-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathfunc.py
312 lines (241 loc) · 7.6 KB
/
func.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
## Automatically adapted for numpy Jun 08, 2006 by convertcode.py
#
# func.py: general function objects
# Author: Johann Hibschman <[email protected]>
#
# Copyright (C) Johann Hibschman 1997
#
# Enhanced for use with Numeric functions (ufuncs)
# All of the functions are combined with Numeric ufuncs; this loses
# some performance when the functions are used on scalar arguments,
# but should give a big win on vectors.
import numpy as Numeric
#from Numeric import *
from numpy import *
import operator
import math
from types import *
ArrayType = type(asarray(1.0))
UfuncType = type(Numeric.add)
# unary function objects (maybe rename to UN_FUNC?)
class FuncOps:
"""
Common mix-in operations for function objects.
Normal function classes are assumed to implement a call routine,
which will be chained to in the __call__ method.
"""
def compose(self, f):
return UnCompose(self, f)
def __add__(self, f):
return BinCompose(Numeric.add, self, f)
def __sub__(self, f):
return BinCompose(Numeric.subtract, self, f)
def __mul__(self, f):
return BinCompose(Numeric.multiply, self, f)
def __div__(self, f):
return BinCompose(Numeric.divide, self, f)
def __neg__(self):
return UnCompose(Numeric.negative, self)
def __pow__(self, f):
return BinCompose(Numeric.power, self, f)
def __coerce__(self, x):
#if type(x) in [IntType, FloatType, LongType, ComplexType]:
if type(x) in [int, float, long, complex]:
return (self, UnConstant(x))
else:
return (self, x)
def __call__(self, arg):
"Default call routine, used for ordinary functions."
if type(arg) == ArrayType:
return array_map(self.call, arg)
else:
return self.call(arg)
def exp(self):
return UnCompose(Numeric.exp, self)
def log(self):
return UnCompose(Numeric.log, self)
# Bind a normal function
# Should check if the argument is a function.
class FuncBinder(FuncOps):
def __init__(self, a_f):
if ((type(a_f) == UfuncType)
or
(type(a_f) == InstanceType and
FuncOps in a_f.__class__.__bases__)):
self.__call__ = a_f # overwrite the existing call method
self.call = a_f
# wrap a constant in a Function class
class UnConstant(FuncOps):
def __init__(self, x):
self.constant = x
def __call__(self, x):
return self.constant
# just return the argument: f(x) = x
# This is used to build up more complex expressions.
class Identity(FuncOps):
def __init__(self):
pass
def __call__(self, arg):
return arg
# compose two unary functions
class UnCompose(FuncOps):
def __init__(self, a_f, a_g):
self.f = a_f
self.g = a_g
def __call__(self, arg):
return self.f(self.g(arg))
# -------------------------------------------------
# binary function objects
# classes of composition:
# a,b,c,d: binary functions m,n,o: unary functions
# d=c.compose(a,b) - c(a(x,y),b(x,y)) - used for a/b, a*b, etc.
# m=c.compose(n,o) - c(n(x), o(x))
# d=c.compose(n,o) - c(n(x), o(y))
# d=m.compose(c) - m(c(x,y))
class BinFuncOps:
# returns self(f(x), g(x)), a unary function
def compose(self, f, g):
return BinCompose(self, f, g)
# returns self(f(x), g(y)), a binary function
def compose2(self, f, g):
return BinUnCompose(self, f, g)
# returns f(self(x,y)), a binary function
def compose_by(self, f):
return UnBinCompose(f, self)
def __add__(self, f):
return BinBinCompose(operator.add, self, f)
def __sub__(self, f):
return BinBinCompose(operator.sub, self, f)
def __mul__(self, f):
return BinBinCompose(operator.mul, self, f)
def __div__(self, f):
return BinBinCompose(operator.div, self, f)
def __pow__(self, f):
return BinBinCompose(pow, self, f)
def __neg__(self):
return UnBinCompose(operator.neg, self)
def reduce(self, a, axis=0):
result = take(a, [0], axis)
for i in range(1, a.shape[axis]):
result = self(result, take(a, [i], axis))
return result
def accumulate(self, a, axis=0):
n = len(a.shape)
sum = take(a, [0], axis)
out = zeros(a.shape, a.dtype.char)
for i in range(1, a.shape[axis]):
out[all_but_axis(i, axis, n)] = self(sum, take(a, [i], axis))
return out
def outer(self, a, b):
n_a = len(a.shape)
n_b = len(b.shape)
a2 = reshape(a, a.shape + (1,)*n_b)
b2 = reshape(b, (1,)*n_a + b.shape)
# duplicate each array in the appropriate directions
a3 = a2
for i in range(n_b):
a3 = repeat(a3, (b.shape[i],), n_a+i)
b3 = b2
for i in range(n_a):
b3 = repeat(b3, (a.shape[i],), i)
answer = array_map_2(self, a3, b3)
return answer
def all_but_axis(i, axis, num_axes):
"""
Return a slice covering all combinations with coordinate i along
axis. (Effectively the hyperplane perpendicular to axis at i.)
"""
the_slice = ()
for j in range(num_axes):
if j == axis:
the_slice = the_slice + (i,)
else:
the_slice = the_slice + (slice(None),)
return the_slice
# bind a binary function
class BinFuncBinder(BinFuncOps):
def __init__(self, a_f):
self.f = a_f
def __call__(self, arg1, arg2):
return self.f(arg1, arg2)
# bind single variables
class BinVar1(BinFuncOps):
def __init__(self):
pass
def __call__(self, arg1, arg2):
return arg1
class BinVar2(BinFuncOps):
def __init__(self):
pass
def __call__(self, arg1, arg2):
return arg2
# bind individual variables within a binary function
class Bind1st(FuncOps):
def __init__(self, a_f, an_arg1):
self.f = a_f
self.arg1 = an_arg1
def __call__(self, x):
return self.f(self.arg1, x)
class Bind2nd(FuncOps):
def __init__(self, a_f, an_arg2):
self.f = a_f
self.arg2 = an_arg2
def __call__(self, x):
return self.f(x, self.arg2)
# compose binary function with two unary functions (=> unary fcn)
# i.e. given a(x,y), b(x), c(x), : d(x) = a(b(x),c(x))
# (what about e(x,y) = a(b(x), c(y)?)
class BinCompose(FuncOps):
def __init__(self, a_binop, a_f, a_g):
self.binop = a_binop
self.f = a_f
self.g = a_g
self.temp = lambda x, op=a_binop, f=a_f, g=a_g: op(f(x),g(x))
def __call__(self, arg):
# return self.binop(self.f(arg), self.g(arg))
return self.temp(arg)
# compose a unary function with a binary function to get a binary
# function: f(g(x,y))
class UnBinCompose(BinFuncOps):
def __init__(self, a_f, a_g):
self.f = a_f
self.g = a_g
def __call__(self, arg1, arg2):
return self.f(self.g(arg1, arg2))
# compose a two unary functions with a binary function to get a binary
# function: f(g(x), h(y))
class BinUnCompose(BinFuncOps):
def __init__(self, a_f, a_g, a_h):
self.f = a_f
self.g = a_g
self.h = a_h
def __call__(self, arg1, arg2):
return self.f(self.g(arg1), self.h(arg2))
# compose two binary functions together, using a third binary function
# to make the composition: h(f(x,y), g(x,y))
class BinBinCompose(BinFuncOps):
def __init__(self, a_h, a_f, a_g):
self.f = a_f
self.g = a_g
self.h = a_h
def __call__(self, arg1, arg2):
return self.h(self.f(arg1, arg2), self.g(arg1, arg2))
# ----------------------------------------------------
# Array mapping routines
def array_map(f, ar):
"Apply an ordinary function to all values in an array."
flat_ar = ravel(ar)
out = zeros(len(flat_ar), flat_ar.dtype.char)
for i in xrange(len(flat_ar)):
out[i] = f(flat_ar[i])
out.shape = ar.shape
return out
def array_map_2(f, a, b):
if a.shape != b.shape:
raise ShapeError
flat_a = ravel(a)
flat_b = ravel(b)
out = zeros(len(flat_a), a.dtype.char)
for i in xrange(len(flat_a)):
out[i] = f(flat_a[i], flat_b[i])
return reshape(out, a.shape)