Develop Reference

How to increase FPS in ursina python - python

I want to create survival games with infinite block terrain(like Minecraft). So i using ursina python game engine, you can see it here
So i using perlin noise to create the terrain with build-in ursina block model. I test for first 25 block and it work pretty good with above 100 FPS, so i start increase to 250 block and more because I want a infinite terrain. But i ran to some problem, when i increase to 100 block or more, my FPS start to decrease below 30 FPS (With i create just one layer).
Here is my code:
#-------------------------------Noise.py(I got on the github)-------------------------
# Copyright (c) 2008, Casey Duncan (casey dot duncan at gmail dot com)
# see LICENSE.txt for details
"""Perlin noise -- pure python implementation"""
__version__ = '$Id: perlin.py 521 2008-12-15 03:03:52Z casey.duncan $'
from math import floor, fmod, sqrt
from random import randint
# 3D Gradient vectors
_GRAD3 = ((1,1,0),(-1,1,0),(1,-1,0),(-1,-1,0),
(1,0,1),(-1,0,1),(1,0,-1),(-1,0,-1),
(0,1,1),(0,-1,1),(0,1,-1),(0,-1,-1),
(1,1,0),(0,-1,1),(-1,1,0),(0,-1,-1),
)
# 4D Gradient vectors
_GRAD4 = ((0,1,1,1), (0,1,1,-1), (0,1,-1,1), (0,1,-1,-1),
(0,-1,1,1), (0,-1,1,-1), (0,-1,-1,1), (0,-1,-1,-1),
(1,0,1,1), (1,0,1,-1), (1,0,-1,1), (1,0,-1,-1),
(-1,0,1,1), (-1,0,1,-1), (-1,0,-1,1), (-1,0,-1,-1),
(1,1,0,1), (1,1,0,-1), (1,-1,0,1), (1,-1,0,-1),
(-1,1,0,1), (-1,1,0,-1), (-1,-1,0,1), (-1,-1,0,-1),
(1,1,1,0), (1,1,-1,0), (1,-1,1,0), (1,-1,-1,0),
(-1,1,1,0), (-1,1,-1,0), (-1,-1,1,0), (-1,-1,-1,0))
# A lookup table to traverse the simplex around a given point in 4D.
# Details can be found where this table is used, in the 4D noise method.
_SIMPLEX = (
(0,1,2,3),(0,1,3,2),(0,0,0,0),(0,2,3,1),(0,0,0,0),(0,0,0,0),(0,0,0,0),(1,2,3,0),
(0,2,1,3),(0,0,0,0),(0,3,1,2),(0,3,2,1),(0,0,0,0),(0,0,0,0),(0,0,0,0),(1,3,2,0),
(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),
(1,2,0,3),(0,0,0,0),(1,3,0,2),(0,0,0,0),(0,0,0,0),(0,0,0,0),(2,3,0,1),(2,3,1,0),
(1,0,2,3),(1,0,3,2),(0,0,0,0),(0,0,0,0),(0,0,0,0),(2,0,3,1),(0,0,0,0),(2,1,3,0),
(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),(0,0,0,0),
(2,0,1,3),(0,0,0,0),(0,0,0,0),(0,0,0,0),(3,0,1,2),(3,0,2,1),(0,0,0,0),(3,1,2,0),
(2,1,0,3),(0,0,0,0),(0,0,0,0),(0,0,0,0),(3,1,0,2),(0,0,0,0),(3,2,0,1),(3,2,1,0))
# Simplex skew constants
_F2 = 0.5 * (sqrt(3.0) - 1.0)
_G2 = (3.0 - sqrt(3.0)) / 6.0
_F3 = 1.0 / 3.0
_G3 = 1.0 / 6.0
class BaseNoise:
"""Noise abstract base class"""
permutation = (151,160,137,91,90,15,
131,13,201,95,96,53,194,233,7,225,140,36,103,30,69,142,8,99,37,240,21,10,23,
190,6,148,247,120,234,75,0,26,197,62,94,252,219,203,117,35,11,32,57,177,33,
88,237,149,56,87,174,20,125,136,171,168,68,175,74,165,71,134,139,48,27,166,
77,146,158,231,83,111,229,122,60,211,133,230,220,105,92,41,55,46,245,40,244,
102,143,54,65,25,63,161,1,216,80,73,209,76,132,187,208,89,18,169,200,196,
135,130,116,188,159,86,164,100,109,198,173,186,3,64,52,217,226,250,124,123,
5,202,38,147,118,126,255,82,85,212,207,206,59,227,47,16,58,17,182,189,28,42,
223,183,170,213,119,248,152,2,44,154,163,70,221,153,101,155,167,43,172,9,
129,22,39,253,9,98,108,110,79,113,224,232,178,185,112,104,218,246,97,228,
251,34,242,193,238,210,144,12,191,179,162,241, 81,51,145,235,249,14,239,107,
49,192,214,31,181,199,106,157,184,84,204,176,115,121,50,45,127,4,150,254,
138,236,205,93,222,114,67,29,24,72,243,141,128,195,78,66,215,61,156,180)
period = len(permutation)
# Double permutation array so we don't need to wrap
permutation = permutation * 2
randint_function = randint
def __init__(self, period=None, permutation_table=None, randint_function=None):
"""Initialize the noise generator. With no arguments, the default
period and permutation table are used (256). The default permutation
table generates the exact same noise pattern each time.
An integer period can be specified, to generate a random permutation
table with period elements. The period determines the (integer)
interval that the noise repeats, which is useful for creating tiled
textures. period should be a power-of-two, though this is not
enforced. Note that the speed of the noise algorithm is indpendent of
the period size, though larger periods mean a larger table, which
consume more memory.
A permutation table consisting of an iterable sequence of whole
numbers can be specified directly. This should have a power-of-two
length. Typical permutation tables are a sequnce of unique integers in
the range [0,period) in random order, though other arrangements could
prove useful, they will not be "pure" simplex noise. The largest
element in the sequence must be no larger than period-1.
period and permutation_table may not be specified together.
A substitute for the method random.randint(a, b) can be chosen. The
method must take two integer parameters a and b and return an integer N
such that a <= N <= b.
"""
if randint_function is not None: # do this before calling randomize()
if not hasattr(randint_function, '__call__'):
raise TypeError(
'randint_function has to be a function')
self.randint_function = randint_function
if period is None:
period = self.period # enforce actually calling randomize()
if period is not None and permutation_table is not None:
raise ValueError(
'Can specify either period or permutation_table, not both')
if period is not None:
self.randomize(period)
elif permutation_table is not None:
self.permutation = tuple(permutation_table) * 2
self.period = len(permutation_table)
def randomize(self, period=None):
"""Randomize the permutation table used by the noise functions. This
makes them generate a different noise pattern for the same inputs.
"""
if period is not None:
self.period = period
perm = list(range(self.period))
perm_right = self.period - 1
for i in list(perm):
j = self.randint_function(0, perm_right)
perm[i], perm[j] = perm[j], perm[i]
self.permutation = tuple(perm) * 2
class SimplexNoise(BaseNoise):
"""Perlin simplex noise generator
Adapted from Stefan Gustavson's Java implementation described here:
http://staffwww.itn.liu.se/~stegu/simplexnoise/simplexnoise.pdf
To summarize:
"In 2001, Ken Perlin presented 'simplex noise', a replacement for his classic
noise algorithm. Classic 'Perlin noise' won him an academy award and has
become an ubiquitous procedural primitive for computer graphics over the
years, but in hindsight it has quite a few limitations. Ken Perlin himself
designed simplex noise specifically to overcome those limitations, and he
spent a lot of good thinking on it. Therefore, it is a better idea than his
original algorithm. A few of the more prominent advantages are:
* Simplex noise has a lower computational complexity and requires fewer
multiplications.
* Simplex noise scales to higher dimensions (4D, 5D and up) with much less
computational cost, the complexity is O(N) for N dimensions instead of
the O(2^N) of classic Noise.
* Simplex noise has no noticeable directional artifacts. Simplex noise has
a well-defined and continuous gradient everywhere that can be computed
quite cheaply.
* Simplex noise is easy to implement in hardware."
"""
def noise2(self, x, y):
"""2D Perlin simplex noise.
Return a floating point value from -1 to 1 for the given x, y coordinate.
The same value is always returned for a given x, y pair unless the
permutation table changes (see randomize above).
"""
# Skew input space to determine which simplex (triangle) we are in
s = (x + y) * _F2
i = floor(x + s)
j = floor(y + s)
t = (i + j) * _G2
x0 = x - (i - t) # "Unskewed" distances from cell origin
y0 = y - (j - t)
if x0 > y0:
i1 = 1; j1 = 0 # Lower triangle, XY order: (0,0)->(1,0)->(1,1)
else:
i1 = 0; j1 = 1 # Upper triangle, YX order: (0,0)->(0,1)->(1,1)
x1 = x0 - i1 + _G2 # Offsets for middle corner in (x,y) unskewed coords
y1 = y0 - j1 + _G2
x2 = x0 + _G2 * 2.0 - 1.0 # Offsets for last corner in (x,y) unskewed coords
y2 = y0 + _G2 * 2.0 - 1.0
# Determine hashed gradient indices of the three simplex corners
perm = self.permutation
ii = int(i) % self.period
jj = int(j) % self.period
gi0 = perm[ii + perm[jj]] % 12
gi1 = perm[ii + i1 + perm[jj + j1]] % 12
gi2 = perm[ii + 1 + perm[jj + 1]] % 12
# Calculate the contribution from the three corners
tt = 0.5 - x0**2 - y0**2
if tt > 0:
g = _GRAD3[gi0]
noise = tt**4 * (g[0] * x0 + g[1] * y0)
else:
noise = 0.0
tt = 0.5 - x1**2 - y1**2
if tt > 0:
g = _GRAD3[gi1]
noise += tt**4 * (g[0] * x1 + g[1] * y1)
tt = 0.5 - x2**2 - y2**2
if tt > 0:
g = _GRAD3[gi2]
noise += tt**4 * (g[0] * x2 + g[1] * y2)
return noise * 70.0 # scale noise to [-1, 1]
def noise3(self, x, y, z):
"""3D Perlin simplex noise.
Return a floating point value from -1 to 1 for the given x, y, z coordinate.
The same value is always returned for a given x, y, z pair unless the
permutation table changes (see randomize above).
"""
# Skew the input space to determine which simplex cell we're in
s = (x + y + z) * _F3
i = floor(x + s)
j = floor(y + s)
k = floor(z + s)
t = (i + j + k) * _G3
x0 = x - (i - t) # "Unskewed" distances from cell origin
y0 = y - (j - t)
z0 = z - (k - t)
# For the 3D case, the simplex shape is a slightly irregular tetrahedron.
# Determine which simplex we are in.
if x0 >= y0:
if y0 >= z0:
i1 = 1; j1 = 0; k1 = 0
i2 = 1; j2 = 1; k2 = 0
elif x0 >= z0:
i1 = 1; j1 = 0; k1 = 0
i2 = 1; j2 = 0; k2 = 1
else:
i1 = 0; j1 = 0; k1 = 1
i2 = 1; j2 = 0; k2 = 1
else: # x0 < y0
if y0 < z0:
i1 = 0; j1 = 0; k1 = 1
i2 = 0; j2 = 1; k2 = 1
elif x0 < z0:
i1 = 0; j1 = 1; k1 = 0
i2 = 0; j2 = 1; k2 = 1
else:
i1 = 0; j1 = 1; k1 = 0
i2 = 1; j2 = 1; k2 = 0
# Offsets for remaining corners
x1 = x0 - i1 + _G3
y1 = y0 - j1 + _G3
z1 = z0 - k1 + _G3
x2 = x0 - i2 + 2.0 * _G3
y2 = y0 - j2 + 2.0 * _G3
z2 = z0 - k2 + 2.0 * _G3
x3 = x0 - 1.0 + 3.0 * _G3
y3 = y0 - 1.0 + 3.0 * _G3
z3 = z0 - 1.0 + 3.0 * _G3
# Calculate the hashed gradient indices of the four simplex corners
perm = self.permutation
ii = int(i) % self.period
jj = int(j) % self.period
kk = int(k) % self.period
gi0 = perm[ii + perm[jj + perm[kk]]] % 12
gi1 = perm[ii + i1 + perm[jj + j1 + perm[kk + k1]]] % 12
gi2 = perm[ii + i2 + perm[jj + j2 + perm[kk + k2]]] % 12
gi3 = perm[ii + 1 + perm[jj + 1 + perm[kk + 1]]] % 12
# Calculate the contribution from the four corners
noise = 0.0
tt = 0.6 - x0**2 - y0**2 - z0**2
if tt > 0:
g = _GRAD3[gi0]
noise = tt**4 * (g[0] * x0 + g[1] * y0 + g[2] * z0)
else:
noise = 0.0
tt = 0.6 - x1**2 - y1**2 - z1**2
if tt > 0:
g = _GRAD3[gi1]
noise += tt**4 * (g[0] * x1 + g[1] * y1 + g[2] * z1)
tt = 0.6 - x2**2 - y2**2 - z2**2
if tt > 0:
g = _GRAD3[gi2]
noise += tt**4 * (g[0] * x2 + g[1] * y2 + g[2] * z2)
tt = 0.6 - x3**2 - y3**2 - z3**2
if tt > 0:
g = _GRAD3[gi3]
noise += tt**4 * (g[0] * x3 + g[1] * y3 + g[2] * z3)
return noise * 32.0
def lerp(t, a, b):
return a + t * (b - a)
def grad3(hash, x, y, z):
g = _GRAD3[hash % 16]
return x*g[0] + y*g[1] + z*g[2]
class TileableNoise(BaseNoise):
"""Tileable implemention of Perlin "improved" noise. This
is based on the reference implementation published here:
http://mrl.nyu.edu/~perlin/noise/
"""
def noise3(self, x, y, z, repeat, base=0.0):
"""Tileable 3D noise.
repeat specifies the integer interval in each dimension
when the noise pattern repeats.
base allows a different texture to be generated for
the same repeat interval.
"""
i = int(fmod(floor(x), repeat))
j = int(fmod(floor(y), repeat))
k = int(fmod(floor(z), repeat))
ii = (i + 1) % repeat
jj = (j + 1) % repeat
kk = (k + 1) % repeat
if base:
i += base; j += base; k += base
ii += base; jj += base; kk += base
x -= floor(x); y -= floor(y); z -= floor(z)
fx = x**3 * (x * (x * 6 - 15) + 10)
fy = y**3 * (y * (y * 6 - 15) + 10)
fz = z**3 * (z * (z * 6 - 15) + 10)
perm = self.permutation
A = perm[i]
AA = perm[A + j]
AB = perm[A + jj]
B = perm[ii]
BA = perm[B + j]
BB = perm[B + jj]
return lerp(fz, lerp(fy, lerp(fx, grad3(perm[AA + k], x, y, z),
grad3(perm[BA + k], x - 1, y, z)),
lerp(fx, grad3(perm[AB + k], x, y - 1, z),
grad3(perm[BB + k], x - 1, y - 1, z))),
lerp(fy, lerp(fx, grad3(perm[AA + kk], x, y, z - 1),
grad3(perm[BA + kk], x - 1, y, z - 1)),
lerp(fx, grad3(perm[AB + kk], x, y - 1, z - 1),
grad3(perm[BB + kk], x - 1, y - 1, z - 1))))
#--------------------------Math.py(For InverseLefp)--------------------------------
def Clamp(t: float, minimum: float, maximum: float):
"""Float result between a min and max values."""
value = t
if t < minimum:
value = minimum
elif t > maximum:
value = maximum
return value
def InverseLefp(a: float, b: float, value: float):
if a != b:
return Clamp((value - a) / (b - a), 0, 1)
return 0
#-----------------------------Game.py(Main code)----------------------
from ursina import *
from ursina.prefabs import *
from ursina.prefabs.first_person_controller import *
from Math import InverseLefp
import Noise
app = Ursina()
#The maximum height of the terrain
maxHeight = 10
#Control the width and height of the map
mapWidth = 10
mapHeight = 10
#A class that create a block
class Voxel(Button):
def __init__(self, position=(0,0,0)):
super().__init__(
parent = scene,
position = position,
model = 'cube',
origin_y = .5,
texture = 'white_cube',
color = color.color(0, 0, random.uniform(.9, 1.0)),
highlight_color = color.lime,
)
#Detect user key input
def input(self, key):
if self.hovered:
if key == 'right mouse down':
#Place block if user right click
voxel = Voxel(position=self.position + mouse.normal)
if key == 'left mouse down':
#Break block if user left click
destroy(self)
if key == 'escape':
#Exit the game if user press the esc key
app.userExit()
#Return perlin noise value between 0 and 1 with x, y position with scale = noiseScale
def GeneratedNoiseMap(y: int, x: int, noiseScale: float):
#Check if the noise scale was invalid or not
if noiseScale <= 0:
noiseScale = 0.001
sampleX = x / noiseScale
sampleY = y / noiseScale
#The Noise.SimplexNoise().noise2 will return the value between -1 and 1
perlinValue = Noise.SimplexNoise().noise2(sampleX, sampleY)
#The InverseLefp will make the value scale to between 0 and 1
perlinValue = InverseLefp(-1, 1, perlinValue)
return perlinValue
for z in range(mapHeight):
for x in range(mapWidth):
#Calculating the height of the block and round it to integer
height = round(GeneratedNoiseMap(z, x, 20) * maxHeight)
#Place the block and make it always below the player
block = Voxel(position=(x, height - maxHeight - 1, z))
#Set the collider of the block
block.collider = 'mesh'
#Character movement
player = FirstPersonController()
#Run the game
app.run()
All file in same folder.
It was working fine but the FPS is very low, so can anyone help?

I'm not able to test this code at the moment but this should serve as a starting point:
level_parent = Entity(model=Mesh(vertices=[], uvs=[]))
for z in range(mapHeight):
for x in range(mapWidth):
height = round(GeneratedNoiseMap(z, x, 20) * maxHeight)
block = Voxel(position=(x, height - maxHeight - 1, z))
level_parent.model.vertices.extend(block.model.vertices)
level_parent.collider = 'mesh' # call this only once after all vertices are set up
For texturing, you might have to add the block.uvs from each block to level_parent.model.uvs as well. Alternatively, call level_parent.model.project_uvs() after setting up the vertices.

On my version of ursina engine (5.0.0) only this code:
`
level_parent = Entity(model=Mesh(vertices=[], uvs=[]))
for z in range(mapHeight):
for x in range(mapWidth):
height = round(GeneratedNoiseMap(z, x, 20) * maxHeight)
block = Voxel(position=(x, height - maxHeight - 1, z))
#level_parent.model.vertices.extend(block.model.vertices)
level_parent.combine().vertices.extend(block.combine().vertices)
level_parent.collider = 'mesh'
`
is working.

Related

I'm trying to compute the continuous function hidden behind the points, but it shows a graph that looks like it actually coutns the points in-between as zeros.
Here's the plot that shows up (100 vectors, red dots - data set, blue plot - my Fourier series):
Here's the python code:
import matplotlib.pyplot as plt
import numpy as np
import math
step = (np.pi * 2) / 5
start = -np.pi
xDiscrete = [start, start + step, start + 2 * step, start + 3 * step, start + 4 * step, np.pi]
yDiscrete = [2.88, 2.98, 3.24, 3.42, 3.57, 3.79]
ak = []
bk = []
a0 = 0
precisionSize = 0.001
n = 100
avgError = 0
def getAN(k):
sum = 0
for ind in range(1, len(yDiscrete)):
sum += yDiscrete[ind] * math.cos(k * xDiscrete[ind])
an = (2.0 / n) * sum
print('a' + str(k) + ' = ' + str(an))
return an
def getBN(k):
sum = 0
for ind in range(1, len(yDiscrete)):
sum += yDiscrete[ind] * math.sin(k * xDiscrete[ind])
bn = (2.0 / n) * sum
print('b' + str(k) + ' = ' + str(bn))
return bn
def getA0():
sum = 0
for ind in range(1, len(yDiscrete)):
sum += yDiscrete[ind]
a0 = (2.0 / n) * sum
print('a0 = ' + str(a0))
return a0
def getFourierOneSum(x, i):
return ak[i - 1] * math.cos(i * x) + bk[i - 1] * math.sin(i * x)
def getFourierAtPoint(x):
sum = a0 / 2
for i in range(1, n + 1):
sum += getFourierOneSum(x, i)
return sum
for i in range(1, n + 1):
ak.append(getAN(i))
bk.append(getBN(i))
a0 = getA0()
x2 = np.arange(-np.pi, np.pi, precisionSize)
y2 = []
for coor in x2:
y2.append(getFourierAtPoint(coor))
plt.plot(xDiscrete, yDiscrete, 'ro', alpha=0.6)
plt.plot(x2, y2)
plt.grid()
plt.title('Approximation')
plt.show()
I've checked where is the problem, and I'm pretty sure it's with the coefficients (functions getAN, getBN, getA0), but I'm not sure how to fix it.

I'm looking for an efficient python implementation of Somers'D, for which I need to compute the number of concordant, discordant and tied pairs between two random variables X and Y. Two pairs (X_i, Y_i), (X_j, Y_j) are concordant if the ranks of both elements agree; that is, x_i > x_j and y_i > y_j or x_i < x_j and y_i < y_j. Two pairs are called discordant if the ranks of both elements do not agree: x_i > x_j and y_i < y_j or x_i < x_j and y_i > y_j. Two pairs are said to be tied in X (Y) when x_i = x_j y_i = y_j.
Somers'D is then computed as D = (N_C - N_D) / (N_tot - N_Ty).
(See: https://en.wikipedia.org/wiki/Somers%27_D.)
I wrote a naive implementation using nested for-loops. Here, S contains my predictions and Y the realized outcomes.
def concordance_computer(Y, S):
N_C = 0
N_D = 0
N_T_y = 0
N_T_x = 0
for i in range(0, len(S)):
for j in range(i+1, len(Y)):
Y1 = Y[i]
X1 = S[i]
Y2 = Y[j]
X2 = S[j]
if Y1 > Y2 and X1 > X2:
N_C += 1
elif Y1 < Y2 and X1 < X2:
N_C += 1
elif Y1 > Y2 and X1 < X2:
N_D += 1
elif Y1 < Y2 and X1 > X2:
N_D += 1
elif Y1 == Y2:
N_T_y += 1
elif X1 == X2:
N_T_x += 1
N_tot = len(S)*(len(S)-1) / 2
SomersD = (N_C - N_D) / (N_tot - N_T_y)
return SomersD
Obviously, this is gonna be very slow when (Y,S) have a lot of rows. I stumbled upon the use of bisect while searching the net for solutions:
merge['Y'] = Y
merge['S'] = S
zeros2 = merge.loc[merge['Y'] == 0]
ones2 = merge.loc[merge['Y'] == 1]
from bisect import bisect_left, bisect_right
def bin_conc(zeros2, ones2):
zeros2_list = sorted([zeros2.iloc[j, 1] for j in range(len(zeros2))])
zeros2_length = len(zeros2_list)
conc = disc = ties = 0
for i in range(len(ones2)):
cur_conc = bisect_left(zeros2_list, ones2.iloc[i, 1])
cur_ties = bisect_right(zeros2_list, ones2.iloc[i, 1]) - cur_conc
conc += cur_conc
ties += cur_ties
disc += zeros2_length - cur_ties - cur_conc
pairs_tested = zeros2_length * len(ones2.index)
return conc, disc, ties, pairs_tested
This is very efficient, but only works for binary variables Y. Now my question is: how can I implement the concordance_computer in an efficient way for ungrouped Y?

I was able to solve this with some help. A friend pointed me to the fact that there already exists a kendall tau implementation in scipy, which is very efficient. The code can be adapted to create a very fast Somers' D implementation. See code below. On my laptop, it runs in ~60ms, which makes it fast enough to use for e.g. bootstrapping confidence intervals with n_boots ~ O(10^3).
import pandas as pd
import numpy as np
import time
from scipy.stats._stats import _kendall_dis
# import data
df_ = pd.read_csv('Data/CR_mockData_EAD.csv')
df = df_[['realized_ead', 'pred_value']][0:100]
def SomersD(x, y):
x = np.asarray(x).ravel()
y = np.asarray(y).ravel()
if x.size != y.size:
raise ValueError("All inputs must be of the same size, "
"found x-size %s and y-size %s" % (x.size, y.size))
def count_rank_tie(ranks):
cnt = np.bincount(ranks).astype('int64', copy=False)
cnt = cnt[cnt > 1]
return ((cnt * (cnt - 1) // 2).sum(),
(cnt * (cnt - 1.) * (cnt - 2)).sum(),
(cnt * (cnt - 1.) * (2*cnt + 5)).sum())
size = x.size
perm = np.argsort(y) # sort on y and convert y to dense ranks
x, y = x[perm], y[perm]
y = np.r_[True, y[1:] != y[:-1]].cumsum(dtype=np.intp)
# stable sort on x and convert x to dense ranks
perm = np.argsort(x, kind='mergesort')
x, y = x[perm], y[perm]
x = np.r_[True, x[1:] != x[:-1]].cumsum(dtype=np.intp)
dis = _kendall_dis(x, y) # discordant pairs
obs = np.r_[True, (x[1:] != x[:-1]) | (y[1:] != y[:-1]), True]
cnt = np.diff(np.where(obs)[0]).astype('int64', copy=False)
ntie = (cnt * (cnt - 1) // 2).sum() # joint ties
xtie, x0, x1 = count_rank_tie(x) # ties in x, stats
ytie, y0, y1 = count_rank_tie(y) # ties in y, stats
tot = (size * (size - 1)) // 2
# Note that tot = con + dis + (xtie - ntie) + (ytie - ntie) + ntie
# = con + dis + xtie + ytie - ntie
#con_minus_dis = tot - xtie - ytie + ntie - 2 * dis
SD = (tot - xtie - ytie + ntie - 2 * dis) / (tot - ntie)
return (SD, dis)
start_time = time.time()
SD, dis = SomersD(df.realized_ead, df.pred_value)
print("--- %s seconds ---" % (time.time() - start_time))

Current code with "effective" realization produces wrong answers.
For example, check it with x=[1,1,0,1,0] and y=[1,1,0,1,1]
SAS and scipy.stats.somersd provide right answers.

Is it possible to change the formula of the mandelbrot set (which is f(z) = z^2 + c by default) to a different one ( f(z) = z^2 + c * e^(-z) is what i need) when using the escape time algorithm and if possible how?
I'm currently using this code by FB36
# Multi-threaded Mandelbrot Fractal (Do not run using IDLE!)
# FB - 201104306
import threading
from PIL import Image
w = 512 # image width
h = 512 # image height
image = Image.new("RGB", (w, h))
wh = w * h
maxIt = 256 # max number of iterations allowed
# drawing region (xa < xb & ya < yb)
xa = -2.0
xb = 1.0
ya = -1.5
yb = 1.5
xd = xb - xa
yd = yb - ya
numThr = 5 # number of threads to run
# lock = threading.Lock()
class ManFrThread(threading.Thread):
def __init__ (self, k):
self.k = k
threading.Thread.__init__(self)
def run(self):
# each thread only calculates its own share of pixels
for i in range(k, wh, numThr):
kx = i % w
ky = int(i / w)
a = xa + xd * kx / (w - 1.0)
b = ya + yd * ky / (h - 1.0)
x = a
y = b
for kc in range(maxIt):
x0 = x * x - y * y + a
y = 2.0 * x * y + b
x = x0
if x * x + y * y > 4:
# various color palettes can be created here
red = (kc % 8) * 32
green = (16 - kc % 16) * 16
blue = (kc % 16) * 16
# lock.acquire()
global image
image.putpixel((kx, ky), (red, green, blue))
# lock.release()
break
if __name__ == "__main__":
tArr = []
for k in range(numThr): # create all threads
tArr.append(ManFrThread(k))
for k in range(numThr): # start all threads
tArr[k].start()
for k in range(numThr): # wait until all threads finished
tArr[k].join()
image.save("MandelbrotFractal.png", "PNG")

From the code I infer that z = x + y * i and c = a + b * i. That corresponds f(z) - z ^2 + c. You want f(z) = z ^2 + c * e^(-z).
Recall that e^(-z) = e^-(x + yi) = e^(-x) * e^i(-y) = e^(-x)(cos(y) - i*sin(y)) = e^(-x)cos(y) - i (e^(-x)sin(y)). Thus you should update your lines to be the following:
x0 = x * x - y * y + a * exp(-x) * cos(y) + b * exp(-x) * sin(y);
y = 2.0 * x * y + a * exp(-x) * sin(y) - b * exp(-x) * cos(y)
x = x0
You might need to adjust maxIt if you don't get the level of feature differentiation you're after (it might take more or fewer iterations to escape now, on average) but this should be the mathematical expression you're after.
As pointed out in the comments, you might need to adjust the criterion itself and not just the maximum iterations in order to get the desired level of differentiation: changing the max doesn't help for ones that never escape.
You can try deriving a good escape condition or just try out some things and see what you get.

I am computing the mandelbrot set recursively and attempting to perform linear interpolation using the smooth coloring algorithm. However, this returns floating point RGB values which I can't put into the ppm image I am using so I am having to round off using int(), creating a smoother but yet still banded image.
Are there any simpler ways that will produce a better non-banded image?
The second function is an extremely bad hack just playing around with ideas as the smooth algorithim seems to be producing rgb values in the range 256**3
Commented out the linear interpolation I was doing.
Here are my three functions:
def linear_interp(self, color_1, color_2, i):
r = (color_1[0] * (1 - i)) + (color_2[0] * i)
g = (color_1[1] * (1 - i)) + (color_2[1] * i)
b = (color_1[2] * (1 - i)) + (color_2[2] * i)
return (int(abs(r)), int(abs(g)), int(abs(b)))
def mandel(self, x, y, z, iteration = 0):
mod_z = sqrt((z.real * z.real) + (z.imag * z.imag))
#If its not in the set or we have reached the maximum depth
if abs(z) >= 2.00 or iteration == DEPTH:
if iteration == DEPTH:
mu = iteration
else:
mu = iteration + 1 - log(log(mod_z)) / log(2)
else:
mu = 0
z = (z * z) + self.c
self.mandel(x, y, z, iteration + 1)
return mu
def create_image(self):
begin = time.time() #For computing how long it took (start time)
self.rgb.palette = []
for y in range(HEIGHT):
self.rgb.palette.append([]) #Need to create the rows of our ppm
for x in range(WIDTH):
self.c = complex(x * ((self.max_a - self.min_a) / WIDTH) + self.min_a,
y * ((self.max_b - self.min_b) / HEIGHT) + self.min_b)
z = self.c
q = (self.c.real - 0.25)**2 + (self.c.imag * self.c.imag)
x = self.c.real
y2 = self.c.imag * self.c.imag
if not (q*(q + (x - 0.25)) < y2 / 4.0 or (x + 1.0)**2 + y2 <0.0625):
mu = self.mandel(x, y, z, iteration = 0)
rgb = self.linear_interp((255, 255, 0), (55, 55, 0), mu)
self.rgb.palette[y].append(rgb)
else:
self.rgb.palette[y].append((55, 55, 0))
if self.progress_bar != None:
self.progress_bar["value"] = y
self.canvas.update()
The image I am getting is below:

I think this is the culprit:
else:
mu = 0
self.mandel(x, y, z, iteration + 1)
return mu
This isn't passing down the value of mu from the recursive call correctly, so you're getting black for everything that doesn't bottom out after 1 call. Try
else:
...
mu = self.mandel(x, y, z, iteration + 1)
return mu

I'm trying to adapt this noise module to a project I'm working on but I'm not getting the results I was expecting:
https://pypi.python.org/pypi/noise/
Instead of a varied height-map, each row tends to have the exact same value. If I create something 250x250 it will generate the same value 250 times and then generate a new value 250 times until it ends.
Here is the function I'm currently using. I understand this function fairly well but I'm just not sure how to get more "interesting" results. Thank you for your help.
class SimplexNoise(BaseNoise):
def noise2(self, x, y):
"""2D Perlin simplex noise.
Return a floating point value from -1 to 1 for the given x, y coordinate.
The same value is always returned for a given x, y pair unless the
permutation table changes (see randomize above).
"""
# Skew input space to determine which simplex (triangle) we are in
s = (x + y) * _F2
i = floor(x + s)
j = floor(y + s)
t = (i + j) * _G2
x0 = x - (i - t) # "Unskewed" distances from cell origin
y0 = y - (j - t)
if x0 > y0:
i1 = 1; j1 = 0 # Lower triangle, XY order: (0,0)->(1,0)->(1,1)
else:
i1 = 0; j1 = 1 # Upper triangle, YX order: (0,0)->(0,1)->(1,1)
x1 = x0 - i1 + _G2 # Offsets for middle corner in (x,y) unskewed coords
y1 = y0 - j1 + _G2
x2 = x0 + _G2 * 2.0 - 1.0 # Offsets for last corner in (x,y) unskewed coords
y2 = y0 + _G2 * 2.0 - 1.0
# Determine hashed gradient indices of the three simplex corners
perm = BaseNoise.permutation
ii = int(i) % BaseNoise.period
jj = int(j) % BaseNoise.period
gi0 = perm[ii + perm[jj]] % 12
gi1 = perm[ii + i1 + perm[jj + j1]] % 12
gi2 = perm[ii + 1 + perm[jj + 1]] % 12
# Calculate the contribution from the three corners
tt = 0.5 - x0**2 - y0**2
if tt > 0:
g = _GRAD3[gi0]
noise = tt**4 * (g[0] * x0 + g[1] * y0)
else:
noise = 0.0
tt = 0.5 - x1**2 - y1**2
if tt > 0:
g = _GRAD3[gi1]
noise += tt**4 * (g[0] * x1 + g[1] * y1)
tt = 0.5 - x2**2 - y2**2
if tt > 0:
g = _GRAD3[gi2]
noise += tt**4 * (g[0] * x2 + g[1] * y2)
return noise * 70.0 # scale noise to [-1, 1]
win = pygcurse.PygcurseWindow(85, 70, 'Generate')
octaves = 2
ysize = 150
xsize = 150
freq = 32.0 * octaves
for y in range(ysize):
for x in range(xsize):
tile = SimplexNoise.noise2(x / freq, y / freq, octaves)
win.write(str(tile) + "\n")