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
|
import numpy as np
from sklearn.preprocessing import LabelBinarizer
class Transformer:
"""Base class for all 1-D transformers."""
def fit(self, X):
return self
def transform(self, X):
raise NotImplementedError
def inverse_transform(self, X):
raise NotImplementedError
class Identity(Transformer):
"""Identity transform."""
def transform(self, X):
return X
def inverse_transform(self, Xt):
return Xt
class StringEncoder(Transformer):
"""StringEncoder transform.
The transform will cast everything to a string and the inverse
transform will cast to the type defined in dtype.
"""
def __init__(self, dtype=str):
super().__init__()
self.dtype = dtype
def fit(self, X):
"""Fit a list or array of categories. All elements must be from the same type.
Parameters
----------
X : array-like, shape=(n_categories,)
List of categories.
"""
if len(X) > 0:
self.dtype = type(X[0])
def transform(self, X):
"""Transform an array of categories to a string encoded representation.
Parameters
----------
X : array-like, shape=(n_samples,)
List of categories.
Returns
-------
Xt : array-like, shape=(n_samples,)
The string encoded categories.
"""
return [str(x) for x in X]
def inverse_transform(self, Xt):
"""Inverse transform string encoded categories back to their original
representation.
Parameters
----------
Xt : array-like, shape=(n_samples,)
String encoded categories.
Returns
-------
X : array-like, shape=(n_samples,)
The original categories.
"""
return [self.dtype(x) for x in Xt]
class LogN(Transformer):
"""Base N logarithm transform."""
def __init__(self, base):
self._base = base
def transform(self, X):
return np.log10(np.asarray(X, dtype=float)) / np.log10(self._base)
def inverse_transform(self, Xt):
return self._base ** np.asarray(Xt, dtype=float)
class CategoricalEncoder(Transformer):
"""OneHotEncoder that can handle categorical variables."""
def __init__(self):
"""Convert labeled categories into one-hot encoded features."""
self._lb = LabelBinarizer()
def fit(self, X):
"""Fit a list or array of categories.
Parameters
----------
X : array-like, shape=(n_categories,)
List of categories.
"""
self.mapping_ = {v: i for i, v in enumerate(X)}
self.inverse_mapping_ = {i: v for v, i in self.mapping_.items()}
self._lb.fit([self.mapping_[v] for v in X])
self.n_classes = len(self._lb.classes_)
return self
def transform(self, X):
"""Transform an array of categories to a one-hot encoded representation.
Parameters
----------
X : array-like, shape=(n_samples,)
List of categories.
Returns
-------
Xt : array-like, shape=(n_samples, n_categories)
The one-hot encoded categories.
"""
return self._lb.transform([self.mapping_[v] for v in X])
def inverse_transform(self, Xt):
"""Inverse transform one-hot encoded categories back to their original
representation.
Parameters
----------
Xt : array-like, shape=(n_samples, n_categories)
One-hot encoded categories.
Returns
-------
X : array-like, shape=(n_samples,)
The original categories.
"""
Xt = np.asarray(Xt)
return [self.inverse_mapping_[i] for i in self._lb.inverse_transform(Xt)]
class LabelEncoder(Transformer):
"""LabelEncoder that can handle categorical variables."""
def __init__(self, X=None):
if X is not None:
self.fit(X)
def fit(self, X):
"""Fit a list or array of categories.
Parameters
----------
X : array-like, shape=(n_categories,)
List of categories.
"""
X = np.asarray(X)
if X.dtype == object:
self.mapping_ = {v: i for i, v in enumerate(X)}
else:
i = 0
self.mapping_ = {}
_, indexes = np.unique(X, return_index=True)
for index in sorted(indexes):
self.mapping_[X[index]] = i
i += 1
self.inverse_mapping_ = {i: v for v, i in self.mapping_.items()}
return self
def transform(self, X):
"""Transform an array of categories to a one-hot encoded representation.
Parameters
----------
X : array-like, shape=(n_samples,)
List of categories.
Returns
-------
Xt : array-like, shape=(n_samples, n_categories)
The integer categories.
"""
X = np.asarray(X)
return [self.mapping_[v] for v in X]
def inverse_transform(self, Xt):
"""Inverse transform integer categories back to their original representation.
Parameters
----------
Xt : array-like, shape=(n_samples, n_categories)
Integer categories.
Returns
-------
X : array-like, shape=(n_samples,)
The original categories.
"""
if isinstance(Xt, (float, np.float64)):
Xt = [Xt]
else:
Xt = np.asarray(Xt)
return [self.inverse_mapping_[int(np.round(i))] for i in Xt]
class Normalize(Transformer):
"""Scales each dimension into the interval [0, 1].
Parameters
----------
low : float
Lower bound.
high : float
Higher bound.
is_int : bool, default=False
Round and cast the return value of `inverse_transform` to integer. Set
to `True` when applying this transform to integers.
"""
def __init__(self, low, high, is_int=False):
self.low = float(low)
self.high = float(high)
self.is_int = is_int
self._eps = 1e-8
def transform(self, X):
X = np.asarray(X)
if self.is_int:
if np.any(np.round(X) > self.high):
raise ValueError(
"All integer values should" "be less than %f" % self.high
)
if np.any(np.round(X) < self.low):
raise ValueError(
"All integer values should" "be greater than %f" % self.low
)
else:
if np.any(X > self.high + self._eps):
raise ValueError("All values should" "be less than %f" % self.high)
if np.any(X < self.low - self._eps):
raise ValueError("All values should" "be greater than %f" % self.low)
if (self.high - self.low) == 0.0:
return X * 0.0
if self.is_int:
return (np.round(X).astype(int) - self.low) / (self.high - self.low)
else:
return (X - self.low) / (self.high - self.low)
def inverse_transform(self, X):
X = np.asarray(X)
if np.any(X > 1.0 + self._eps):
raise ValueError("All values should be less than 1.0")
if np.any(X < 0.0 - self._eps):
raise ValueError("All values should be greater than 0.0")
X_orig = X * (self.high - self.low) + self.low
if self.is_int:
return np.round(X_orig).astype(int)
return X_orig
class Pipeline(Transformer):
"""A lightweight pipeline to chain transformers.
Parameters
----------
transformers : list
A list of Transformer instances.
"""
def __init__(self, transformers):
self.transformers = list(transformers)
for transformer in self.transformers:
if not isinstance(transformer, Transformer):
raise ValueError(
"Provided transformers should be a Transformer "
"instance. Got %s" % transformer
)
def fit(self, X):
for transformer in self.transformers:
transformer.fit(X)
return self
def transform(self, X):
for transformer in self.transformers:
X = transformer.transform(X)
return X
def inverse_transform(self, X):
for transformer in self.transformers[::-1]:
X = transformer.inverse_transform(X)
return X
|
