-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathSparseNDArray.py
355 lines (291 loc) · 13.4 KB
/
SparseNDArray.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
"""
Module for sparse arrays using dictionaries. Inspired in part
by ndsparse (https://launchpad.net/ndsparse) by Pim Schellart
Jan Erik Solem, Feb 9 2010.
[email protected] (bug reports and feedback welcome)
"""
import numpy
class sparray(object):
""" Class for n-dimensional sparse array objects using
Python's dictionary structure.
"""
def __init__(self, shape, default=0, dtype=float):
self.__default = default #default value of non-assigned elements
self.shape = tuple(shape)
self.ndim = len(shape)
self.dtype = dtype
self.__data = {}
def __setitem__(self, index, value):
""" set value to position given in index, where index is a tuple. """
self.__data[index] = value
def __getitem__(self, index):
""" get value at position given in index, where index is a tuple. """
return self.__data.get(index,self.__default)
def __delitem__(self, index):
""" index is tuples of element to be deleted. """
if self.__data.has_key(index):
del(self.__data[index])
def __add__(self, other):
""" Add two arrays. """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] + other.__default
out.__default = self.__default + other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val + other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __sub__(self, other):
""" Subtract two arrays. """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] - other.__default
out.__default = self.__default - other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val - other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __mul__(self, other):
""" Multiply two arrays (element wise). """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] * other.__default
out.__default = self.__default * other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val * other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __div__(self, other):
""" Divide two arrays (element wise).
Type of division is determined by dtype. """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] / other.__default
out.__default = self.__default / other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val / other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __truediv__(self, other):
""" Divide two arrays (element wise).
Type of division is determined by dtype. """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] / other.__default
out.__default = self.__default / other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val / other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __floordiv__(self, other):
""" Floor divide ( // ) two arrays (element wise). """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] // other.__default
out.__default = self.__default // other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val // other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __mod__(self, other):
""" mod of two arrays (element wise). """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] % other.__default
out.__default = self.__default % other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val % other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __pow__(self, other):
""" power (**) of two arrays (element wise). """
if self.shape == other.shape:
out = self.__class__(self.shape, self.dtype)
out.__data = self.__data.copy()
for k in set.difference(set(out.__data.keys()),set(other.__data.keys())):
out.__data[k] = out.__data[k] ** other.__default
out.__default = self.__default ** other.__default
for k in other.__data.keys():
old_val = out.__data.setdefault(k,self.__default)
out.__data[k] = old_val ** other.__data[k]
return out
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __iadd__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] + other.__default
self.__default = self.__default + other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val + other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __isub__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] - other.__default
self.__default = self.__default - other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val - other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __imul__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] * other.__default
self.__default = self.__default * other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val * other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __idiv__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] / other.__default
self.__default = self.__default / other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val / other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __itruediv__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] / other.__default
self.__default = self.__default / other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val / other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __ifloordiv__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] // other.__default
self.__default = self.__default // other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val // other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __imod__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] % other.__default
self.__default = self.__default % other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val % other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __ipow__(self, other):
if self.shape == other.shape:
for k in set.difference(set(self.__data.keys()),set(other.__data.keys())):
self.__data[k] = self.__data[k] ** other.__default
self.__default = self.__default ** other.__default
for k in other.__data.keys():
old_val = self.__data.setdefault(k,self.__default)
self.__data[k] = old_val ** other.__data[k]
return self
else:
raise ValueError('Array sizes do not match. '+str(self.shape)+' versus '+str(other.shape))
def __str__(self):
return str(self.dense())
def dense(self):
""" Convert to dense NumPy array. """
out = self.__default * numpy.ones(self.shape)
for ind in self.__data:
out[ind] = self.__data[ind]
return out
def sum(self):
""" Sum of elements."""
s = self.__default * numpy.array(self.shape).prod()
for ind in self.__data:
s += (self.__data[ind] - self.__default)
return s
if __name__ == "__main__":
#test cases
#create a sparse array
A = sparray((3,3))
print 'shape =', A.shape, 'ndim =', A.ndim
A[(1,1)] = 10
A[2,2] = 10
#access an element
print A[2,2]
print 'remove an element...'
print A
del(A[2,2])
print A
print 'array with different default value...'
B = sparray((3,3),default=3)
print B
print 'adding...'
print A+A
print A+B
print B+B
print 'subtracting...'
print A-A
print A-B
print B-B
print 'multiplication...'
print A*A
print A*B
print B*B
print 'division...'
print A/B
print B/B
print 'mod...'
print B%B
print A%B
print 'power...'
print A**B
print 'iadd...'
A+=B
print A
A+=A
print A
print 'sum of elements...'
print A.sum()
print 'mix with NumPy arrays...'
print A.dense() * numpy.ones((3,3))
print 'Frobenius norm...'
print sum( (A.dense().flatten()-B.dense().flatten())**2 )
print ((A-B)*(A-B)).sum()