/usr/share/pyshared/brian/utils/circular.py is in python-brian 1.3.1-1build1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
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 | # ----------------------------------------------------------------------------------
# Copyright ENS, INRIA, CNRS
# Contributors: Romain Brette (brette@di.ens.fr) and Dan Goodman (goodman@di.ens.fr)
#
# Brian is a computer program whose purpose is to simulate models
# of biological neural networks.
#
# This software is governed by the CeCILL license under French law and
# abiding by the rules of distribution of free software. You can use,
# modify and/ or redistribute the software under the terms of the CeCILL
# license as circulated by CEA, CNRS and INRIA at the following URL
# "http://www.cecill.info".
#
# As a counterpart to the access to the source code and rights to copy,
# modify and redistribute granted by the license, users are provided only
# with a limited warranty and the software's author, the holder of the
# economic rights, and the successive licensors have only limited
# liability.
#
# In this respect, the user's attention is drawn to the risks associated
# with loading, using, modifying and/or developing or reproducing the
# software by the user in light of its specific status of free software,
# that may mean that it is complicated to manipulate, and that also
# therefore means that it is reserved for developers and experienced
# professionals having in-depth computer knowledge. Users are therefore
# encouraged to load and test the software's suitability as regards their
# requirements in conditions enabling the security of their systems and/or
# data to be ensured and, more generally, to use and operate it in the
# same conditions as regards security.
#
# The fact that you are presently reading this means that you have had
# knowledge of the CeCILL license and that you accept its terms.
# ----------------------------------------------------------------------------------
#
'''
Circular arrays
Ideas for speed improvements: use put, putmask and take with mode='wrap' and out=...
'''
from numpy import *
from scipy import weave
import bisect
import os
import warnings
from ..globalprefs import get_global_preference
__all__ = ['CircularVector', 'SpikeContainer']
class CircularVector(object):
'''
A vector with circular structure.
Variables:
* X = the data (array of size n)
* cursor = current position in the array (where the 0 index is)
'''
def __init__(self, n, dtype=float, useweave=False, compiler=None): # pylint: disable-msg=W0621
'''
n is the size of the vector.
'''
self.X = zeros(n, dtype=dtype)
self.dtype = dtype
self.cursor = 0
self.n = n
self._useweave = useweave
if useweave:
self._optimisedreturnarray = zeros(n, dtype=dtype)
self._cpp_compiler = compiler
self._extra_compile_args = ['-O3']
if self._cpp_compiler == 'gcc':
self._extra_compile_args += get_global_preference('gcc_options') # ['-march=native', '-ffast-math']
else:
self._cpp_compiler = ''
def reinit(self):
self.X[:] = zeros(self.n, self.dtype)
self.cursor = 0
def advance(self, k):
self.cursor = (self.cursor + k) % self.n
def __len__(self):
return self.n
def __getitem__(self, i):
'''
V[i]
'''
return self.X[(self.cursor + i) % self.n]
def __setitem__(self, i, x):
'''
V[i]=x
'''
self.X[(self.cursor + i) % self.n] = x
def __getslice__(self, i, j):
n = self.n
i0 = (self.cursor + i) % n # pylint: disable-msg=W0621
j0 = (self.cursor + j) % n # pylint: disable-msg=W0621
if j0 >= i0:
return self.X[i0:j0]
else:
#return self.X[range(i0,n)+range(0,j0)]
return concatenate((self.X[i0:], self.X[0:j0])) # this version is MUCH faster
def get_conditional(self, i, j, min, max, offset=0):
"""
Returns only those vectors with values between min and max
This rather specialised usage of a circular vector is for the
benefit of the SpikeContainer class get_spikes method, which is
in turn used by the NeuronGroup.get_spikes method.
It returns v-offset for those elements v in self[i:j] such that min<v<max.
"""
if self._useweave:
n = self.n
i0 = int((self.cursor + i) % n) # pylint: disable-msg=W0612,W0621
j0 = int((self.cursor + j) % n) # pylint: disable-msg=W0612
X = self.X # pylint: disable-msg=W0612
ret = self._optimisedreturnarray
code = """
int numgot = 0;
for(int k=i0;k!=j0;k=(k+1)%n)
{
int Xk = X(k);
if(Xk>=min && Xk<max)
ret(numgot++)=Xk-offset;
}
return_val = numgot;
"""
numgot = weave.inline(code, ['n', 'i0', 'j0', 'X', 'ret', 'offset', 'min', 'max'],
compiler=self._cpp_compiler,
type_converters=weave.converters.blitz,
extra_compile_args=self._extra_compile_args)
return ret[0:numgot]
else:
spikes = self[i:j]
spikes = spikes[bisect.bisect_left(spikes, min):bisect.bisect_left(spikes, max)]
if offset: spikes = spikes - offset
return spikes
def __setslice__(self, i, j, W):
# NB: S[4:1] does not give a circular slice but only []
# Should we change that behaviour?
# TODO: speed improvements
if j > i:
n = self.n
i0 = (self.cursor + i) % n # pylint: disable-msg=W0621
j0 = (self.cursor + j) % n
if j0 > i0:
self.X[i0:j0] = W
elif isinstance(W, ndarray):
self.X[i0:] = W[0:n - i0]
self.X[0:j0] = W[n - i0:n - i0 + j0]
def __repr__(self):
return repr(hstack((self[0:self.n - 1], self[self.n - 1:self.n])))
def __print__(self):
return (hstack((self[0:self.n - 1], self[self.n - 1:self.n]))).__print__()
class SpikeContainer(object):
'''
An object that stores previous spikes.
S[0] is an array of the last spikes (neuron indexes).
S[1] is an array with the spikes at time t-dt, etc.
S[0:50] contains all spikes in last 50 bins.
'''
def __init__(self, m, useweave=False, compiler=None):
'''
n = maximum number of spikes stored (not used anymore)
m = maximum number of bins stored
'''
if m < 2: m = 2
self.S = CircularVector(2, dtype=int, useweave=useweave, compiler=compiler)
self.ind = CircularVector(m + 1, dtype=int, useweave=useweave, compiler=compiler) # indexes of bins
self.remaining_space = 1
self._useweave = useweave
def reinit(self):
self.S.reinit()
self.ind.reinit()
def push(self, spikes):
'''
Stores spikes in the array at time dt.
'''
ns = len(spikes)
self.remaining_space += (self.ind[2] - self.ind[1]) % self.S.n
while ns >= self.remaining_space:
# double size of array
S = self.S
newS = CircularVector(2 * S.n, dtype=int, useweave=S._useweave, compiler=S._cpp_compiler)
newS.X[:S.n - S.cursor] = S.X[S.cursor:]
newS.X[S.n - S.cursor:S.n] = S.X[:S.cursor]
newS.cursor = S.n
self.S = newS
self.ind.X = (self.ind.X - S.cursor) % S.n
self.ind.X[self.ind.X == 0] = S.n
self.remaining_space += S.n
self.S[0:ns] = spikes
self.S.advance(ns)
self.ind.advance(1)
self.ind[0] = self.S.cursor
self.remaining_space -= ns
def lastspikes(self):
'''
Returns S[0].
'''
return self.S[self.ind[-1] - self.S.cursor:self.S.n]
def __getitem__(self, i):
'''
S[i]: returns the spikes at time t-i*dt.
'''
# NB: this could be optimized
return self.S[self.ind[-i - 1] - self.S.cursor:self.ind[-i] - self.S.cursor + self.S.n]
# optimised version of the above, but the speed improvement is not very much, might be
# better to just wait and write a fully C/C++ version of the whole library
def get_spikes(self, delay, origin, N):
"""
Returns those spikes in self[delay] between origin and origin+N
"""
return self.S.get_conditional(self.ind[-delay - 1] - self.S.cursor, \
self.ind[-delay] - self.S.cursor + self.S.n, \
origin, origin + N, origin)
def __getslice__(self, i, j):
return self.S[self.ind[-j] - self.S.cursor:self.ind[-i] - self.S.cursor + self.S.n]
def __repr__(self):
return "Spike container."
def __print__(self):
return self.__repr__()
try:
import ccircular.ccircular as _ccircular
class SpikeContainer(_ccircular.SpikeContainer):
def __init__(self, m, useweave=False, compiler=None):
_ccircular.SpikeContainer.__init__(self, m)
#warnings.warn('Using C++ SpikeContainer')
except ImportError:
pass
# I am not sure that class below is useful!
class ModInt(object):
'''
A number in Z/nZ, i.e., modulo n.
Variables:
* x = number modulo n
* n
Implemented: additions and subtraction.
N.B.: not a very useful class.
'''
def __init__(self, x, n):
self.x = x
self.n = n
def __add__(self, m):
if isinstance(m, ModInt):
assert self.n == m.n
return ModInt((self.x + m.x) % self.n, self.n)
else:
return ModInt((self.x + m) % self.n, self.n)
def __radd__(self, m):
if isinstance(m, ModInt):
assert self.n == m.n
return ModInt((self.x + m.x) % self.n, self.n)
else:
return ModInt((self.x + m) % self.n, self.n)
def __sub__(self, m):
if isinstance(m, ModInt):
assert self.n == m.n
return ModInt((self.x - m.x) % self.n, self.n)
else:
return ModInt((self.x - m) % self.n, self.n)
def __repr__(self):
return str(self.x) + ' [' + str(self.n) + ']'
def __print__(self):
return self.__repr__()
|