-
Notifications
You must be signed in to change notification settings - Fork 0
/
brain.py
244 lines (204 loc) · 7.09 KB
/
brain.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
from __future__ import print_function
import sys
sys.path.append('/Library/Frameworks/Python.framework/Versions/3.5/lib/python3.5/site-packages/')
#TODO: if want other people to use this figure out how to get them this dependency!
import deap.gp as gp
import pygraphviz as pgv
import numpy as np
import vector_operations
import math, random
import uuid
''' TODO: rename module to processor Tree or something like that'''
'''
inputs:(collision event) rel_sphere_pos, rel_parent_pos
constants: ? maybe a constant vector like z up?
output: new_node_position
'''
def scale_protected(array,scalar):
if scalar == 0.0:
return array
return array*scalar
def scale(array,scalar):
return array*scalar
def subtractnz(array_x,array_y): return array_x if (array_x==array_y).all() else array_x-array_y
def mean_vec(x,y):
c = np.column_stack([x,y])
return np.mean(c,axis=1)
def add_scalar(x,y):
return x+y
def mult_scalar(x,y):
return x*y
def if_greater_vec(w,x,y,z):
return if_greater(w,x,y,z)
def if_greater_float(w,x,y,z):
return if_greater(w,x,y,z)
def if_greater(w,x,y,z):
'''
w float
x float
y any type
z any type
'''
return if_else(greater(w,x),y,z)
def greater(x,y):
return x >= y
def if_else(x,y,z):
return y if x else z
def larger(): pass
def unit_vector(vector):
""" Returns the unit vector of the vector. """
if not vector.any():
return vector
return vector / np.linalg.norm(vector)
def angle_between(v1, v2):
""" Returns the angle in radians between vectors 'v1' and 'v2'::
>>> angle_between((1, 0, 0), (0, 1, 0))
1.5707963267948966
>>> angle_between((1, 0, 0), (1, 0, 0))
0.0
>>> angle_between((1, 0, 0), (-1, 0, 0))
3.141592653589793
"""
v1_u = unit_vector(v1)
v2_u = unit_vector(v2)
return np.arccos(np.clip(np.dot(v1_u, v2_u), -1.0, 1.0))
return np.cross(x,y)
def x_comp(v):
return v[0]
def y_comp(v):
return v[1]
def z_comp(v):
return v[2]
def big_pset():
vec = np.ndarray
two_in = [vec ,vec ]
pset = gp.PrimitiveSetTyped("main",two_in,vec )
''' vec '''
pset.addPrimitive(np.dot,two_in,float)
pset.addPrimitive(np.cross,two_in,vec)
pset.addPrimitive(mean_vec,two_in,vec )
pset.addPrimitive(vector_operations.rotate_vec,[vec ,vec ,float],vec )
pset.addPrimitive(scale,[vec ,float],vec )
pset.addPrimitive(np.linalg.norm, [vec ],float)
pset.addPrimitive(angle_between,[vec ,vec ],float)
pset.addPrimitive(unit_vector,[vec],vec)
pset.addPrimitive(if_greater_vec, [float,float,vec ,vec], vec)
pset.addPrimitive(x_comp, [vec],float)
pset.addPrimitive(y_comp, [vec],float)
pset.addPrimitive(z_comp, [vec],float)
'''non vec'''
pset.addPrimitive(add_scalar, [float,float], float)
pset.addPrimitive(if_greater_float, [float, float, float, float ], float)
pset.addPrimitive(mult_scalar, [float, float], float)
pset.addTerminal(.5,float,'c1')
pset.addTerminal(np.array([0.,0.,-1.0]),vec,'downVec')
pset.addEphemeralConstant(str(uuid.uuid1()),lambda: random.uniform(0, math.pi*2.),float)
pset.addEphemeralConstant(str(uuid.uuid1()),lambda: random.uniform(-1,1),float)
pset.renameArguments(ARG0="x")
pset.renameArguments(ARG1="y")
return pset
def make_vec_pset():
'''
primite set designed to take in vectors and scalars
'''
array_type = np.ndarray
#unit_x = np.array((1.0,0.0,0.0))
two_in = [array_type,array_type]
one_out = array_type
pset = gp.PrimitiveSetTyped("main",two_in,array_type)
#pset.addPrimitive(np.add,two_in,array_type)
#pset.addPrimitive(subtractnz,two_in,array_type)
#pset.addPrimitive(np.multiply,two_in,array_type)
pset.addPrimitive(np.dot,two_in,float)
#pset.addPrimitive(np.maximum,two_in,array_type)
#pset.addPrimitive(np.minimum,two_in,array_type)
pset.addPrimitive(mean_vec,two_in,array_type)
pset.addPrimitive(vector_operations.rotate_vec,[array_type,array_type,float],array_type)
pset.addPrimitive(scale,[array_type,float],array_type)
pset.addPrimitive(np.linalg.norm, [array_type],float)
#pset.addPrimitive(np.divide,two_in,array_type) gets divide by zero errors
#pset.addPrimitive(np.reciprocal,two_in,array_type) gets divide by zero errors
#pset.addTerminal(unit_x,array_type,"unit_x")
pset.addTerminal(.5,float,'c1')
e_name = str(uuid.uuid1())
#pset.addEphemeralConstant(e_name,lambda: random.uniform(0, math.pi*2.),float)
#pset.addEphemeralConstant(str(uuid.uuid1()),lambda: random.uniform(0,1),float)
pset.renameArguments(ARG0="x")
pset.renameArguments(ARG1="y")
return pset
'''
expr = gp.genFull(pset,min_=2,max_=10)
tree = gp.PrimitiveTree(expr)
print(tree)
function = gp.compile(tree,pset)
x = np.array((1.1,0.3,0.2))
y = np.array((2.0,1.9,1.5))
print(function(x,y))
'''
def plot_processor_tree(expression):
'''
creates a pdf image of the processor tree
defined by expression
'''
nodes,edges,labels = gp.graph(expression)
g = pgv.AGraph()
g.add_nodes_from(nodes)
g.add_edges_from(edges)
g.layout(prog="dot")
for i in nodes:
n = g.get_node(i)
l = labels[i]
if type(l) == float:
l = round(l,2)
n.attr["label"] = l
g.draw("cat.pdf")
import matplotlib.pyplot as plt
import networkx
def plot_genealogy_tree(tree, genealogy_history=None):
#NOTE: currently not behaving as desired!
graph = networkx.DiGraph(tree)
graph = graph.reverse() # Make the grah top-down
#need to find way to store fitness for each individual in pop
#colors = [toolbox.evaluate(genealogy_history[i])[0] for i in graph]
#networkx.draw(graph, node_color=colors)
pos = networkx.nx_pydot.graphviz_layout(graph, prog='dot')
networkx.draw(graph,pos,with_labels=True)
#plt.show()
plt.savefig('genealogy_cat.pdf',bbox_inches='tight')
plt.close()
def generate_processor_tree(pset,minDepth,maxDepth):
'''
generates a function that is composed of randomly selected primitives
'''
expr = gp.genGrow(pset,min_=minDepth,max_=maxDepth)
tree = gp.PrimitiveTree(expr)
return tree,gp.compile(tree,pset)
defualt_filename = 'tree.txt'
def load_text(filename):
f = open(filename,'r')
return f.read()
def save_processor_tree(tree,filename=defualt_filename):
'''
saves string representaiton of the processor tree
as a text file
'''
with open(filename, "w") as text_file:
print(str(tree), file=text_file)
def resurrect_processor_tree(pset,tree_string=None):
if tree_string==None:
tree_string = load_text(defualt_filename)
tree = gp.PrimitiveTree.from_string(tree_string,pset)
return tree,gp.compile(tree,pset)
def get_callable(expr,pset):
tree = gp.PrimitiveTree(expr)
return gp.compile(tree,pset)
if __name__=="__main__":
pset = big_pset()
expr = gp.genGrow(pset,min_=3,max_=3)
tree = gp.PrimitiveTree(expr)
processor = gp.compile(tree,pset)
print(tree)
x = np.array((1.1,0.3,0.2))
y = np.array((2.0,1.9,1.5))
print(processor(x,y))
plot_processor_tree(expr)