Circle-Packings/fractal_dimension/mcmullen_stuff/fancymcmullen.py
2021-07-16 00:59:31 -04:00

140 lines
No EOL
4.2 KiB
Python

from node import Node
import math
import numpy as np
import time
from scipy.sparse import csr_matrix
from scipy import stats
def functions(n, z):
if n == 0:
return np.conj((2-math.sqrt(3))**2 / z)
elif n ==1:
return 2j + np.conj(3 / (z - 2j))
elif n == 2:
return -1 * math.sqrt(3) - 1j + np.conj(3 / (z + math.sqrt(3) + 1j))
elif n == 3:
return math.sqrt(3) - 1j + np.conj(3 / (z - math.sqrt(3) + 1j))
def derivatives(n, z):
if n == 0:
return abs(z**2 / (2-math.sqrt(3))**2)
elif n == 1:
return abs((z-2j)**2 / 3)
elif n == 2:
return abs((z+math.sqrt(3)+1j)**2 / 3)
elif n == 3:
return abs((z-math.sqrt(3)+1j)**2 / 3)
def samplePoint(word):
if word[-1] == 0:
p = 0 + 0j
elif word[-1] == 1:
p = 0 + 0.5j
elif word[-1] == 2:
p = -1 * math.sqrt(3) / 4 - 0.25j
elif word[-1] == 3:
p = math.sqrt(3) / 4 - 0.25j
for letter in word[-2::-1]:
p = functions(letter, p)
return p
def sampleValue(word):
return derivatives(word[0], samplePoint(word))
def generateTree(words, dc):
generators = [np.array([[-1,2,2,2],[0,1,0,0],[0,0,1,0],[0,0,0,1]]),
np.array([[1,0,0,0],[2,-1,2,2],[0,0,1,0],[0,0,0,1]]),
np.array([[1,0,0,0],[0,1,0,0],[2,2,-1,2],[0,0,0,1]]),
np.array([[1,0,0,0],[0,1,0,0],[0,0,1,0],[2,2,2,-1]])]
root = Node([-1., 2. + math.sqrt(3), 2. + math.sqrt(3), 2. + math.sqrt(3)], [], words, False)
current_leaves = [root]
nodes = 1
while True:
new_leaves = []
for leaf in current_leaves:
next_gen = leaf.next_generation(words, dc, generators)
new_leaves += next_gen
nodes += len(next_gen) if len(next_gen) > 1 else 0
if current_leaves == new_leaves:
break
else:
current_leaves = new_leaves
for i,leaf in enumerate(current_leaves):
words[str(leaf.word)] = i
print(len(current_leaves), "partitions")
print(nodes,"nodes")
return current_leaves
def constructMatrix(words, dc):
leave = generateTree(words, dc)
row = []
col = []
data = []
for i,leaf in enumerate(leave):
thing = words[str(leaf.word[1:])]
if isinstance(thing,int):
row.append(i)
col.append(thing)
data.append(sampleValue(leaf.word))
else:
for wor in thing.leaves():
row.append(i)
col.append(words[str(wor.word)])
data.append(sampleValue(leaf.word))
return csr_matrix((data,(row,col)),shape=(len(leave),len(leave)))
def secant(x0,y0,x1,y1,z):
return x0 - (y0-z) * ((x1-x0)/(y1-y0))
def matrixFunction(matrix,l,a):
matrix = matrix.power(a)
vec = np.ones(l)
previous_entry = vec[0]
previous_val = 0
current = matrix * vec
current_val = current[0] / previous_entry
count = 0
while count < 1000000000 and abs(current_val - previous_val) > 1e-15:
previous_val = current_val
previous_entry = current[0]
current = matrix * current
current_val = current[0] / previous_entry
count += 1
print("power method:", count)
return current_val
def secantMethod(matrix,l,z,x1,x2,e,its):
k1 = x1
k2 = x2
y1 = matrixFunction(matrix,l,k1)
y2 = matrixFunction(matrix,l,k2)
#y1 = testFunction(k1)
#y2 = testFunction(k2)
count = 1
print(count,k1,y1)
while abs(y1-z)>e and count<its:
k3 = secant(k1,y1,k2,y2,z)
k1 = k2
y1 = y2
k2 = k3
y2 = matrixFunction(matrix,l,k2)
#y2 = testFunction(k2)
count += 1
print(count,k1,y1)
def main():
start = time.time()
words = {}
matrix = constructMatrix(words,400)
#print(matrix.toarray())
print("construction (s): %f\nconstruction (m): %f" % (time.time()-start, (time.time()-start)/60))
#print(csr_matrix.transpose(matrix))
print(csr_matrix.count_nonzero(matrix), (matrix.shape[0])*3)
secantMethod(matrix,matrix.shape[0],1,1.30,1.31,10**(-10),1000)
print("total (s): %f\ntotal (m): %f" % (time.time()-start, (time.time()-start)/60))
main()