/
compositions.py
348 lines (290 loc) · 13.9 KB
/
compositions.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
# Copyright (c) 2019. yoshida-lab. All rights reserved.
# Use of this source code is governed by a BSD-style
# license that can be found in the LICENSE file.
import numpy as np
from .base import BaseDescriptor, BaseCompositionFeaturizer
class Counting(BaseCompositionFeaturizer):
def __init__(self, *, one_hot_vec=False, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
one_hot_vec : bool
Set ``true`` to using one-hot-vector encoding.
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
self.one_hot_vec = one_hot_vec
self._elems = self._elements.index.tolist()
self.__authors__ = ['TsumiNa']
def mix_function(self, elems, nums):
vec = np.zeros(len(self._elems), dtype=np.int)
for i, e in enumerate(elems):
if self.one_hot_vec:
vec[self._elems.index(e)] = 1
else:
vec[self._elems.index(e)] = nums[i]
return vec
@property
def feature_labels(self):
return self._elems
class WeightedAverage(BaseCompositionFeaturizer):
def __init__(self, *, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
def mix_function(self, elems, nums):
elems_ = self._elements.loc[elems, :].values
w_ = nums / np.sum(nums)
return w_.dot(elems_)
@property
def feature_labels(self):
return ['ave:' + s for s in self._elements]
class WeightedSum(BaseCompositionFeaturizer):
def __init__(self, *, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
def mix_function(self, elems, nums):
elems_ = self._elements.loc[elems, :].values
w_ = np.array(nums)
return w_.dot(elems_)
@property
def feature_labels(self):
return ['sum:' + s for s in self._elements]
class GeometricMean(BaseCompositionFeaturizer):
def __init__(self, *, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
def mix_function(self, elems, nums):
elems_ = self._elements.loc[elems, :].values
w_ = np.array(nums).reshape(-1, 1)
tmp = elems_ ** w_
return np.power(tmp.prod(axis=0), 1 / sum(w_))
@property
def feature_labels(self):
return ['gmean:' + s for s in self._elements]
class HarmonicMean(BaseCompositionFeaturizer):
def __init__(self, *, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
def mix_function(self, elems, nums):
elems_ = 1 / self._elements.loc[elems, :].values
w_ = np.array(nums)
tmp = w_.dot(elems_)
return sum(w_) / tmp
@property
def feature_labels(self):
return ['hmean:' + s for s in self._elements]
class WeightedVariance(BaseCompositionFeaturizer):
def __init__(self, *, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
def mix_function(self, elems, nums):
elems_ = self._elements.loc[elems, :].values
w_ = nums / np.sum(nums)
mean_ = w_.dot(elems_)
var_ = elems_ - mean_
return w_.dot(var_ ** 2)
@property
def feature_labels(self):
return ['var:' + s for s in self._elements]
class MaxPooling(BaseCompositionFeaturizer):
def __init__(self, *, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
def mix_function(self, elems, _):
elems_ = self._elements.loc[elems, :]
return elems_.max().values
@property
def feature_labels(self):
return ['max:' + s for s in self._elements]
class MinPooling(BaseCompositionFeaturizer):
def __init__(self, *, n_jobs=-1, on_errors='raise', return_type='any'):
"""
Parameters
----------
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'raise' which will raise up the exception.
return_type: str
Specific the return type.
Can be ``any``, ``array`` and ``df``.
``array`` and ``df`` force return type to ``np.ndarray`` and ``pd.DataFrame`` respectively.
If ``any``, the return type dependent on the input type.
Default is ``any``
"""
super().__init__(n_jobs=n_jobs, on_errors=on_errors, return_type=return_type)
def mix_function(self, elems, _):
elems_ = self._elements.loc[elems, :]
return elems_.min().values
@property
def feature_labels(self):
return ['min:' + s for s in self._elements]
class Compositions(BaseDescriptor):
"""
Calculate elemental descriptors from compound's composition.
"""
def __init__(self, *, n_jobs=-1, featurizers='all', on_errors='nan'):
"""
Parameters
----------
elements: panda.DataFrame
Elements information in `pandas.DataFrame` object. indexed by element symbol.
n_jobs: int
The number of jobs to run in parallel for both fit and predict.
Set -1 to use all cpu cores (default).
Inputs ``X`` will be split into some blocks then run on each cpu cores.
featurizers: list[str] or 'all'
Featurizers that will be used.
Default is 'all'.
on_errors: string
How to handle exceptions in feature calculations. Can be 'nan', 'keep', 'raise'.
When 'nan', return a column with ``np.nan``.
The length of column corresponding to the number of feature labs.
When 'keep', return a column with exception objects.
The default is 'nan' which will raise up the exception.
"""
super().__init__(featurizers=featurizers)
self.n_jobs = n_jobs
self.composition = Counting(n_jobs=n_jobs, on_errors=on_errors)
self.composition = WeightedAverage(n_jobs=n_jobs, on_errors=on_errors)
self.composition = WeightedSum(n_jobs=n_jobs, on_errors=on_errors)
self.composition = WeightedVariance(n_jobs=n_jobs, on_errors=on_errors)
self.composition = GeometricMean(n_jobs=n_jobs, on_errors=on_errors)
self.composition = HarmonicMean(n_jobs=n_jobs, on_errors=on_errors)
self.composition = MaxPooling(n_jobs=n_jobs, on_errors=on_errors)
self.composition = MinPooling(n_jobs=n_jobs, on_errors=on_errors)