/
preprocessor.html
1575 lines (1464 loc) · 62.2 KB
/
preprocessor.html
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
<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, minimum-scale=1" />
<meta name="generator" content="pdoc 0.10.0" />
<title>ktrain.tabular.preprocessor API documentation</title>
<meta name="description" content="" />
<link rel="preload stylesheet" as="style" href="https://cdnjs.cloudflare.com/ajax/libs/10up-sanitize.css/11.0.1/sanitize.min.css" integrity="sha256-PK9q560IAAa6WVRRh76LtCaI8pjTJ2z11v0miyNNjrs=" crossorigin>
<link rel="preload stylesheet" as="style" href="https://cdnjs.cloudflare.com/ajax/libs/10up-sanitize.css/11.0.1/typography.min.css" integrity="sha256-7l/o7C8jubJiy74VsKTidCy1yBkRtiUGbVkYBylBqUg=" crossorigin>
<link rel="stylesheet preload" as="style" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/10.1.1/styles/github.min.css" crossorigin>
<style>:root{--highlight-color:#fe9}.flex{display:flex !important}body{line-height:1.5em}#content{padding:20px}#sidebar{padding:30px;overflow:hidden}#sidebar > *:last-child{margin-bottom:2cm}.http-server-breadcrumbs{font-size:130%;margin:0 0 15px 0}#footer{font-size:.75em;padding:5px 30px;border-top:1px solid #ddd;text-align:right}#footer p{margin:0 0 0 1em;display:inline-block}#footer p:last-child{margin-right:30px}h1,h2,h3,h4,h5{font-weight:300}h1{font-size:2.5em;line-height:1.1em}h2{font-size:1.75em;margin:1em 0 .50em 0}h3{font-size:1.4em;margin:25px 0 10px 0}h4{margin:0;font-size:105%}h1:target,h2:target,h3:target,h4:target,h5:target,h6:target{background:var(--highlight-color);padding:.2em 0}a{color:#058;text-decoration:none;transition:color .3s ease-in-out}a:hover{color:#e82}.title code{font-weight:bold}h2[id^="header-"]{margin-top:2em}.ident{color:#900}pre code{background:#f8f8f8;font-size:.8em;line-height:1.4em}code{background:#f2f2f1;padding:1px 4px;overflow-wrap:break-word}h1 code{background:transparent}pre{background:#f8f8f8;border:0;border-top:1px solid #ccc;border-bottom:1px solid #ccc;margin:1em 0;padding:1ex}#http-server-module-list{display:flex;flex-flow:column}#http-server-module-list div{display:flex}#http-server-module-list dt{min-width:10%}#http-server-module-list p{margin-top:0}.toc ul,#index{list-style-type:none;margin:0;padding:0}#index code{background:transparent}#index h3{border-bottom:1px solid #ddd}#index ul{padding:0}#index h4{margin-top:.6em;font-weight:bold}@media (min-width:200ex){#index .two-column{column-count:2}}@media (min-width:300ex){#index .two-column{column-count:3}}dl{margin-bottom:2em}dl dl:last-child{margin-bottom:4em}dd{margin:0 0 1em 3em}#header-classes + dl > dd{margin-bottom:3em}dd dd{margin-left:2em}dd p{margin:10px 0}.name{background:#eee;font-weight:bold;font-size:.85em;padding:5px 10px;display:inline-block;min-width:40%}.name:hover{background:#e0e0e0}dt:target .name{background:var(--highlight-color)}.name > span:first-child{white-space:nowrap}.name.class > span:nth-child(2){margin-left:.4em}.inherited{color:#999;border-left:5px solid #eee;padding-left:1em}.inheritance em{font-style:normal;font-weight:bold}.desc h2{font-weight:400;font-size:1.25em}.desc h3{font-size:1em}.desc dt code{background:inherit}.source summary,.git-link-div{color:#666;text-align:right;font-weight:400;font-size:.8em;text-transform:uppercase}.source summary > *{white-space:nowrap;cursor:pointer}.git-link{color:inherit;margin-left:1em}.source pre{max-height:500px;overflow:auto;margin:0}.source pre code{font-size:12px;overflow:visible}.hlist{list-style:none}.hlist li{display:inline}.hlist li:after{content:',\2002'}.hlist li:last-child:after{content:none}.hlist .hlist{display:inline;padding-left:1em}img{max-width:100%}td{padding:0 .5em}.admonition{padding:.1em .5em;margin-bottom:1em}.admonition-title{font-weight:bold}.admonition.note,.admonition.info,.admonition.important{background:#aef}.admonition.todo,.admonition.versionadded,.admonition.tip,.admonition.hint{background:#dfd}.admonition.warning,.admonition.versionchanged,.admonition.deprecated{background:#fd4}.admonition.error,.admonition.danger,.admonition.caution{background:lightpink}</style>
<style media="screen and (min-width: 700px)">@media screen and (min-width:700px){#sidebar{width:30%;height:100vh;overflow:auto;position:sticky;top:0}#content{width:70%;max-width:100ch;padding:3em 4em;border-left:1px solid #ddd}pre code{font-size:1em}.item .name{font-size:1em}main{display:flex;flex-direction:row-reverse;justify-content:flex-end}.toc ul ul,#index ul{padding-left:1.5em}.toc > ul > li{margin-top:.5em}}</style>
<style media="print">@media print{#sidebar h1{page-break-before:always}.source{display:none}}@media print{*{background:transparent !important;color:#000 !important;box-shadow:none !important;text-shadow:none !important}a[href]:after{content:" (" attr(href) ")";font-size:90%}a[href][title]:after{content:none}abbr[title]:after{content:" (" attr(title) ")"}.ir a:after,a[href^="javascript:"]:after,a[href^="#"]:after{content:""}pre,blockquote{border:1px solid #999;page-break-inside:avoid}thead{display:table-header-group}tr,img{page-break-inside:avoid}img{max-width:100% !important}@page{margin:0.5cm}p,h2,h3{orphans:3;widows:3}h1,h2,h3,h4,h5,h6{page-break-after:avoid}}</style>
<script defer src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/10.1.1/highlight.min.js" integrity="sha256-Uv3H6lx7dJmRfRvH8TH6kJD1TSK1aFcwgx+mdg3epi8=" crossorigin></script>
<script>window.addEventListener('DOMContentLoaded', () => hljs.initHighlighting())</script>
</head>
<body>
<main>
<article id="content">
<header>
<h1 class="title">Module <code>ktrain.tabular.preprocessor</code></h1>
</header>
<section id="section-intro">
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">from .. import utils as U
from ..imports import *
from ..preprocessor import Preprocessor
class TabularPreprocessor(Preprocessor):
"""
```
Tabular preprocessing base class
```
"""
def __init__(
self,
predictor_columns,
label_columns,
date_columns=[],
is_regression=False,
procs=[],
max_card=20,
):
self.is_regression = is_regression
self.c = None
self.pc = predictor_columns
self.lc = label_columns
self.lc = [self.lc] if isinstance(self.lc, str) else self.lc
self.dc = date_columns
self.label_columns = None
self.cat_names = []
self.cont_names = []
self.date_names = []
self.label_transform = None
self.procs = procs
self.max_card = max_card
@property
def na_names(self):
return [n for n in self.cat_names if n[-3:] == "_na"]
def get_preprocessor(self):
return (self.label_transform, self.procs)
def get_classes(self):
return self.label_columns if not self.is_regression else []
def preprocess(self, df):
return self.preprocess_test(df)
def _validate_columns(self, df):
missing_columns = []
for col in df.columns.values:
if col not in self.lc and col not in self.pc:
missing_columns.append(col)
if len(missing_columns) > 0:
raise ValueError("df is missing columns: %s" % (missing_columns))
return
def denormalize(self, df):
normalizer = None
for proc in self.procs:
if type(proc).__name__ == "Normalize":
normalizer = proc
break
if normalizer is None:
return df
return normalizer.revert(df)
# def codify(self, df):
# df = df.copy()
# for lab in self.lc:
# df[lab] = df[lab].cat.codes
# return df
def preprocess_train(self, df, mode="train", verbose=1):
"""
```
preprocess training set
```
"""
df = df.copy()
clean_df(df, pc=self.pc, lc=self.lc, check_labels=mode == "train")
if not isinstance(df, pd.DataFrame):
raise ValueError("df must be a pd.DataFrame")
# validate columns
self._validate_columns(df)
# validate mode
# if mode != 'train' and self.label_transform is None:
# raise ValueError('self.label_transform is None but mode is %s: are you sure preprocess_train was invoked first?' % (mode))
# verbose
if verbose:
print(
"processing %s: %s rows x %s columns" % (mode, df.shape[0], df.shape[1])
)
# convert date fields
for field in self.dc:
df = df.copy() # TODO: fix this
df, date_names = add_datepart(df, field)
self.date_names = date_names
# preprocess labels and data
if mode == "train":
label_columns = self.lc[:]
# label_columns.sort() # leave label columns sorted in same order as in DataFrame
self.label_transform = U.YTransformDataFrame(
label_columns, is_regression=self.is_regression
)
df = self.label_transform.apply_train(df)
self.label_columns = (
self.label_transform.get_classes()
if not self.is_regression
else self.label_transform.label_columns
)
self.cont_names, self.cat_names = cont_cat_split(
df, label_columns=self.label_columns, max_card=self.max_card
)
self.procs = [
proc(self.cat_names, self.cont_names) for proc in self.procs
] # "objectivy"
else:
df = self.label_transform.apply_test(df)
for proc in self.procs:
proc(df, test=mode != "train") # apply processors
from .dataset import TabularDataset
return TabularDataset(df, self.cat_names, self.cont_names, self.label_columns)
def preprocess_valid(self, df, verbose=1):
"""
```
preprocess validation set
```
"""
return self.preprocess_train(df, mode="valid", verbose=verbose)
def preprocess_test(self, df, verbose=1):
"""
```
preprocess test set
```
"""
return self.preprocess_train(df, mode="test", verbose=verbose)
def pd_data_types(df, return_df=False):
"""
```
infers data type of each column in Pandas DataFrame
Args:
df(pd.DataFrame): pandas DataFrame
return_df(bool): If True, returns columns and types in DataFrame.
Otherwise, a dictionary is returned.
```
"""
infer_type = lambda x: pd.api.types.infer_dtype(x, skipna=True)
df.apply(infer_type, axis=0)
# DataFrame with column names & new types
df_types = (
pd.DataFrame(df.apply(pd.api.types.infer_dtype, axis=0))
.reset_index()
.rename(columns={"index": "column", 0: "type"})
)
if return_df:
return df_types
cols = list(df_types["column"].values)
col_types = list(df_types["type"].values)
return dict(list(zip(cols, col_types)))
def clean_df(
train_df, val_df=None, pc=[], lc=[], check_labels=True, return_types=False
):
train_type_dict = pd_data_types(train_df)
for k, v in train_type_dict.items():
if v != "string":
continue
train_df[k] = train_df[k].str.strip()
if val_df is not None:
if k not in val_df.columns:
raise ValueError("val_df is missing %s column" % (k))
val_df[k] = val_df[k].str.strip()
if (pc and not lc) or (not pc and lc):
raise ValueError("pc and lc: both or neither must exist")
if pc and lc:
inp_cols = (
train_df.columns.values
if check_labels
else [col for col in train_df.columns.values if col not in lc]
)
original_cols = pc + lc if check_labels else pc
if set(original_cols) != set(inp_cols):
raise ValueError(
"DataFrame is either missing columns or includes extra columns: \n"
+ "expected: %s\nactual: %s" % (original_cols, inp_cols)
)
if return_types:
return train_type_dict
return
# --------------------------------------------------------------------
# These are helper functions adapted from fastai:
# https://github.com/fastai/fastai
# -------------------------------------------------------------------
from numbers import Number
from types import SimpleNamespace
from typing import (
Any,
AnyStr,
Callable,
Collection,
Dict,
Hashable,
Iterator,
List,
Mapping,
NewType,
Optional,
Sequence,
Tuple,
TypeVar,
Union,
)
from pandas.api.types import is_categorical_dtype, is_numeric_dtype
def ifnone(a, b):
"`a` if `a` is not None, otherwise `b`."
return b if a is None else a
def make_date(df, date_field):
"""
Make sure `df[field_name]` is of the right date type.
Reference: https://github.com/fastai/fastai
"""
field_dtype = df[date_field].dtype
if isinstance(field_dtype, pd.core.dtypes.dtypes.DatetimeTZDtype):
field_dtype = np.datetime64
if not np.issubdtype(field_dtype, np.datetime64):
df[date_field] = pd.to_datetime(df[date_field], infer_datetime_format=True)
return
def cont_cat_split(df, max_card=20, label_columns=[]):
"Helper function that returns column names of cont and cat variables from given df."
cont_names, cat_names = [], []
for col in df:
if col in label_columns:
continue
if (
df[col].dtype == int
and df[col].unique().shape[0] > max_card
or df[col].dtype == float
):
cont_names.append(col)
else:
cat_names.append(col)
return cont_names, cat_names
def add_datepart(
df: pd.DataFrame,
field_name: str,
prefix: str = None,
drop: bool = True,
time: bool = False,
return_added_columns=True,
):
"Helper function that adds columns relevant to a date in the column `field_name` of `df`."
make_date(df, field_name)
field = df[field_name]
prefix = ifnone(prefix, re.sub("[Dd]ate$", "", field_name))
attr = [
"Year",
"Month",
"Week",
"Day",
"Dayofweek",
"Dayofyear",
"Is_month_end",
"Is_month_start",
"Is_quarter_end",
"Is_quarter_start",
"Is_year_end",
"Is_year_start",
]
if time:
attr = attr + ["Hour", "Minute", "Second"]
added_columns = []
for n in attr:
df[prefix + n] = getattr(field.dt, n.lower())
added_columns.append(prefix + n)
df[prefix + "Elapsed"] = field.astype(np.int64) // 10**9
if drop:
df.drop(field_name, axis=1, inplace=True)
if return_added_columns:
return (df, added_columns)
else:
return df
def cyclic_dt_feat_names(time: bool = True, add_linear: bool = False) -> List[str]:
"Return feature names of date/time cycles as produced by `cyclic_dt_features`."
fs = ["cos", "sin"]
attr = [
f"{r}_{f}" for r in "weekday day_month month_year day_year".split() for f in fs
]
if time:
attr += [f"{r}_{f}" for r in "hour clock min sec".split() for f in fs]
if add_linear:
attr.append("year_lin")
return attr
def cyclic_dt_features(d, time: bool = True, add_linear: bool = False) -> List[float]:
"Calculate the cos and sin of date/time cycles."
tt, fs = d.timetuple(), [np.cos, np.sin]
day_year, days_month = tt.tm_yday, calendar.monthrange(d.year, d.month)[1]
days_year = 366 if calendar.isleap(d.year) else 365
rs = (
d.weekday() / 7,
(d.day - 1) / days_month,
(d.month - 1) / 12,
(day_year - 1) / days_year,
)
feats = [f(r * 2 * np.pi) for r in rs for f in fs]
if time and isinstance(d, datetime) and type(d) != date:
rs = tt.tm_hour / 24, tt.tm_hour % 12 / 12, tt.tm_min / 60, tt.tm_sec / 60
feats += [f(r * 2 * np.pi) for r in rs for f in fs]
if add_linear:
if type(d) == date:
feats.append(d.year + rs[-1])
else:
secs_in_year = (
datetime(d.year + 1, 1, 1) - datetime(d.year, 1, 1)
).total_seconds()
feats.append(
d.year + ((d - datetime(d.year, 1, 1)).total_seconds() / secs_in_year)
)
return feats
def add_cyclic_datepart(
df: pd.DataFrame,
field_name: str,
prefix: str = None,
drop: bool = True,
time: bool = False,
add_linear: bool = False,
):
"Helper function that adds trigonometric date/time features to a date in the column `field_name` of `df`."
make_date(df, field_name)
field = df[field_name]
prefix = ifnone(prefix, re.sub("[Dd]ate$", "", field_name))
series = field.apply(partial(cyclic_dt_features, time=time, add_linear=add_linear))
columns = [prefix + c for c in cyclic_dt_feat_names(time, add_linear)]
df_feats = pd.DataFrame(
[item for item in series], columns=columns, index=series.index
)
for column in columns:
df[column] = df_feats[column]
if drop:
df.drop(field_name, axis=1, inplace=True)
return df
class TabularProc:
"A processor for tabular dataframes."
def __init__(self, cat_names, cont_names):
self.cat_names = cat_names
self.cont_names = cont_names
def __call__(self, df, test=False):
"Apply the correct function to `df` depending on `test`."
func = self.apply_test if test else self.apply_train
func(df)
def apply_train(self, df):
"Function applied to `df` if it's the train set."
raise NotImplementedError
def apply_test(self, df):
"Function applied to `df` if it's the test set."
self.apply_train(df)
class Categorify(TabularProc):
def __init__(self, cat_names, cont_names):
super().__init__(cat_names, cont_names)
self.categories = None
def apply_train(self, df):
self.categories = {}
for n in self.cat_names:
df.loc[:, n] = df.loc[:, n].astype("category").cat.as_ordered()
self.categories[n] = df[n].cat.categories
def apply_test(self, df):
for n in self.cat_names:
df.loc[:, n] = pd.Categorical(
df[n], categories=self.categories[n], ordered=True
)
FILL_MEDIAN = "median"
FILL_CONSTANT = "constant"
class FillMissing(TabularProc):
"Fill the missing values in continuous columns."
def __init__(
self,
cat_names,
cont_names,
fill_strategy=FILL_MEDIAN,
add_col=True,
fill_val=0.0,
):
super().__init__(cat_names, cont_names)
self.fill_strategy = fill_strategy
self.add_col = add_col
self.fill_val = fill_val
self.na_dict = None
def apply_train(self, df):
self.na_dict = {}
self.filler_dict = {}
for name in self.cont_names:
if self.fill_strategy == FILL_MEDIAN:
filler = df[name].median()
elif self.fill_strategy == FILL_CONSTANT:
filler = self.fill_val
else:
filler = df[name].dropna().value_counts().idxmax()
self.filler_dict[name] = filler
if pd.isnull(df[name]).sum():
if self.add_col:
df[name + "_na"] = pd.isnull(df[name])
if name + "_na" not in self.cat_names:
self.cat_names.append(name + "_na")
df[name] = df[name].fillna(filler)
self.na_dict[name] = True
def apply_test(self, df):
"Fill missing values in `self.cont_names` like in `apply_train`."
for name in self.cont_names:
if name in self.na_dict:
if self.add_col:
df[name + "_na"] = pd.isnull(df[name])
if name + "_na" not in self.cat_names:
self.cat_names.append(name + "_na")
df[name] = df[name].fillna(self.filler_dict[name])
elif pd.isnull(df[name]).sum() != 0:
warnings.warn(
f"""There are nan values in field {name} but there were none in the training set.
Filled with {self.fill_strategy}."""
)
df[name] = df[name].fillna(self.filler_dict[name])
# raise Exception(f"""There are nan values in field {name} but there were none in the training set.
# Please fix those manually.""")
class Normalize(TabularProc):
"Normalize the continuous variables."
def __init__(self, cat_names, cont_names):
super().__init__(cat_names, cont_names)
self.means = None
self.stds = None
def apply_train(self, df):
"Compute the means and stds of `self.cont_names` columns to normalize them."
self.means, self.stds = {}, {}
for n in self.cont_names:
assert is_numeric_dtype(
df[n]
), f"""Cannot normalize '{n}' column as it isn't numerical.
Are you sure it doesn't belong in the categorical set of columns?"""
self.means[n], self.stds[n] = df[n].mean(), df[n].std()
df[n] = (df[n] - self.means[n]) / (1e-7 + self.stds[n])
def apply_test(self, df):
"Normalize `self.cont_names` with the same statistics as in `apply_train`."
for n in self.cont_names:
df[n] = (df[n] - self.means[n]) / (1e-7 + self.stds[n])
def revert(self, df):
"""
Undoes normalization and returns reverted dataframe
"""
out_df = df.copy()
for n in self.cont_names:
out_df[n] = (df[n] * (1e-7 + self.stds[n])) + self.means[n]
return out_df</code></pre>
</details>
</section>
<section>
</section>
<section>
</section>
<section>
<h2 class="section-title" id="header-functions">Functions</h2>
<dl>
<dt id="ktrain.tabular.preprocessor.add_cyclic_datepart"><code class="name flex">
<span>def <span class="ident">add_cyclic_datepart</span></span>(<span>df: pandas.core.frame.DataFrame, field_name: str, prefix: str = None, drop: bool = True, time: bool = False, add_linear: bool = False)</span>
</code></dt>
<dd>
<div class="desc"><p>Helper function that adds trigonometric date/time features to a date in the column <code>field_name</code> of <code>df</code>.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def add_cyclic_datepart(
df: pd.DataFrame,
field_name: str,
prefix: str = None,
drop: bool = True,
time: bool = False,
add_linear: bool = False,
):
"Helper function that adds trigonometric date/time features to a date in the column `field_name` of `df`."
make_date(df, field_name)
field = df[field_name]
prefix = ifnone(prefix, re.sub("[Dd]ate$", "", field_name))
series = field.apply(partial(cyclic_dt_features, time=time, add_linear=add_linear))
columns = [prefix + c for c in cyclic_dt_feat_names(time, add_linear)]
df_feats = pd.DataFrame(
[item for item in series], columns=columns, index=series.index
)
for column in columns:
df[column] = df_feats[column]
if drop:
df.drop(field_name, axis=1, inplace=True)
return df</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.add_datepart"><code class="name flex">
<span>def <span class="ident">add_datepart</span></span>(<span>df: pandas.core.frame.DataFrame, field_name: str, prefix: str = None, drop: bool = True, time: bool = False, return_added_columns=True)</span>
</code></dt>
<dd>
<div class="desc"><p>Helper function that adds columns relevant to a date in the column <code>field_name</code> of <code>df</code>.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def add_datepart(
df: pd.DataFrame,
field_name: str,
prefix: str = None,
drop: bool = True,
time: bool = False,
return_added_columns=True,
):
"Helper function that adds columns relevant to a date in the column `field_name` of `df`."
make_date(df, field_name)
field = df[field_name]
prefix = ifnone(prefix, re.sub("[Dd]ate$", "", field_name))
attr = [
"Year",
"Month",
"Week",
"Day",
"Dayofweek",
"Dayofyear",
"Is_month_end",
"Is_month_start",
"Is_quarter_end",
"Is_quarter_start",
"Is_year_end",
"Is_year_start",
]
if time:
attr = attr + ["Hour", "Minute", "Second"]
added_columns = []
for n in attr:
df[prefix + n] = getattr(field.dt, n.lower())
added_columns.append(prefix + n)
df[prefix + "Elapsed"] = field.astype(np.int64) // 10**9
if drop:
df.drop(field_name, axis=1, inplace=True)
if return_added_columns:
return (df, added_columns)
else:
return df</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.clean_df"><code class="name flex">
<span>def <span class="ident">clean_df</span></span>(<span>train_df, val_df=None, pc=[], lc=[], check_labels=True, return_types=False)</span>
</code></dt>
<dd>
<div class="desc"></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def clean_df(
train_df, val_df=None, pc=[], lc=[], check_labels=True, return_types=False
):
train_type_dict = pd_data_types(train_df)
for k, v in train_type_dict.items():
if v != "string":
continue
train_df[k] = train_df[k].str.strip()
if val_df is not None:
if k not in val_df.columns:
raise ValueError("val_df is missing %s column" % (k))
val_df[k] = val_df[k].str.strip()
if (pc and not lc) or (not pc and lc):
raise ValueError("pc and lc: both or neither must exist")
if pc and lc:
inp_cols = (
train_df.columns.values
if check_labels
else [col for col in train_df.columns.values if col not in lc]
)
original_cols = pc + lc if check_labels else pc
if set(original_cols) != set(inp_cols):
raise ValueError(
"DataFrame is either missing columns or includes extra columns: \n"
+ "expected: %s\nactual: %s" % (original_cols, inp_cols)
)
if return_types:
return train_type_dict
return</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.cont_cat_split"><code class="name flex">
<span>def <span class="ident">cont_cat_split</span></span>(<span>df, max_card=20, label_columns=[])</span>
</code></dt>
<dd>
<div class="desc"><p>Helper function that returns column names of cont and cat variables from given df.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def cont_cat_split(df, max_card=20, label_columns=[]):
"Helper function that returns column names of cont and cat variables from given df."
cont_names, cat_names = [], []
for col in df:
if col in label_columns:
continue
if (
df[col].dtype == int
and df[col].unique().shape[0] > max_card
or df[col].dtype == float
):
cont_names.append(col)
else:
cat_names.append(col)
return cont_names, cat_names</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.cyclic_dt_feat_names"><code class="name flex">
<span>def <span class="ident">cyclic_dt_feat_names</span></span>(<span>time: bool = True, add_linear: bool = False) ‑> List[str]</span>
</code></dt>
<dd>
<div class="desc"><p>Return feature names of date/time cycles as produced by <code><a title="ktrain.tabular.preprocessor.cyclic_dt_features" href="#ktrain.tabular.preprocessor.cyclic_dt_features">cyclic_dt_features()</a></code>.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def cyclic_dt_feat_names(time: bool = True, add_linear: bool = False) -> List[str]:
"Return feature names of date/time cycles as produced by `cyclic_dt_features`."
fs = ["cos", "sin"]
attr = [
f"{r}_{f}" for r in "weekday day_month month_year day_year".split() for f in fs
]
if time:
attr += [f"{r}_{f}" for r in "hour clock min sec".split() for f in fs]
if add_linear:
attr.append("year_lin")
return attr</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.cyclic_dt_features"><code class="name flex">
<span>def <span class="ident">cyclic_dt_features</span></span>(<span>d, time: bool = True, add_linear: bool = False) ‑> List[float]</span>
</code></dt>
<dd>
<div class="desc"><p>Calculate the cos and sin of date/time cycles.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def cyclic_dt_features(d, time: bool = True, add_linear: bool = False) -> List[float]:
"Calculate the cos and sin of date/time cycles."
tt, fs = d.timetuple(), [np.cos, np.sin]
day_year, days_month = tt.tm_yday, calendar.monthrange(d.year, d.month)[1]
days_year = 366 if calendar.isleap(d.year) else 365
rs = (
d.weekday() / 7,
(d.day - 1) / days_month,
(d.month - 1) / 12,
(day_year - 1) / days_year,
)
feats = [f(r * 2 * np.pi) for r in rs for f in fs]
if time and isinstance(d, datetime) and type(d) != date:
rs = tt.tm_hour / 24, tt.tm_hour % 12 / 12, tt.tm_min / 60, tt.tm_sec / 60
feats += [f(r * 2 * np.pi) for r in rs for f in fs]
if add_linear:
if type(d) == date:
feats.append(d.year + rs[-1])
else:
secs_in_year = (
datetime(d.year + 1, 1, 1) - datetime(d.year, 1, 1)
).total_seconds()
feats.append(
d.year + ((d - datetime(d.year, 1, 1)).total_seconds() / secs_in_year)
)
return feats</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.ifnone"><code class="name flex">
<span>def <span class="ident">ifnone</span></span>(<span>a, b)</span>
</code></dt>
<dd>
<div class="desc"><p><code>a</code> if <code>a</code> is not None, otherwise <code>b</code>.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def ifnone(a, b):
"`a` if `a` is not None, otherwise `b`."
return b if a is None else a</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.make_date"><code class="name flex">
<span>def <span class="ident">make_date</span></span>(<span>df, date_field)</span>
</code></dt>
<dd>
<div class="desc"><p>Make sure <code>df[field_name]</code> is of the right date type.
Reference: <a href="https://github.com/fastai/fastai">https://github.com/fastai/fastai</a></p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def make_date(df, date_field):
"""
Make sure `df[field_name]` is of the right date type.
Reference: https://github.com/fastai/fastai
"""
field_dtype = df[date_field].dtype
if isinstance(field_dtype, pd.core.dtypes.dtypes.DatetimeTZDtype):
field_dtype = np.datetime64
if not np.issubdtype(field_dtype, np.datetime64):
df[date_field] = pd.to_datetime(df[date_field], infer_datetime_format=True)
return</code></pre>
</details>
</dd>
<dt id="ktrain.tabular.preprocessor.pd_data_types"><code class="name flex">
<span>def <span class="ident">pd_data_types</span></span>(<span>df, return_df=False)</span>
</code></dt>
<dd>
<div class="desc"><pre><code>infers data type of each column in Pandas DataFrame
Args:
df(pd.DataFrame): pandas DataFrame
return_df(bool): If True, returns columns and types in DataFrame.
Otherwise, a dictionary is returned.
</code></pre></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def pd_data_types(df, return_df=False):
"""
```
infers data type of each column in Pandas DataFrame
Args:
df(pd.DataFrame): pandas DataFrame
return_df(bool): If True, returns columns and types in DataFrame.
Otherwise, a dictionary is returned.
```
"""
infer_type = lambda x: pd.api.types.infer_dtype(x, skipna=True)
df.apply(infer_type, axis=0)
# DataFrame with column names & new types
df_types = (
pd.DataFrame(df.apply(pd.api.types.infer_dtype, axis=0))
.reset_index()
.rename(columns={"index": "column", 0: "type"})
)
if return_df:
return df_types
cols = list(df_types["column"].values)
col_types = list(df_types["type"].values)
return dict(list(zip(cols, col_types)))</code></pre>
</details>
</dd>
</dl>
</section>
<section>
<h2 class="section-title" id="header-classes">Classes</h2>
<dl>
<dt id="ktrain.tabular.preprocessor.Categorify"><code class="flex name class">
<span>class <span class="ident">Categorify</span></span>
<span>(</span><span>cat_names, cont_names)</span>
</code></dt>
<dd>
<div class="desc"><p>A processor for tabular dataframes.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">class Categorify(TabularProc):
def __init__(self, cat_names, cont_names):
super().__init__(cat_names, cont_names)
self.categories = None
def apply_train(self, df):
self.categories = {}
for n in self.cat_names:
df.loc[:, n] = df.loc[:, n].astype("category").cat.as_ordered()
self.categories[n] = df[n].cat.categories
def apply_test(self, df):
for n in self.cat_names:
df.loc[:, n] = pd.Categorical(
df[n], categories=self.categories[n], ordered=True
)</code></pre>
</details>
<h3>Ancestors</h3>
<ul class="hlist">
<li><a title="ktrain.tabular.preprocessor.TabularProc" href="#ktrain.tabular.preprocessor.TabularProc">TabularProc</a></li>
</ul>
<h3>Inherited members</h3>
<ul class="hlist">
<li><code><b><a title="ktrain.tabular.preprocessor.TabularProc" href="#ktrain.tabular.preprocessor.TabularProc">TabularProc</a></b></code>:
<ul class="hlist">
<li><code><a title="ktrain.tabular.preprocessor.TabularProc.apply_test" href="#ktrain.tabular.preprocessor.TabularProc.apply_test">apply_test</a></code></li>
<li><code><a title="ktrain.tabular.preprocessor.TabularProc.apply_train" href="#ktrain.tabular.preprocessor.TabularProc.apply_train">apply_train</a></code></li>
</ul>
</li>
</ul>
</dd>
<dt id="ktrain.tabular.preprocessor.FillMissing"><code class="flex name class">
<span>class <span class="ident">FillMissing</span></span>
<span>(</span><span>cat_names, cont_names, fill_strategy='median', add_col=True, fill_val=0.0)</span>
</code></dt>
<dd>
<div class="desc"><p>Fill the missing values in continuous columns.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">class FillMissing(TabularProc):
"Fill the missing values in continuous columns."
def __init__(
self,
cat_names,
cont_names,
fill_strategy=FILL_MEDIAN,
add_col=True,
fill_val=0.0,
):
super().__init__(cat_names, cont_names)
self.fill_strategy = fill_strategy
self.add_col = add_col
self.fill_val = fill_val
self.na_dict = None
def apply_train(self, df):
self.na_dict = {}
self.filler_dict = {}
for name in self.cont_names:
if self.fill_strategy == FILL_MEDIAN:
filler = df[name].median()
elif self.fill_strategy == FILL_CONSTANT:
filler = self.fill_val
else:
filler = df[name].dropna().value_counts().idxmax()
self.filler_dict[name] = filler
if pd.isnull(df[name]).sum():
if self.add_col:
df[name + "_na"] = pd.isnull(df[name])
if name + "_na" not in self.cat_names:
self.cat_names.append(name + "_na")
df[name] = df[name].fillna(filler)
self.na_dict[name] = True
def apply_test(self, df):
"Fill missing values in `self.cont_names` like in `apply_train`."
for name in self.cont_names:
if name in self.na_dict:
if self.add_col:
df[name + "_na"] = pd.isnull(df[name])
if name + "_na" not in self.cat_names:
self.cat_names.append(name + "_na")
df[name] = df[name].fillna(self.filler_dict[name])
elif pd.isnull(df[name]).sum() != 0:
warnings.warn(
f"""There are nan values in field {name} but there were none in the training set.
Filled with {self.fill_strategy}."""
)
df[name] = df[name].fillna(self.filler_dict[name])</code></pre>
</details>
<h3>Ancestors</h3>
<ul class="hlist">
<li><a title="ktrain.tabular.preprocessor.TabularProc" href="#ktrain.tabular.preprocessor.TabularProc">TabularProc</a></li>
</ul>
<h3>Methods</h3>
<dl>
<dt id="ktrain.tabular.preprocessor.FillMissing.apply_test"><code class="name flex">
<span>def <span class="ident">apply_test</span></span>(<span>self, df)</span>
</code></dt>
<dd>
<div class="desc"><p>Fill missing values in <code>self.cont_names</code> like in <code>apply_train</code>.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">def apply_test(self, df):
"Fill missing values in `self.cont_names` like in `apply_train`."
for name in self.cont_names:
if name in self.na_dict:
if self.add_col:
df[name + "_na"] = pd.isnull(df[name])
if name + "_na" not in self.cat_names:
self.cat_names.append(name + "_na")
df[name] = df[name].fillna(self.filler_dict[name])
elif pd.isnull(df[name]).sum() != 0:
warnings.warn(
f"""There are nan values in field {name} but there were none in the training set.
Filled with {self.fill_strategy}."""
)
df[name] = df[name].fillna(self.filler_dict[name])</code></pre>
</details>
</dd>
</dl>
<h3>Inherited members</h3>
<ul class="hlist">
<li><code><b><a title="ktrain.tabular.preprocessor.TabularProc" href="#ktrain.tabular.preprocessor.TabularProc">TabularProc</a></b></code>:
<ul class="hlist">
<li><code><a title="ktrain.tabular.preprocessor.TabularProc.apply_train" href="#ktrain.tabular.preprocessor.TabularProc.apply_train">apply_train</a></code></li>
</ul>
</li>
</ul>
</dd>
<dt id="ktrain.tabular.preprocessor.Normalize"><code class="flex name class">
<span>class <span class="ident">Normalize</span></span>
<span>(</span><span>cat_names, cont_names)</span>
</code></dt>
<dd>
<div class="desc"><p>Normalize the continuous variables.</p></div>
<details class="source">
<summary>
<span>Expand source code</span>
</summary>
<pre><code class="python">class Normalize(TabularProc):
"Normalize the continuous variables."
def __init__(self, cat_names, cont_names):
super().__init__(cat_names, cont_names)
self.means = None
self.stds = None
def apply_train(self, df):
"Compute the means and stds of `self.cont_names` columns to normalize them."