-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathrencode.py
More file actions
349 lines (304 loc) · 11.8 KB
/
Copy pathrencode.py
File metadata and controls
349 lines (304 loc) · 11.8 KB
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
import logging
from functools import partial
from bitstring import BitStream
import math
import arn
from evodevo import Problem, Agent
from utils import *
from utils.bitstrutils import *
from utils.mathlogic import *
log = logging.getLogger(__name__)
def printdotcircuit(circuit, labels=None):
circuit = circuit.circuit
s = 'digraph best {\nordering = out;\n'
for c in circuit:
s += '%i [label="%s"];\n' % (c[0], c[1])# if not labels
#else labels[c[1]])
for inp in c[2]:
aux = "dir=back"
if inp < 0:
aux += ",style=dotted"
s += '%i -> %i [%s];\n' % (c[0],abs(inp),aux)
s += '}'
return s
#TODO: remove! behavior is identical to nnlikefun
def regressionfun(mapped, node_inputs, inputs ):
if not node_inputs:
return eval(mapped)
mainmod = __import__('__main__')
if mapped in ['if_']:
return getattr(mainmod, mapped)(*node_inputs)
#print r
# return r
#else:
if len(node_inputs) == 1:
return getattr(mainmod, mapped)(node_inputs[0])
return reduce(lambda m,n: getattr(mainmod, mapped)(m,n),
node_inputs)
def nnlikefun(mapped, node_inputs, inputs):
mainmod = __import__('__main__')
if not node_inputs:
try:
return eval(mapped)
except NameError:
return getattr(mainmod, mapped)(inputs)
if len(node_inputs) == 1:
return getattr(mainmod, mapped)(node_inputs[0])
if mapped not in ReNCoDeProb.funs:
return getattr(mainmod, mapped)(*node_inputs)
#print mapped, node_inputs
return reduce(lambda m,n: getattr(mainmod, mapped)(m,n),
node_inputs)
def mergefun(mapped, node_inputs, inputs):
if not node_inputs:
return mapped
else:
result = mapped + '('
result += reduce(lambda m, n: m + ',' + n, node_inputs)
result += ')'
return result
def defaultnodemap(signature, mappingset):
if len(mappingset) < 2:
return mappingset[0]
index = BitArray(bin=applymajority(
signature,
int(math.ceil(math.log(len(mappingset),2)))))
intindex = index.uint
if intindex >= len(mappingset):
intindex -= len(mappingset)
return mappingset[intindex]
def evaluatecircuit(circuit, circuitmap, resultdict, *inputs,**kwargs):
try:
nout = kwargs['nout']
except KeyError:
nout = 1
if resultdict == None:
resultdict = dict()
for i in range(len(circuit)):
inputvalues = [resultdict[abs(v)] for v in circuit[-1-i][2]]
result = circuitmap(circuit[-1-i][1],
inputvalues,
inputs)
resultdict[circuit[-1-i][0]] = result
if nout == 1:
return result
else:
results = list()
for i in range(nout):
if len(circuit) > i:
results.append(resultdict[circuit[i][0]])
else:
results.append(0)
return results
def buildcircuit(agent, problem, **kwargs):
"""Returns the circuit to be fed into the evaluation function"""
arn = agent.genotype
if not arn.promlist:
return []
#arn.simulate()
#orderedps = sorted(arn.proteins, key = lambda x: x[-1], reverse=True)
graph = arn.ebindings - arn.ibindings
cleanpairs(graph)
promlist = [(p[0],
_getinputlist(
arn.promlist,
graph[:,arn.promlist.index(p[0])].tolist()))
for p in arn.proteins]
promlist.sort(key=lambda x: len(x[1]),reverse=True)
circuit = []
pdict = dict(zip(arn.promlist,arn.proteins))
recursivebuild(circuit, problem, pdict, dict(promlist),
graph, [promlist[0][0]], [], [])
return circuit
def recursivebuild(circuit, problem, proteindict, inputdict, graph,
pqueue, secondqueue, blacklist):
if not pqueue:
if not secondqueue: return circuit
else:
secondqueue.sort(key=lambda x: len(inputdict[x]),
reverse=True)
pqueue.append(secondqueue.pop(0))
next = pqueue.pop(0)
pnext = proteindict[next]
blacklist.append(next)
inputs = []
for i in inputdict[next]:
if i in blacklist:
if problem.feedback:
inputs.append(-i)
else:
inputs.append(i)
if inputs:
fun = problem.nodemap_(pnext[4],problem.funs)
ar = problem.arity[fun]
if ar > 0 and len(inputs) > ar:
del inputs[ar:]
else:
fun = problem.nodemap_(pnext[4],problem.terms)
circuit.append((next, fun , inputs))
#update input dict if not using feedback
if not problem.feedback:
for k,v in inputdict.items():
inputdict[k] = filter(
lambda i: i not in blacklist, v)
secondqueue.extend([inp for inp in inputs
if inp not in (pqueue+secondqueue+blacklist)
and (inp >= 0)])
pqueue, secondqueue = _mergequeues(pqueue, secondqueue, inputdict)
pqueue.sort(key=lambda x: len(inputdict[x]), reverse = True)
return recursivebuild(circuit, problem, proteindict, inputdict,
graph, pqueue,secondqueue, blacklist)
def cleanpairs(matrix):
s = len(matrix)
for i in range(s):
for j in range(i):
if matrix[i][j] >= matrix[j][i]:
matrix[j][i] = 0
else:
matrix[i][j] = 0
def _mergequeues(q1, q2,inputdict):
disjunction = q2[:]
demoted = []
for e in q1:
dependent = False
if e in [p
for pq_el in q1
for p in inputdict[pq_el]
if pq_el != e]:
dependent = True
if e in [p
for pq_el in q2
for p in inputdict[pq_el]]:
dependent = True
if dependent:
demoted.append(e)
for e in q2:
dependent = False
if e in [p
for pq_el in q1
for p in inputdict[pq_el]]:
dependent = True
if e in [p
for pq_el in q2
for p in inputdict[pq_el]
if pq_el != e]:
dependent = True
if dependent:
disjunction.remove(e)
for e in demoted: q1.remove(e)
q2.extend(demoted)
for e in disjunction: q2.remove(e)
q1.extend(disjunction)
return q1,q2
def _getinputlist(promlist, weights):
"""Returns the relevant inputs given the weights."""
inputs=[]
pmap = zip(promlist,weights)
for p,w in pmap:
if w > 0:
inputs.append(p)
return inputs
def printcircuit(circuit):
s = ''
for c in circuit:
s += "%i [%s]: %s\n" % (c[0], c[1],
reduce(lambda m,n: "%s %s " % (m,n),c[2],
""))
return s[:-2]
### Problem base to use with ReNCoDe
class ReNCoDeProb(Problem):
#TODO: read fun set from config file
funs = [ 'add_', 'sub_', 'mul_', 'div_']
terms = [ 'inputs[0]' ]
labels = {'add_':'+', 'sub_':'-', 'mul_':'*', 'div_':'/',
'inputs[0]':'x', 'inputs[1]':'1.0'}
arity = {}#{'add_':0, 'sub_':0, 'mul_':0, 'div_':0}
feedback = False
nout=1
def __init__(self, evaluate, nodemap = defaultnodemap):
Problem.__init__(self, evaluate)
self.nodemap_ = nodemap
#callable phenotype
class P:
def __init__(self, arnet, circuit, problem, skeleton = regressionfun):
self.geno = arnet
self.circuit = circuit
self.problem = problem
#memory (disabled by default)
self.memory = None
self.funskel = skeleton
if not problem.feedback:
self._str = compile(evaluatecircuit(self.circuit, mergefun, self.memory, *problem.terms),
'<string>',
'eval')
def getcircuit(self, *args):
return self.circuit
def __call__(self, *inputs):
return evaluatecircuit(self.circuit,self.funskel , self.memory,
*inputs, nout = self.problem.nout)
def __len__(self):
return len(self.circuit)
def __eq__(self,other):
return self.circuit == other.circuit
def printgraph(self):
return "GRAPH NOT AVAILABLE"
### Agent model to use with this CoDe module
class ReNCoDeAgent(Agent):
genotype = None
phenotype = None
fitness = None
def __init__(self, config, problem, gcode = None, parent = None):
Agent.__init__(self, parent)
generator = arn.bindparams(config, self.generate)
if gcode == None:
gcode = generator()
arnet = arn.ARNetwork(gcode,config)
self.genotype = arnet
self.phenotype = P(arnet,buildcircuit(self,problem), problem)
self.problem = problem
self.fitness = 1e6
def __str__(self):
return "### Agent ###\n%s\n%s: %f" % (self.arn,self.phenotype,
self.fitness)
def pickled(self):
return self.print_()
def print_(self):
return printdotcircuit(self.phenotype)
class DMAgent(ReNCoDeAgent):
def __init__(self, config, problem, gcode = None, parent = None):
self.generate = arn.generatechromo
ReNCoDeAgent.__init__(self, config, problem, gcode, parent)
class RndAgent(ReNCoDeAgent):
def __init__(self, config, problem, gcode = None, parent = None):
self.generate = partial(arn.generatechromo_rnd,
genomesize = 32 * pow(2,config.getint('default','initdm')))
ReNCoDeAgent.__init__(self, config, problem, gcode, parent)
###########################################################################
### Test ###
###########################################################################
if __name__ == '__main__':
import ConfigParser
import random
from bitstring import BitStream
from bitstringutils import dm_event
from arn import ARNetwork
from evodevo import Problem
log.setLevel(logging.DEBUG)
cfg = ConfigParser.ConfigParser()
cfg.readfp(open('test.cfg'))
arncfg = ConfigParser.ConfigParser()
arncfg.readfp(open('arnsim.cfg'))
proteins=[]
nump = 0
while nump < 3:
genome = BitStream(float=random.random(), length=32)
for i in range(5):
genome = dm_event(genome,
.02)
arnet = ARNetwork(genome, arncfg)
nump = len(arnet.promlist)
for p in arnet.proteins: print p
prob = Problem(defaultnodemap,regressionfun, None)
circuit = buildcircuit(arnet, prob,False)
_printcircuit(circuit)
print printdotcircuit(circuit, prob.labels)