-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathdeodel2.py
1139 lines (993 loc) · 43 KB
/
deodel2.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
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
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
"""Deodata Delanga Classification"""
# c4pub@git 2023 - 2024
#
# Latest version available at: https://github.com/c4pub/deodel
#
import warnings
import statistics
import pandas as pd
import numpy as np
import math
class DeodelSecond :
"""Predictor implementing a variant of the Deodata Delanga classifier.
Read more in the reference document.
Parameters
----------
aux_param : dict, default=None
Auxiliary configuration parameters.
Configuration dictionary keywords:
'split_no' : int, default=3
Number of categories, or bins, to split numerical attributes.
'tbreak_depth' : int, default=0
Limit to the depth of searching for tie breakers. O means no limit.
Attributes
----------
version : float
Version of algorithm implementation
Notes
-----
See online documentation for a discussion of the method.
https://doi.org/10.13140/RG.2.2.33413.06880
Examples
--------
>>> X = [[0], [1], [2], [3]]
>>> y = [0, 0, 1, 1]
>>> from deodel2 import DeodelSecond
>>> deodel2 = DeodelSecond()
>>> deodel2.fit(X, y)
DeodelSecond(...)
>>> print(deodel2.predict([[1.1]]))
[0]
"""
def __init__(
self,
aux_param = None
):
default_aux = {
'split_no': 0,
'split_mode': 'eq_width',
'tbreak_depth': None,
'predict_mode': 'auto',
'score_factor': 2,
'legacy_match': False,
'int_is_num': True,
}
if aux_param == None :
self.aux_param = {}
self.cfg_param = default_aux
else :
self.aux_param = aux_param
cfg_param = {}
for crt_elem in default_aux :
if crt_elem in aux_param :
cfg_param[crt_elem] = aux_param[crt_elem]
else :
cfg_param[crt_elem] = default_aux[crt_elem]
self.cfg_param = cfg_param
version = 2.17
def __repr__(self):
'''Returns representation of the object'''
return("{}({!r})".format(self.__class__.__name__, self.aux_param))
def fit(self, X, y):
"""Fit the classifier with the training dataset.
Parameters
----------
X : array-like matrix of shape (n_samples, n_features)
Training data.
y : array-like matrix of shape (n_samples,)
Target values.
Returns
-------
self : DeodelSecond
The fitted classifier.
"""
ret = Working.WorkFit( self, X, y )
return ret
def predict(self, X):
"""Predict the class labels for the provided data.
Parameters
----------
X : array-like of shape n_queries, n_features
Test samples.
Returns
-------
y : list of (n_queries,)
Class labels for each data sample.
"""
ret = Working.WorkPredict( self, X )
return ret
# >-----------------------------------------------------------------------------
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# > Working - Begin
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
class Working:
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def WorkFit(object, in_X, in_y):
data_y = CasetDeodel.ListDataConvert(in_y)
if not data_y == [] :
if (isinstance(data_y[0], list)) :
if len(data_y[0]) == 1 :
# must be a one column matrix
transpose_data = Working.MatrixTranspose(data_y)
data_y = transpose_data[0]
else :
data_y = None
data_X = CasetDeodel.ListDataConvert(in_X)
split_no = object.cfg_param['split_no']
split_mode = object.cfg_param['split_mode']
predict_mode = object.cfg_param['predict_mode']
score_factor = object.cfg_param['score_factor']
int_is_num = object.cfg_param['int_is_num']
legacy_match = object.cfg_param['legacy_match']
ret_item = Working.DicretizeTable( data_X, split_no, split_mode, int_is_num )
(ret_cat_tbl, ret_num_tbl, ret_mask_tbl, ret_attr_num_cfg, ret_attr_dict_list) = ret_item
object.attr_cat_X = np.array(ret_cat_tbl, dtype='int')
object.attr_num_X = np.array(ret_num_tbl, dtype='float')
object.attr_mask_X = np.array(ret_mask_tbl, dtype='int')
object.attr_num_cfg = ret_attr_num_cfg
object.attr_dict_list = ret_attr_dict_list
object.attr_split_no = split_no
object.attr_split_mode = split_mode
object.attr_score_factor = score_factor
object.attr_int_is_num = int_is_num
object.attr_legacy_match = legacy_match
if predict_mode == 'auto' :
regress_flag = Working.RegressParse(data_y, int_is_num)
elif predict_mode == 'regress' :
regress_flag = True
else :
# 'classif'
regress_flag = False
object.regress_flag = regress_flag
object.targ_y = data_y
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def DicretizeTable(in_tbl, in_split_no, in_split_mode, in_int_is_num):
data_tbl = in_tbl
entry_no = len(data_tbl)
# determine column dimension (for incomplete lists of lists)
max_col_no = 0
for crt_row in data_tbl :
crt_rlen = len(crt_row)
if crt_rlen > max_col_no :
max_col_no = crt_rlen
attr_no = max_col_no
attr_cat_tbl = []
attr_num_tbl = []
attr_mask_tbl = []
attr_dict_list = []
attr_num_thresh = []
split_num = in_split_no
for crt_idx in range(attr_no) :
crt_col = Working.GetCol( data_tbl, crt_idx )
ret_tuple = Working.ProcessVector(crt_col, in_int_is_num)
(conv_cat, conv_num, conv_mask, shadow_dict, shadrev_dict, numerical_list) = ret_tuple
# create numerical thresholds if numerical elements are present
if len(numerical_list) > 0 :
if in_split_mode == 'eq_width' :
ord_num_list = sorted(numerical_list)
if ord_num_list[0] == ord_num_list[-1] :
crt_attr_thresh = [False, []]
else :
crt_attr_thresh = [True, [ord_num_list[0], ord_num_list[-1]]]
elif in_split_mode == 'eq_legacy_width' :
crt_attr_thresh = Working.NumSplit(numerical_list, in_split_no, in_split_mode)
else :
crt_attr_thresh = Working.NumSplit(numerical_list, in_split_no, in_split_mode)
else :
crt_attr_thresh = [False, []]
attr_num_thresh.append(crt_attr_thresh)
attr_dict_list.append(shadow_dict)
attr_cat_tbl.append(conv_cat)
attr_num_tbl.append(conv_num)
attr_mask_tbl.append(conv_mask)
# Replace numerical values with discretized/processed values
for crt_idx_1 in range(attr_no) :
crt_attr_mask_list = attr_mask_tbl[crt_idx_1]
crt_attr_num_list = attr_num_tbl[crt_idx_1]
for crt_idx_2 in range(entry_no) :
crt_mask_elem = crt_attr_mask_list[crt_idx_2]
crt_num_elem = crt_attr_num_list[crt_idx_2]
# [crt_flag, crt_thresh_list] = attr_num_thresh[crt_idx_1]
crt_thresh_list = attr_num_thresh[crt_idx_1]
if crt_mask_elem == 1 :
process_val = Working.DiscretizeAttrVal(crt_num_elem, crt_thresh_list, in_split_mode, in_split_no)
new_id = process_val
attr_num_tbl[crt_idx_1][crt_idx_2] = new_id
ret_cat_tbl = Working.MatrixTranspose(attr_cat_tbl)
ret_num_tbl = Working.MatrixTranspose(attr_num_tbl)
ret_mask_tbl = Working.MatrixTranspose(attr_mask_tbl)
ret_attr_num_cfg = attr_num_thresh
ret_attr_dict_list = attr_dict_list
ret_tuple = (ret_cat_tbl, ret_num_tbl, ret_mask_tbl, ret_attr_num_cfg, ret_attr_dict_list)
return ret_tuple
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def WorkPredict(object, in_query):
query_req = CasetDeodel.ListDataConvert(in_query)
# attr_no = len(object.attr_X[0])
# if attr_no != len( object.attr_num_cfg ) :
# raise ValueError( "Mismatch in the number of attributes" )
result_lst = []
for row in query_req :
crt_result = Working.PredictOne(object, row)
result_lst.append(crt_result)
return result_lst
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def DiscretizeAttrVal(in_cont_value, in_discretize_cfg, in_split_mode, in_split_no) :
# Discretize a continuous value using threshold list
crt_num_elem = in_cont_value
[crt_multi_flag, crt_thresh_list] = in_discretize_cfg
# start_no_id = len( attr_dict_list[crt_idx_1] ) + 1
start_no_id = 0.0
thresh_elem_no = len(crt_thresh_list)
if thresh_elem_no == 0 :
new_id = start_no_id
else :
if not crt_multi_flag :
# monocolor attribute
if in_cont_value == crt_thresh_list[0] :
proc_val = start_no_id + 1
else :
proc_val = start_no_id
else :
if in_split_mode == 'eq_width' :
proc_val = Working.EqualWidthDiscretize(crt_num_elem, crt_thresh_list, in_split_no)
elif in_split_mode == 'eq_legacy_width' :
thresh_div = len(crt_thresh_list)
upper_idx = Working.GetElemIdxInOrderList(crt_num_elem, crt_thresh_list)
proc_val = upper_idx / (1.0 * thresh_div)
else :
thresh_div = len(crt_thresh_list)
upper_idx = Working.GetElemIdxInOrderList(crt_num_elem, crt_thresh_list)
proc_val = upper_idx / (1.0 * thresh_div)
new_id = start_no_id + proc_val
return new_id
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def TranslateAttrEntry(object, in_attr_list):
# translate to conforming list
attr_entry = in_attr_list
attr_no = len(attr_entry)
transl_cat_query = []
transl_mask_query = []
transl_num_query = []
int_is_num = object.attr_int_is_num
for crt_idx in range(attr_no) :
crt_attr = attr_entry[crt_idx]
crt_num_cfg = object.attr_num_cfg[crt_idx]
crt_dict = object.attr_dict_list[crt_idx]
if crt_attr == None :
# None is considered to represent missing attribute values, will be ignored.
new_cat_id = -1
new_num_id = 0
new_mask_id = 0
else :
ret_tuple = Working.NumericalCheck(crt_attr, int_is_num)
is_numerical, translate_value = ret_tuple
if is_numerical :
# numerical attribute, discretize with interval
process_val = Working.DiscretizeAttrVal(translate_value, crt_num_cfg, object.attr_split_mode, object.attr_split_no)
new_num_id = process_val
new_mask_id = 1
new_cat_id = 0
else :
new_num_id = 0
new_mask_id = 0
if translate_value in crt_dict :
new_cat_id = crt_dict[translate_value]
else :
new_cat_id = 0
transl_cat_query.append(int(new_cat_id))
# transl_num_query.append(int(new_num_id + 0.5))
transl_num_query.append(float(new_num_id))
transl_mask_query.append(int(new_mask_id))
ret_item = (
np.array(transl_cat_query, dtype='int'),
np.array(transl_num_query, dtype='float'),
np.array(transl_mask_query, dtype='int'),
)
return ret_item
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def PredictOne(object, in_query):
# prediction for one query
query_len = len(in_query)
train_len = len(object.attr_cat_X[0])
int_is_num = object.attr_int_is_num
if query_len == train_len :
adjusted_query = in_query[:]
elif query_len > train_len :
adjusted_query = in_query[:train_len]
else :
# list comprehension
adjusted_query = in_query + [None for i in range(train_len - query_len)]
attr_no = query_len
query_cat_req, query_num_req, query_mask_req = Working.TranslateAttrEntry(object, adjusted_query)
# list comprehension
# match_score_list = [[] for i in range(attr_no + 1)]
match_score_list = [[] for i in range(attr_no * object.attr_score_factor + 1)]
# shadow_score_list = [[] for i in range(attr_no + 1)]
shadow_score_list = [[] for i in range(attr_no * object.attr_score_factor + 1)]
attr_rows = len(object.attr_cat_X)
targ_no = len(object.targ_y)
train_no = min(attr_rows, targ_no)
for crt_idx in range(train_no) :
crt_train_cat_attr = object.attr_cat_X[crt_idx]
compare_cat_vect = (np.equal(crt_train_cat_attr, query_cat_req)).astype(int)
entry_cat_match_score = int(np.count_nonzero(compare_cat_vect))
crt_train_mask_attr = object.attr_mask_X[crt_idx]
crt_train_num_attr = object.attr_num_X[crt_idx]
# discretize_div = object.attr_split_no - 1
discretize_div = 1.0
compare_num_vect = 1.0 - abs((crt_train_num_attr - query_num_req)/(1.0 * discretize_div))
compare_num_vect *= crt_train_mask_attr
compare_num_vect *= query_mask_req
if object.attr_legacy_match :
tmp_vector = compare_num_vect.astype(int)
compare_num_vect = np.where(tmp_vector == 1, 1, 0)
entry_num_match_score = sum(compare_num_vect)
entry_agg_match_score = int( (entry_cat_match_score + entry_num_match_score) * object.attr_score_factor + 0.5 )
if object.regress_flag :
crt_targ_elem = object.targ_y[crt_idx]
ret_tuple = Working.NumericalCheck(crt_targ_elem, int_is_num)
is_numerical, translate_value = ret_tuple
if is_numerical :
match_score_list[entry_agg_match_score].append((None, 0))
shadow_score_list[entry_agg_match_score].append(object.targ_y[crt_idx])
else :
match_score_list[entry_agg_match_score].append(object.targ_y[crt_idx])
else :
match_score_list[entry_agg_match_score].append(object.targ_y[crt_idx])
aux_cfg_data = {'top_first': False}
if 'tbreak_depth' in object.cfg_param :
if object.cfg_param['tbreak_depth'] == None :
pass
elif object.cfg_param['tbreak_depth'] > 0 :
aux_cfg_data['eval_limit'] = object.cfg_param['tbreak_depth']
ret_tuple = CasetDeodel.HelperRecurseTieBreaker(match_score_list, None, aux_cfg_data)
champ_sel = ret_tuple[2]
if object.regress_flag :
if champ_sel == (None, 0) :
# This indicates that the predicted outcome is numerical.
# Find top numerical entry.
# Top score is last non-empty entry
shadow_len = len(shadow_score_list)
for crt_idx in range(shadow_len) :
complmnt_idx = (shadow_len - 1) - crt_idx
crt_list = shadow_score_list[complmnt_idx]
if not crt_list == [] :
break
top_score = crt_idx
top_num_list = crt_list
vect_mean = statistics.mean(top_num_list)
champ_sel = vect_mean
predict_y = champ_sel
return predict_y
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def GetElemIdxInOrderList( in_elem, in_thresh_list ) :
ordered_list = in_thresh_list
new_element = in_elem
fn_ret_status = False
fn_ret_insert_idx = None
# one iteration loop to allow unified return through loop breaks
for dummy_idx in range(1) :
no_of_elem = len( ordered_list )
if no_of_elem == 0 :
fn_ret_status = True
fn_ret_insert_idx = 0
break
interval_len = no_of_elem
offset_idx = 0
while True :
half_len = int(interval_len/2)
middle_idx = half_len
middle_idx += offset_idx
crt_elem = ordered_list[middle_idx]
ret_compare = ( new_element < crt_elem )
if ret_compare :
# Correct order
interval_len = middle_idx - offset_idx
else :
# Not correct order
interval_len = interval_len - ( middle_idx - offset_idx ) - 1
offset_idx = middle_idx + 1
# check break condition
if interval_len <= 0 :
break
# After iterations completed the appropiate position is in offset_idx
fn_ret_status = True
fn_ret_insert_idx = offset_idx
ret_item = fn_ret_insert_idx
return(ret_item)
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def EqualWidthDiscretize( in_elem, in_minmax_list, in_split_no ) :
# list_minmax = sorted(in_minmax_list)
list_minmax = in_minmax_list
crt_num_elem = in_elem
val_span = (list_minmax[-1] - list_minmax[0]) * 1.0
if val_span > 0 :
if crt_num_elem < list_minmax[0] :
crt_num_elem = list_minmax[0]
if crt_num_elem > list_minmax[-1] :
crt_num_elem = list_minmax[-1]
input_offset = (crt_num_elem - list_minmax[0])
if in_split_no <= 1 :
# no discretization
convert_val = input_offset/(1.0 * val_span)
else :
unit_width = (val_span * 1.0)/(in_split_no - 1)
# list comprehension
# near_discr_offset = int(input_offset/(1.0 * unit_width) + 0.5)
# near_discr_offset = int(input_offset/(1.0 * unit_width) + 0.0)
interm_val = input_offset/(1.0 * unit_width)
discr_offset = int(interm_val + 0.5)
near_scaled_offset = discr_offset / (1.0 * (in_split_no - 1))
convert_val = near_scaled_offset
else :
if crt_num_elem >= list_minmax[0] :
convert_val = 1
else :
convert_val = 0
# convert_val = 0
ret_item = convert_val
return(ret_item)
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def GetCol( in_array, in_col ) :
ret_item = []
if not isinstance(in_array, list) :
ret_item = None
else :
row_no = len(in_array)
if not isinstance(in_array[0], list) :
ret_item = None
else :
ret_item = []
for crt_idx_row in range(row_no) :
crt_row_len = len(in_array[crt_idx_row])
if in_col < crt_row_len :
ret_item.append((in_array[crt_idx_row][in_col]))
else :
ret_item.append(None)
return(ret_item)
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def MatrixTranspose( in_array ) :
if in_array == [] : return []
if not isinstance(in_array[0], list) :
transp_data = list(in_array)
else :
transp_data = []
col_no = len(in_array[0])
for crt_idx_col in range(col_no) :
crt_vect = Working.GetCol( in_array, crt_idx_col )
transp_data.append(crt_vect)
ret_item = transp_data
return(ret_item)
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def ProcessVector( in_v, in_int_is_num ) :
vect_size = len(in_v)
shadow_id = 1
shadow_dict = {}
shadrev_dict = {}
numerical_list = []
conv_cat = []
conv_num = []
conv_mask = []
for crt_idx in range(vect_size) :
crt_elem = in_v[crt_idx]
is_num_flag, equiv_val = Working.NumericalCheck(crt_elem, in_int_is_num)
if is_num_flag :
numerical_list.append(crt_elem)
conv_num.append(crt_elem)
conv_cat.append(-2)
# mask one is used as a marker for numerical values
conv_mask.append(1)
else :
conv_num.append(-1)
if not equiv_val in shadow_dict :
shadow_dict[crt_elem] = shadow_id
shadrev_dict[shadow_id] = equiv_val
conv_cat.append(shadow_id)
shadow_id += 1
else :
conv_cat.append(shadow_dict[equiv_val])
conv_mask.append(0)
fn_ret_tuple = (conv_cat, conv_num, conv_mask, shadow_dict, shadrev_dict, numerical_list)
return fn_ret_tuple
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def NumSplit( in_num_list, in_split_no = 2, in_split_mode = 'eq_width', no_dupl_flag = True ) :
threshold_list = []
if in_split_no <= 1 :
return [False, threshold_list]
num_len = len(in_num_list)
if num_len == 0 :
return [False, threshold_list]
elif num_len == 1 :
if no_dupl_flag :
threshold_list = [in_num_list[0]]
else :
# list comprehension
threshold_list = [in_num_list[0] for i in range(in_split_no - 1)]
return [True, threshold_list]
if in_split_mode == 'eq_freq' :
threshold_list = Working.NumSplitFreq( in_num_list, in_split_no, no_dupl_flag )
elif in_split_mode == 'eq_width' :
threshold_list = Working.NumSplitWidth( in_num_list, in_split_no, no_dupl_flag )
elif in_split_mode == 'eq_legacy_width' :
threshold_list = Working.NumLegacySplitWidth( in_num_list, in_split_no, no_dupl_flag )
else :
# invalid
return [False, threshold_list]
return [True, threshold_list]
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def NumSplitFreq( in_num_list, in_split_no = 2, no_dupl_flag = True ) :
threshold_list = []
num_len = len(in_num_list)
if num_len == 0 :
return threshold_list
ord_num_list = sorted(in_num_list)
max_split_factor = 2
if in_split_no > max_split_factor * num_len :
split_num = max_split_factor * num_len
else :
split_num = int(in_split_no)
if ord_num_list[0] == ord_num_list[-1] :
if no_dupl_flag :
threshold_list = [ord_num_list[0]]
else :
# list comprehension
threshold_list = [ord_num_list[0] for i in range(in_split_no - 1)]
return threshold_list
# |.....|.....|.....|.....|.....|.....|.....
# x........x........x........x........x........
# .....x........x........x........x........x........
period_split = 1.0/split_num
period_ordrd = 1.0/num_len
sample_point_uscale = -period_ordrd / 2.0
sample_point_uscale += period_split
crt_split_len = 0
thresh_prev = None
while ((sample_point_uscale < (1 - period_ordrd/4.0))
and (crt_split_len <= (split_num - 2))) :
num_near_offset = (sample_point_uscale) / period_ordrd
num_near_idx = int(num_near_offset)
if num_near_idx == num_near_offset :
new_thresh = ord_num_list[num_near_idx]
else:
prev_ordrd_uscale = num_near_idx * period_ordrd
next_ordrd_uscale = (num_near_idx + 1) * period_ordrd
delta_fract = (sample_point_uscale - prev_ordrd_uscale) / period_ordrd
# threshold crossed
# 0.00 0.25 0.50 0.75 1.00
# ^
# a (b-a)*0.25 b
#
if num_near_idx + 1 < num_len :
x_interp = delta_fract
y_a = ord_num_list[num_near_idx]
y_b = ord_num_list[num_near_idx + 1]
y_interp = y_a + (y_b - y_a) * x_interp
new_thresh = y_interp
else :
new_thresh = ord_num_list[num_near_idx]
sampling_add = period_split
if no_dupl_flag :
if num_near_idx < num_len - 1 :
thresh_next = ord_num_list[num_near_idx + 1]
if not thresh_next == new_thresh :
threshold_list.append(new_thresh)
crt_split_len += 1
thresh_prev = new_thresh
else :
# skip append
sample_point_uscale = (num_near_idx + 1) * period_ordrd
else :
if not new_thresh == thresh_prev :
threshold_list.append(new_thresh)
crt_split_len += 1
thresh_prev = new_thresh
else :
# skip append
pass
else :
threshold_list.append(new_thresh)
crt_split_len += 1
thresh_prev = new_thresh
sample_point_uscale += sampling_add
if threshold_list == [] :
if ord_num_list[0] != ord_num_list[-1] :
# at least add the last of ordered list
last_elem = ord_num_list[-1]
threshold_list = [last_elem]
return threshold_list
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def NumSplitWidth( in_num_list, in_split_no = 2, no_dupl_flag = True ) :
threshold_list = []
num_len = len(in_num_list)
if num_len == 0 :
return threshold_list
ord_num_list = sorted(in_num_list)
num_min = ord_num_list[0]
num_max = ord_num_list[-1]
width = (num_max - num_min)/(in_split_no - 1.0)
if width == 0 :
if no_dupl_flag :
threshold_list = [num_min]
else :
# list comprehension
threshold_list = [in_num_list[0] for i in range(in_split_no - 1)]
else :
threshold_list = []
crt_thresh = num_min + width/2.0
for crt_idx in range(in_split_no - 1) :
threshold_list.append(crt_thresh)
crt_thresh += width
return threshold_list
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def NumLegacySplitWidth( in_num_list, in_split_no = 2, no_dupl_flag = True ) :
threshold_list = []
num_len = len(in_num_list)
if num_len == 0 :
return threshold_list
ord_num_list = sorted(in_num_list)
num_min = ord_num_list[0]
num_max = ord_num_list[-1]
width = (num_max - num_min)/(in_split_no * 1.0)
if width == 0 :
if no_dupl_flag :
threshold_list = [num_min]
else :
# list comprehension
threshold_list = [num_min for i in range(in_split_no - 1)]
else :
threshold_list = []
crt_thresh = num_min
for crt_idx in range(in_split_no - 1) :
crt_thresh += width
threshold_list.append(crt_thresh)
return threshold_list
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def RevertVector( in_cat_v, shadrev_dict, in_num_v ) :
vect_size = len(in_cat_v)
revrt_v = []
for crt_idx in range(vect_size) :
crt_elem = in_cat_v[crt_idx]
if crt_elem == -2 :
revrt_v.append(in_num_v[crt_idx])
else :
tr_elem = shadrev_dict[crt_elem]
revrt_v.append(tr_elem)
fn_ret = revrt_v
return fn_ret
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def NumericalCheck( in_value, int_is_num_flag = True ) :
"""
Check if numerical. If non regular float, result is false and
translated value returned
"""
if not int_is_num_flag :
if isinstance(in_value, float) :
float_flag, valid_val = CasetDeodel.ValidateFloat(in_value)
fn_ret = (float_flag, valid_val)
else :
fn_ret = (False, in_value)
else :
if isinstance(in_value, float) :
float_flag, valid_val = CasetDeodel.ValidateFloat(in_value)
fn_ret = (float_flag, valid_val)
elif isinstance(in_value, int) :
fn_ret = (True, in_value)
else :
fn_ret = (False, in_value)
return fn_ret
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def RegressParse( in_vect, in_int_is_num) :
"""
Parse vector and check if vector has valid regress numerical elements.
"""
regress_flag = False
num_list = []
for crt_elem in in_vect :
is_numerical, translate_value = Working.NumericalCheck(crt_elem, in_int_is_num)
if is_numerical :
num_list.append(crt_elem)
def_min_len = 10
def_min_chk = 4
def_check_fract = 1.0/def_min_len
len_num_list = len(num_list)
if len_num_list >= def_min_len :
# check whether list appears to be made of continuous values
ret_tuple = CasetDeodel.SummaryFreqCount(num_list)
crt_types_no, crt_id_list, crt_count_list = ret_tuple
if(crt_types_no >= def_min_len) :
fract_idx = int(crt_types_no * def_check_fract)
chk_idx = min(fract_idx, def_min_chk)
if chk_idx > 0 :
if crt_count_list[chk_idx] == 1 :
# majority of values are unique
regress_flag = True
return regress_flag
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# > Working - End
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# > CasetDeodel - Begin
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
class CasetDeodel:
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def OrderedFreqCount(in_symbol_sequence_list):
"""
Returns a list that contains info about the frequency of
items in parameter input list (in_symbol_sequence_list).
The list is sorted on no of occurences order
Params:
in_symbol_sequence_list
sequence of symbol occurences
returns:
out_list
The output list has the following structure:
each row (first level list) has:
first column the element itself from the list
second column the no of occurences
third column the list of indexes containing the element
"""
from operator import itemgetter
in_len = len(in_symbol_sequence_list)
idx = 0
out_list = []
for in_el in in_symbol_sequence_list:
found_match = 0
for out_el in out_list:
if in_el == out_el[0]:
out_el[1] = out_el[1] + 1
out_el[2].append(idx)
found_match = 1
break
if found_match == 0:
out_list.append([in_el, 1, []])
# append index into third column
out_list[-1][2].append(idx)
idx = idx + 1
out_list.sort(key=itemgetter(1), reverse=True)
return out_list
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def SummaryFreqCount(in_symbol_sequence_list):
"""
Returns a summary about the frequency of
items in parameter input list (in_symbol_sequence_list).
The list is sorted on no of occurences order
Params:
in_symbol_sequence_list
sequence of symbol occurences
returns:
ret_no_of_distinct_elems
ret_elem_list
List of distinct elements
ret_count_list
List with counts of each element matching ret_elem_list
"""
count_data = CasetDeodel.OrderedFreqCount(in_symbol_sequence_list)
distinct_elem_no, elem_list, count_list = CasetDeodel.CountDataToFreqLists(count_data)
return distinct_elem_no, elem_list, count_list
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def CountDataToFreqLists(in_freq_count_data):
"""
Returns a summary of info about the frequency of
items in parameter input list (in_freq_count_data).
The lists are sorted on no of occurences order
Params:
in_freq_count_data
sequence of symbol occurences
returns:
ret_no_of_distinct_elems
ret_elem_list
List of distinct elements
ret_count_list
List with counts of each element matching ret_elem_list
"""
# determine no of distinct elements
distinct_elem_no = len(in_freq_count_data)
# Filter out the index lists
elem_count_pairs = [ elem[:2] for elem in in_freq_count_data ]
elem_list = [ x[0] for x in elem_count_pairs ]
count_list = [ x[1] for x in elem_count_pairs ]
return distinct_elem_no, elem_list, count_list
# >- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@staticmethod
def HelperRecurseTieBreaker(in_ordered_outcome_table, in_sel_id_dict = None, aux_data = None) :
"""
Recursive tie breaker
Param:
in_ordered_outcome_table
in_sel_id_dict
aux_data
Auxilliary data
Return:
ret_status
ret_outcome_id
ret_outcome_count
"""
fn_ret_status = False
fn_tbreak_status = False
fn_ret_outcome_id = None
fn_ret_outcome_count = None
# one iteration loop to allow unified return through loop breaks
for dummy_idx in range(1) :
if (isinstance (in_sel_id_dict, dict) == False
and in_sel_id_dict != None ) :
break
if in_ordered_outcome_table == [] :
# empty list
fn_ret_status = False
fn_tbreak_status = False
fn_ret_outcome_id = None
fn_ret_outcome_count = 0
break
if aux_data == None :
aux_data = {'top_first': True}
else :
if not 'top_first' in aux_data :
aux_data['top_first'] = True
new_aux = dict(aux_data)
if not new_aux['top_first'] :
cumulate_outc_list = in_ordered_outcome_table[-1][:]
else :
cumulate_outc_list = in_ordered_outcome_table[0][:]
ret_tuple = CasetDeodel.SummaryFreqCount(cumulate_outc_list)
crt_types_no, crt_id_list, crt_count_list = ret_tuple
if not new_aux['top_first'] :
matchnum_score_list = in_ordered_outcome_table[:-1]
else :
matchnum_score_list = in_ordered_outcome_table[1:]
if 'eval_limit' in new_aux :
eval_limit = new_aux['eval_limit']
if eval_limit <= 0 :
break
selectid_dict = in_sel_id_dict
# look for the first valid outcome type
crt_max_idx = 0