag = False
if tag == True:
Idx.append(i)
print("@@ Total filtering [%d] samples that exceed max Currency strength of %d" %(count_filtered, threshold))
return Idx
if __name__ == '__main__':
#test()
for i in range(13):
print("- Decompisiotn of [%d]" %i)
print(number_decomposition(i))