def main():
imgpath = "lena.png"
m = "s&p"
cv2.imshow("Image", noise.noise(imgpath, "0"))
cv2.imshow("Image" + m, noise.noise(imgpath, m))
cv2.waitKey(0)
cv2.destroyAllWindows()
def fly(self): s = self s.t += 1 s.w1[1] = -1+u.trapwave(s.t*s.aspd)*3 s.w2[1] = -2+u.trapwave(s.t*s.aspd)*8 s.w3[1] = -1+u.trapwave(s.t*s.aspd+math.pi*0.2)*12 s.w2[0] = -3+math.sin(s.t*s.aspd-math.pi*0.5)*2 s.w3[0] = -5+math.sin(s.t*s.aspd-math.pi*0.5)*3 s.wingCoordToRL(5,s.w1) s.wingCoordToRL(6,s.w2,s.w1,s.skel[5][0]) s.wingCoordToRL(7,s.w3,s.w2,s.skel[6][0],s.skel[5][0]) s.wingCoordToRL(8,s.w1) s.wingCoordToRL(9,s.w2,s.w1,s.skel[5][0]) s.wingCoordToRL(10,s.w3,s.w2,s.skel[6][0],s.skel[5][0]) s.to(4,0,-0+math.sin(s.t*s.aspd+math.pi)*10) s.to(1,0,-5+math.sin(s.t*s.aspd)*1) s.to(12,0,-40) s.to(14,0,-40) s.to(13,0,10+math.sin(s.t*s.aspd)*5 + 10*noise.noise(s.t*s.aspd*0.01,1)-5) s.to(15,0,10+math.sin(s.t*s.aspd)*5 + 10*noise.noise(s.t*s.aspd*0.01,2)-5) #s.x += s.v[0]*s.dir s.y += 0.2*self.s*math.sin(s.t*s.aspd+math.pi)#30.5*s.v[1]+0.5*s.v[1]*(0.5*(math.sin(s.t*s.aspd)+1))
def rest(self): s = self s.t +=1 for i in range(0,len(s.skel)): if i != 6 and i != 1: s.to(i,0,s.ssk[i][0]+noise.noise(i*0.05,s.t*0.05),5) s.to(6,0,30+math.cos(s.t*s.aspd*0.5)*10) noi = max(min((noise.noise(s.t*s.aspd*0.05)-0.4)*40,1),-1) s.to(1,0,-5+noi*50,5) s.to(0,0,-40-noi*40,5)
def walk(self): s = self s.t += 1 s.f1 = 6 + math.cos(s.t*s.aspd*2)*2 +noise.noise(s.t*s.aspd)*2 s.f2 = 8 + math.cos(s.t*s.aspd*2)*2 #print self.animations if self.status[1] == "": for i in range(0,len(s.skel)): for j in range(0,max(0,len(self.animations))): for a in self.animations[j]: if i == a[0][0]: #print "y" return s.to(i,0,s.ssk[i][0]+10*(noise.noise(i*0.05,s.t*0.05)-0.5),2)
def rest(self): s = self #s.t = -1 #math.pi s.t2 += 1 s.to(5,0,20+180*2*((s.skel[5][0]>0)-0.5),10) s.to(6,0,-20+180*2*((s.skel[6][0]>0)-0.5),10) s.to(7,0,20-180*2*((s.skel[10][0]<0)-0.5),10) s.to(8,0,20+180*2*((s.skel[8][0]>0)-0.5),10) s.to(9,0,-20+180*2*((s.skel[9][0]>0)-0.5),10) s.to(10,0,20-180*2*((s.skel[10][0]<0)-0.5),10) s.to(12,0,30,10) s.to(14,0,30,10) s.to(13,0,100,10) s.to(15,0,100,10) noi = max(min((noise.noise(s.t2*s.aspd*0.5)-0.3)*50,1),-1) s.to(1,0,-10+noi*20,5) s.to(1,1,5-noi*2,5) s.to(4,0,-10+noi*10,5) s.x += s.v[0]*s.dir s.y += s.v[1] if s.y>=0: s.v[1] = 0 s.v[0] = 0 s.y = 0 else: s.v[1] += 0.2*s.s if random.random() < 0.02 and s.v[1] == 0: s.v[1]=-1*s.s r = random.choice([1,1]) s.v[0]+=0.5*r*s.s
def rest(self): s = self s.t+=1 noi = noise.noise(s.t*s.aspd*0.5) s.f1 += (6+(noi*5)-s.f1)/2 s.s1 += (1-(noi*4)-s.s1)/2 s.s2 += (1-(noi*4)-s.s2)/2
def generate_noise(width, height):
noise_map = []
points = 256
span = 5.0
speed = 1.0
botom_range = 0
top_range = 0
xoff = BASE_XOFF
for _x in range(width):
new_row = []
xoff = float(_x) * span / points - 0.5 * span
for _y in range(height):
noise_value = noise(xoff + BASE_XOFF)
new_row.append(noise_value)
xoff += random.random()
if noise_value < botom_range:
botom_range = noise_value
elif noise_value > top_range:
top_range = noise_value
noise_map.append(new_row)
difference = float(top_range - botom_range)
for x in range(width):
for y in range(height):
noise_map[x][y] = (noise_map[x][y] - botom_range) / difference
return noise_map
def gen_recover_period(prange, ntimes, npers, randtimes = 0.0, rantnormp = 0.0, snrnoise = 0.0, normpropnoise = 0.0, samples = 100):
"""Generate some periodic data and then try to recover the maximum period under various conditions.
prange = period to analyse (kicking off with a single period) with lower bound and upper bound for search
ntimes number of times to generate signal for
npers number of periods (possibly fractional) to scan for
randtimes whether we vary the sample times 0.0 none, 1 probably max
rantnormp variation in times proortion between uniform (0.0) and gaussian (1.0)
snr signal to noise ration of noise to add 0 for None
normpropnoise proportion of noise uniform (0.0) or gaussian (1.0)
samples is the number of samples for the periodogram
Return fractional error"""
lbound, period, ubound = prange
timelist, amps = siggen.siggen(period, ntimes=ntimes, npers=npers, randv=randtimes, unorm=rantnormp, randphase=True)
if snrnoise > 0.0:
amps = noise.noise(amps, snrnoise, normpropnoise)
# OK now lets do our stuff
model = LombScargle().fit(timelist, amps, 0.001)
periods = np.linspace(lbound, ubound, samples)
pgram = model.periodogram(periods)
# Find maximum
maxes = argmaxmin.maxmaxes(periods, pgram)
if len(maxes) == 0:
return 1.0
maxp = periods[maxes[0]]
return abs(maxp - period) / period
def __init__(self,
seed,
n_state,
n_action,
batch_size=64,
buffer=1e5,
gamma=0.99,
lr_actor=1e-4,
lr_critic=1e-3,
weight_decay=0,
tau=1e-3):
self.batch_size = batch_size
#init actor
self.local_actor = Actor(n_state, n_action, seed).to(device)
self.target_actor = Actor(n_state, n_action, seed).to(device)
self.optim_actor = torch.optim.Adam(self.local_actor.parameters(),
lr=lr_actor)
#init critic
self.local_critic = Critic(n_state, n_action, seed).to(device)
self.target_critic = Critic(n_state, n_action, seed).to(device)
self.optim_critic = torch.optim.Adam(self.local_critic.parameters(),
lr=lr_critic,
weight_decay=weight_decay)
#init memory
self.memory = memory(int(buffer), device, seed)
self.tau = tau
self.gamma = gamma
self.noise = noise(n_action, seed=seed)
def float_noise(*args):
"""Returns a float value which is a (pseudo) random function of its arguments.
This function is imported from the noise module.
"""
la = len(args)
if la==1:
return noise.noise(args[0])
elif la==2:
return noise.noise(args[0],args[1])
elif la==3:
return noise.noise(args[0],args[1],args[2])
elif la==4:
return noise.noise(args[0],args[1],args[2],args[3])
else:
raise TypeError("the function takes between 1 and 4 arguments (%s given)"%(la))
def test_noise(img):
print('\t[Loading...] Calculating..')
start = perf_counter()
result = noise(img, -9, 20)
finish = perf_counter()
cv2.imshow('Add noise', result)
print('\tTime: ' + str(finish - start))
return result
def fill(self):
for y in range(0, Chunk.size):
for x in range(0, Chunk.size):
rand = noise((x + self.x) / 8, (y + self.y) / 8, .1, octaves=2)
if rand < .35 and rand > -.35:
value = 1
else:
value = 0
self.grid[Chunk.size - 1 - y][x] = value
def fill(zone):
noise_f = noise(6)
bg = generate(zone.y, zone.x, noise_f)
for y, row in enumerate(bg):
for x, col in enumerate(row):
val = FILTERS['mountains'](col)
if val > 0.25:
zone.set_field(y, x, '=')
def fill(zone):
noise_f = noise(6)
bg = generate(zone.y, zone.x, noise_f)
for y, row in enumerate(bg):
for x, col in enumerate(row):
val = FILTERS['mountains'](col)
if val > 0.25:
zone.set_field(y, x, '=')
def draw():
global noisy, WIDTH, HEIGHT
set_line_width(1.0)
set_source_rgba(1.0,1.0,1.0,0.2)
rectangle(0,0,WIDTH,HEIGHT)
fill()
for i in range(num):
x, y = pos[i]
# print "Drawing circle at: %s, %s" % (x, y)
set_source_rgba(R1,G1,B1)
arc(x, y, 4, 0, 2 * pi)
stroke()
set_source_rgba(R2,G2,B2,100/255.0)
arc(x, y, 4.5, 0, 2 * pi)
stroke()
set_source_rgba(100/255.0,100/255.0,100/255.0,1.0)
vel[i][0] = 10 * noise(200+x*0.007, 200+y*0.007, noisy*2)*cos(4*pi*noise(x*0.003,y*0.003, noisy))
vel[i][1] = 10 * noise(200+x*0.007, 200+y*0.007, noisy*2)*sin(4*pi*noise(x*0.003,y*0.003, noisy))
'''
# If you want to adjust the acceleration of the circles
for j in range(num):
if j != i:
jx, jy = pos[j]
den1, den2 = dist(x, y, jx, jy), (5+dist(x, y, jx, jy)) ** 2
acc[i][0] += (x-jy)/den1/den2
acc[i][1] += (y-jy)/den1/den2
'''
pos[i] += vel[i]
if not (0 < pos[i][0] < WIDTH) or not (0 < pos[i][1] < HEIGHT):
pos[i][0], pos[i][1] = randrange(0, WIDTH), randrange(0, HEIGHT)
vel[i] = 0
# acc[i] = 0
noisy += 0.007
def hatch(IM, sc=16):
print("hatching...")
PX = IM.load()
w, h = IM.size
lg1 = []
lg2 = []
for x0 in range(w):
for y0 in range(h):
x = x0 * sc
y = y0 * sc
if PX[x0, y0] > 144:
pass
elif PX[x0, y0] > 64:
lg1.append([(x, y + sc / 4), (x + sc, y + sc / 4)])
elif PX[x0, y0] > 16:
lg1.append([(x, y + sc / 4), (x + sc, y + sc / 4)])
lg2.append([(x + sc, y), (x, y + sc)])
else:
lg1.append([(x, y + sc / 4), (x + sc, y + sc / 4)])
lg1.append([(x, y + sc / 2 + sc / 4),
(x + sc, y + sc / 2 + sc / 4)])
lg2.append([(x + sc, y), (x, y + sc)])
lines = [lg1, lg2]
for k in range(0, len(lines)):
for i in range(0, len(lines[k])):
for j in range(0, len(lines[k])):
if lines[k][i] != [] and lines[k][j] != []:
if lines[k][i][-1] == lines[k][j][0]:
lines[k][i] = lines[k][i] + lines[k][j][1:]
lines[k][j] = []
lines[k] = [l for l in lines[k] if len(l) > 0]
lines = lines[0] + lines[1]
for i in range(0, len(lines)):
for j in range(0, len(lines[i])):
lines[i][j] = int(lines[i][j][0] +
sc * noise.noise(i * 0.5, j * 0.1, 1)
), int(lines[i][j][1] +
sc * noise.noise(i * 0.5, j * 0.1, 2)) - j
return lines
def getcontours(IM, sc=2):
print("generating contours...")
IM = find_edges(IM)
IM1 = IM.copy()
IM2 = IM.rotate(-90, expand=True).transpose(Image.FLIP_LEFT_RIGHT)
dots1 = getdots(IM1)
contours1 = connectdots(dots1)
dots2 = getdots(IM2)
contours2 = connectdots(dots2)
for i in range(len(contours2)):
contours2[i] = [(c[1], c[0]) for c in contours2[i]]
contours = contours1 + contours2
for i in range(len(contours)):
for j in range(len(contours)):
if len(contours[i]) > 0 and len(contours[j]) > 0:
if sum_distance(contours[j][0], contours[i][-1]) < 8:
contours[i] = contours[i] + contours[j]
contours[j] = []
for i in range(len(contours)):
contours[i] = [contours[i][j] for j in range(0, len(contours[i]), 8)]
contours = [c for c in contours if len(c) > 1]
for i in range(0, len(contours)):
contours[i] = [(v[0] * sc, v[1] * sc) for v in contours[i]]
for i in range(0, len(contours)):
for j in range(0, len(contours[i])):
contours[i][j] = int(contours[i][j][0] +
10 * noise.noise(i * 0.5, j * 0.1, 1)), int(
contours[i][j][1] +
10 * noise.noise(i * 0.5, j * 0.1, 2))
return contours
def init(self):
self.field = []
xoff = 0
xrand = Prandom(0, 1)
yrand = Prandom(0, 1)
for i in range(self.cols):
yoff = 0
tmp = []
for j in range(self.rows):
theta = Pmap(noise(xoff + xrand, yoff + yrand), 0, 1, 0,
2 * pi)
assert theta < 2 * pi
# Polar to cartesian coordinate transformation to get x and y components of the vector
tmp.append(PVector(cos(theta), sin(theta)))
yoff += 0.01
self.field.append(tmp)
xoff += 0.01
def _generate(self, x: int):
l = []
y = 0
for _ in range(round((noise(x * 0.01) + 1) * 50) + 50):
l.append(blocks.Stone((x, y)))
y += 1
for _ in range(3):
l.append(blocks.Dirt((x, y)))
y += 1
l.append(blocks.Grass((x, y)))
y += 1
for _ in range(5):
l.append(blocks.Air((x, y)))
y += 1
for _ in range(255 - len(l)):
l.append(blocks.Air((x, y)))
y += 1
self._generated[x] = l
for y, row in enumerate(bg):
for x, col in enumerate(row):
val = FILTERS['mountains'](col)
if val > 0.25:
zone.set_field(y, x, '=')
if __name__ == '__main__':
from bmp import Bitmap
size = 500
def rgb(x):
return (x, x, x)
f = Bitmap('output.bmp', size, size)
perlin = Perlin()
pnoise_harmonic = noise(6)
pnoise = perlin.noise
n = generate(size, size, pnoise_harmonic)
for y, row in enumerate(n):
for x, col in enumerate(row):
val = FILTERS['mountains'](col)
f.set_pixel(y, x, rgb(val))
f.flush()
f.close()
saveTime = time.time()
totalTime = saveTime - start_time
print("Totaly data prework time " + str(totalTime) + " seconds")
print(user_venc_in_plane)
#os.system("python /home/condauser/Documents/ao_seg/aorta_seg.py --path=/home/condauser/Documents/ao_seg/MK")
ppp = "/home/condauser/Documents/ao_seg/MK"
image, mrStruct, netflow = eddy(ppp)
eddyTime = time.time()
totalEddyTime = eddyTime - saveTime
print("Totaly eddy time " + str(totalEddyTime) + " seconds")
#gc.collect()
image, mrStruct = noise(image, mrStruct, netflow)
noiseTime = time.time()
totalNoiseTime = noiseTime - eddyTime
print("Totaly Noise time " + str(totalNoiseTime) + " seconds")
#gc.collect()
cc = TFrecord(mrStruct, user_venc_in_plane)
print("TFrecord finish")
io.savemat('MK/vel_struct.mat', {'mrStruct': mrStruct})
del mrStruct
jjj = Unet_test()
#del jj
aliasTime = time.time()
totalAliasTime = aliasTime - noiseTime
print("Totaly Aliasing time " + str(totalAliasTime) + " seconds")
#jjj = io.loadmat("/home/condauser/Documents/ao_seg/new_vel.mat")
print "Unable to open", ewfile, "error was", e.args[1]
sys.exit(11)
except ValueError as e:
print "Conversion error in", ewfile, "error was", e.args[0]
sys.exit(12)
signals = []
for sig in siglist:
nsig = sigdets()
nsig.parse(sig)
signals.append(nsig)
col2 = np.zeros_like(ewf[1])
col4 = col2 + 1.0
col6 = col2 + 1.0
for sig in signals:
col2 += sig.apply(ewf[1], 0)
col4 += sig.apply(ewf[1], 1)
col6 += sig.apply(ewf[1], 2)
if snr != 0:
col2 = noise.noise(col2, snr, gauss)
col4 = noise.noise(col4, snr, gauss)
col6 = noise.noise(col6, snr, gauss)
ewf[2] = col2
ewf[4] = col4
ewf[6] = col6
np.savetxt(outfile, ewf.transpose())
if dp[pointer + 1][1] > dp[pointer + 1][3]:
text += "0"
index = 3
else:
text += "0"
index = 1
elif index == 3:
if dp[pointer + 1][1] > dp[pointer + 1][3]:
text += "1"
index = 3
else:
text += "1"
index = 1
pointer += 1
return text[::-1]
text = "amirhosseinahmadi"
print("Plain Text:", text)
huffman_encoded = huffman_encoder(text)
print("Huffman Encoding:", huffman_encoded)
conv_encoded = convolutional_enccoder(huffman_encoded)
print("Convolutional Encoding:", conv_encoded)
conv_encoded = [int(item) for item in conv_encoded]
noised_data = noise(conv_encoded)
noised_data = list(map(str, noised_data))
noised_data = "".join(noised_data)
viterbi_decoded = viterbi_decoder(noised_data)
print("Viterbi Decoding:", viterbi_decoded)
huffman_decoded = huffman_decoder(viterbi_decoded)
print("Huffman Decoding:", huffman_decoded)
def get_price_at_minutes(self, minutes):
noise_seed(self.cost_seed)
cost = self.cost_min + self.cost_range * noise(self.cost_offset +
minutes / 2000)
return max(1, int(cost))
for y, row in enumerate(bg):
for x, col in enumerate(row):
val = FILTERS['mountains'](col)
if val > 0.25:
zone.set_field(y, x, '=')
if __name__ == '__main__':
from bmp import Bitmap
size = 500
def rgb(x):
return (x, x, x)
f = Bitmap('output.bmp', size, size)
perlin = Perlin()
pnoise_harmonic = noise(6)
pnoise = perlin.noise
n = generate(size, size, pnoise_harmonic)
for y, row in enumerate(n):
for x, col in enumerate(row):
val = FILTERS['mountains'](col)
f.set_pixel(y, x, rgb(val))
f.flush()
f.close()
phasesr = np.arange(0.0, 1.0, res['pstep'])
phases = phasesr * TWOPI
errs = np.zeros_like(times) + err
lo = []
mid = []
hi = []
s1 = []
s2 = []
s3 = []
for p in phases:
sig = a1 * np.sin(f1 * times) + a2 * np.sin(f2 * times + p)
if snr != 0:
sig = noise.noise(sig, snr, gauss)
if usegatspy:
model = LombScargle().fit(times, sig, err)
pgram = model.periodogram(periods)
else:
pgram = lsas(times, sig, errs, freqs)
maxima = argmaxmin.maxmaxes(periods, pgram)
if len(maxima) > 3:
maxima = maxima[0:3]
maxp = list(periods[maxima])
h, m, l = maxp
s1.append(h)
s2.append(m)
s3.append(l)
maxp.sort()
l, m, h = maxp
suff = resargs['suff']
if suff is None:
suff = '.' + string.replace("n%#.3g" % snr, ".", "_")
elif len(suff) == 0:
print "Invalid suffix should be at least 1 char"
sys.exit(11)
elif suff[0] != '.':
suff = '.' + suff
errors = 0
ok = 0
for spec in specs:
try:
indata = np.loadtxt(spec, unpack=True)
indata[ycolumn] = noise.noise(indata[ycolumn], snr, gauss)
np.savetxt(spec + suff, indata.transpose())
except IndexError:
print "Invalid y column", ycolumn, "in", spec
errors += 1
continue
except ValueError:
print "Invalid values in", spec
errors += 1
continue
except IOError as e:
print "IO Error in", spec, "was", e.args[1]
errors += 1
ok += 1
if errors > 0:
import encrypt
import parse_mat
import network_encoding
import noise
import viterbi_decoding
import decrypt
string = 'saharrajabi'
probs = parse_mat.parse('data\\freq.mat').get('freq')
codes = {}
print 'name: ' + string
source_encrypted = encrypt.encrypt(probs, string, codes)
print 'source encoder output: ' + source_encrypted
network_encrypted = network_encoding.encode(source_encrypted)
print 'network encoder output: ' + network_encrypted
network_encrypted = map(lambda x: int(x), list(network_encrypted))
noisy_data = noise.noise(network_encrypted)
noisy_data = ''.join(map(lambda x: str(x), noisy_data))
print 'noisy data: ' + noisy_data
err, network_decoded = viterbi_decoding.decode(noisy_data)
print 'network decoder output: ' + network_decoded
decoded = decrypt.decrypt(network_decoded, codes)
print 'source decoder output: ' + decoded
r = deriv.sin(x)
print "r", deriv.v(r)
print "dr/dx", deriv.v(deriv.d(r, x))
print "dr/dy", deriv.v(deriv.d(r, y))
print "d2r/dx2", deriv.v(deriv.d(r, x, x))
print "d2r/dy2", deriv.v(deriv.d(r, y, y))
print "d2r/dx/dy", deriv.v(deriv.d(r, x, y))
print "d2r/dy/dx", deriv.v(deriv.d(r, y, x))
print
print
print
import noise
x = -1 + noise.noise(3)
y = 2 + noise.noise(7)
r = x, y, x + y, x - y, x / y, x * x, x * y, y * y
print ' '.join("%7.03f" % noise.E(x) for x in r), "E"
print ' '.join("%7.03f" % noise.var(x) for x in r), "variance"
print
print "covariance matrix:"
for row in noise.cov_matrix(r):
for col in row:
print "%7.03f" % col,
print
def mk_noise(self):
self.StatusLabel.hide()
noise.noise(im)
self.StatusLabel.show()
im.last_op = 'noise'
def main():
startTime = time.time()
args = _getparser().parse_args()
_dispParser(args)
tf._stats(args.fname.name)
header, _, offset, obs, dates = tf._read(args.fname.name)
A, cycle = dm(args.fname.name, args.periods, args.fs)._coefficients()
L = obs[:, 0].reshape(len(obs[:, 0]), 1)
dof = len(A[:, 0]) - len(A[0, :])
unkOLS, sUnkOLS, s0OLS, resid = ls._lse(A, L)
WRMS = np.sqrt((resid.T @ resid) / len(A[:, 0]))[0][0]
#print("WRMS: " + str(WRMS))
for line in header:
if 'UNIT' in line:
unit = line.split(": ")[1].split("\n")[0]
break
unitLabel = [unit] * len(unkOLS)
unitLabel[1] = unitLabel[1] + "/year"
unkLabel = ['intercept', 'trend']
offsetLabel = [
"{}{}".format(a, b)
for a, b in zip(["offset at "] * len(offset), offset)
]
unkLabel = np.append(unkLabel, offsetLabel)
seasLabel = []
for s1 in range(len(cycle)):
seasLabel = np.append(seasLabel, [
'sin ' + "{:10.4f}".format(cycle[s1]),
'cos ' + "{:10.4f}".format(cycle[s1])
])
unkLabel = np.append(unkLabel, seasLabel)
if args.ols:
print("\n Ordinary Least-squares estimation\n")
for o in range(len(unkOLS)):
print("%20s: %10.4f +- %9.4f %s" % \
(unkLabel[o], \
unkOLS[o], \
sUnkOLS[o], \
unitLabel[o]))
print("\n")
print("%20s: %14.8f %8s\n" % ('s0', s0OLS, unit))
print("\nElapsed time : %.1f sec\n" % (time.time() - startTime))
return
for line in header:
if 'DATE FORMAT' in line:
dateFormat = line.split(": ")[1].split("\n")[0]
break
_, J = noise(dates, dateFormat, args.kappa, args.fs).mat()
resultTemp = np.zeros((args.nRND**2, 3))
wnaLast = np.zeros((args.repeat, 1), dtype=float)
fnaLast = np.zeros((args.repeat, 1), dtype=float)
for i in range(args.repeat):
printProgressBar(i + 1,
args.repeat,
prefix=' Progress',
suffix='Complete')
_, yWNA, _, yFNA = ss(args.alpha, WRMS, dof, args.nRND)._randomPoints()
idx = 0
for j in range(args.nRND):
cWN = (yWNA[j]**2 * np.eye(len(A[:, 0]), dtype=float))
for k in range(args.nRND):
C = cWN + (yFNA[k]**2 * J)
_, _, s0, _ = ls._lse(A, L, la.cholesky(la.inv(C)).T)
resultTemp[idx, :] = [yWNA[j], yFNA[k], s0]
idx += 1
#np.savetxt('15-s0s.dat', resultTemp, fmt="%10.10f")
xx = np.arange(min(resultTemp[:, 0]), max(resultTemp[:, 0]), args.incr)
yy = np.arange(min(resultTemp[:, 1]), max(resultTemp[:, 1]), args.incr)
XX, YY = np.meshgrid(xx, yy)
ZZ = griddata((resultTemp[:, 0], resultTemp[:, 1]),
resultTemp[:, 2], (XX, YY),
method='cubic')
#np.savetxt('1-XX.dat', XX, fmt="%10.4f")
#np.savetxt('2-YY.dat', YY, fmt="%10.4f")
#np.savetxt('3-ZZ.dat', ZZ, fmt="%10.4f")
if (np.min(np.ma.masked_invalid(ZZ)) < 1) and (np.max(
np.ma.masked_invalid(ZZ)) > 1):
cs = plt.contour(xx, yy, np.ma.masked_invalid(ZZ), [1])
L1 = cs.collections[0].get_paths()[0]
coorL1 = L1.vertices
else:
index = np.where(
abs(np.ma.masked_invalid(ZZ) -
1) == np.amin(abs(np.ma.masked_invalid(ZZ) - 1)))
coorL1 = np.array([[XX[index[0][0],index[1][0]], \
YY[index[0][0],index[1][0]]]])
#np.savetxt('4-coorL1.dat', coorL1, fmt="%10.4f")
s0 = np.zeros((len(coorL1[:, 0]), 1), dtype=float)
for d in range(len(coorL1[:, 0])):
C2 = (coorL1[d,0]**2 * np.eye(len(A[:,0]), dtype=float)) + \
(coorL1[d,1]**2 * J)
_, _, s0[d], _ = ls._lse(A, L, la.cholesky(la.inv(C2)).T)
diffIDX = np.where(abs(s0 - 1) == np.amin(abs(s0 - 1)))[0][0]
wnaLast[i] = coorL1[diffIDX, 0]
fnaLast[i] = coorL1[diffIDX, 1]
#np.savetxt('5-noiseAmp.dat', np.concatenate((wnaLast, fnaLast), axis=1), fmt="%10.4f")
Cfin = (statistics.median(wnaLast)**2 * np.eye(len(A[:,0]), dtype=float)) + \
(statistics.median(fnaLast)**2 * J)
unk_fin, sUnk_fin, s0_fin, _ = ls._lse(A, L, la.cholesky(la.inv(Cfin)).T)
print("\n")
for o in range(len(unk_fin)):
print("%20s: %10.4f +- %9.4f %s" % \
(unkLabel[o], \
unk_fin[o], \
sUnk_fin[o], \
unitLabel[o]))
print("\n")
print("%20s: %10.4f %s" % ('wna', statistics.median(wnaLast), unit))
print("%20s: %10.4f %s" %
('fna', statistics.median(fnaLast), unit + "/year^0.25"))
print("%20s: %14.8f %s\n" % ('s0', s0_fin, unit))
if args.writeModel:
for ii in range(len(header)):
if 'COMMENT' and 'outliers' in header[ii]:
header[ii] = "* COMMENT : Model values" + "\n"
break
modelFilename = (args.fname.name.split(".")[0] + "_model.tse")
tf._write(modelFilename, \
header, \
np.concatenate(((A @ unk_fin), np.zeros((len(A[:,0]),1), dtype=float)), axis=1), \
dates)
print("\nElapsed time : %.1f sec\n" % (time.time() - startTime))
# set env
env = ArmEnv([0., 0., 0.], [0., 0., 0.], [0., 0., 0.], [0., 0., 0.])
s_dim = env.state_dim
a_dim = env.action_dim
a_bound = env.action_bound
# set RL method (continuous)
a_scale = [1000, 3000]
#a_scale = [0, 10]
rl = DDPG(a_dim, s_dim, a_scale)
noise_mean = 0
noise_std_dev = 0.2
noise_theta = 0.15
noise_dt = env.dt
noise = noise(a_dim, noise_mean, noise_std_dev, noise_theta, noise_dt)
steps = []
def train():
# start training
for i in range(MAX_EPISODES):
s = env.reset()
noise.reset()
ep_r = 0.
for j in range(MAX_EP_STEPS):
#j = 0
#while not (s[0] > 400 or s[1] > 400 or s[0] < 0 or s[1] < 0):
#env.render()
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import numpy as np
import random
import array
import cv2
from noise import noise
from loader import MNIST
from save import save
__all__ = [MNIST, ]
absPath = os.path.dirname(__file__)
mnist = MNIST(absPath)
noise = noise()
filesave = save()
def serialize(image):
im_array = np.zeros(784, np.int32)
#im_array = array.array('i', (0 for k in range(784)))
k = 0
for i in xrange(image.__len__()):
for j in xrange(image[i].__len__()):
im_array[k] = int(round(image[i, j]))
if im_array[k] < 0:
im_array[k] = 0
k = k + 1
return im_array
hs = hr = 1.0
results.append([obst, obst, ew, 0.0, hs, 0.0, hr, 0,0])
continue
ewleft, ewright = prof.ewinds
if ewleft == ewright:
print "Error, cannot find EW in", sf
errors += 1
nohorns += 1
ew = 0.0
hs = hr = 1.0
results.append([obst, obst, ew, 0.0, hs, 0.0, hr, 0.0])
continue
if snr > -1000.0:
amps = noise.noise(amps, snr, gauss)
ewsz = si.simps(amps[ewleft:ewright+1]-1.0, wavelengths[ewleft:ewright+1])
ew = ewsz - (wavelengths[ewright] - wavelengths[ewleft])
if prof.twinpeaks:
lhmax, rhmax = prof.maxima
minind = prof.minima[0]
minamp = amps[minind]
below_minamp = np.where(amps < minamp)[0]
lwhere = below_minamp[below_minamp < lhmax][-1] + 1
rwhere = below_minamp[below_minamp > rhmax][0] - 1
lhornsz = si.simps(amps[lwhere:minind+1] - 1.0, wavelengths[lwhere:minind+1])
rhornsz = si.simps(amps[minind:rwhere+1]- 1.0, wavelengths[minind:rwhere+1])
lwl = wavelengths[minind] - wavelengths[lwhere]
rwl = wavelengths[rwhere] - wavelengths[minind]
for letter in letternames:
letterpath = path_original_letters + '/' + letter
newletterpath = path_new_letters + '/' + letter + '/'
if not os.path.exists(newletterpath):
os.mkdir(newletterpath)
imagenames = os.listdir(letterpath)
for imagename in imagenames[0:5]:
imagepath = letterpath + '/' + imagename
image = cv2.imread(imagepath,0) #read as grayscale image
#print(newletterpath)
#Now we will do the follow with all images:
rotation(image, newletterpath + imagename[0:-4], [5,10,15,20,25,30])
noise(image, newletterpath + imagename[0:-4], [1.2,1.4,1.6,1.8,2])
translation(image, newletterpath + imagename[0:-4],[3,5])
shearing(image, newletterpath + imagename[0:-4],[0,0.5,0.8,1.2])
import Image
import math
size = 256
inv = 1.0 / size
im = Image.new("RGB", (size, size))
redOffs = (1.3579, 0.25672, 0.4567)
greenOffs = (2.4786, 1.54567, 2.654)
blueOffs = (3.5795, 2.16713, 1.0912)
# now build the texture
for x in range(size):
for y in range(size):
x1 = inv * x
y1 = inv * y
# get func value based on noise
n = (1 + noise.noise(redOffs[0] + 35 * x1, redOffs[1] + 4 * y1,
redOffs[2])) * 0.15
v = (y1 + n) * 0.5
if v < 0:
v = 1.0
im.putpixel((x, y), (255 * v, 0, 0))
im.show()
im.save("ThirdSpace.bmp", "bmp")
from noise import pnoise2 as noise parts = [] rings = [] height = 40 segs = 30 h_freq_scale = .1 a_freq_scale = 1 for i in range( height ): ring = [] for a in range( segs ): ring.append( (noise( i*h_freq_scale, math.sin((float(a)/segs)*math.pi)*a_freq_scale, 5 )+1)*.7 ) rings.append( ring ) h_scale = lambda h: (height+1 - h)*.25 + 5 a_to_p = lambda a, v, h: [math.cos(a)*h_scale(h)*v,math.sin(a)*h_scale(h)*v,h] for h in range( len( rings ) -1 ): ringa = rings[h] ringb = rings[h+1] points = [] print h for i in range( len( ringa ) ): print ringa[i] points.append( a_to_p( i * ((math.pi*2)/segs), ringa[i], h ) ) for i in range( len( ringb ) ): points.append( a_to_p( i * ((math.pi*2)/segs), ringb[i], h+1 ) )
def n2D(x, y, offs):
return 127 * noise.noise(offs[0] + x, offs[1] + y, offs[2]) + 128
def __init__(self,size=1000):
self.size=size
self.grid=[noise.noise(i) in range(size)]
self.gravity=[0,5]
# nearest neighbor algorithm
from upscale import upscale
p1 = args["factor"] # pixel
p2 = args["factor2"]
new_size = (p1, p2)
output = upscale(input_image, new_size)
save(option, output)
elif option == "crop" or option2 == "crop":
# cropping
from crop import crop
p1 = args["factor"] # pos
p2 = args["factor2"]
pos = (p1, p2) #origin
#end defined in crop.py
output = crop(input_image, pos)
save(option, output)
elif option == "contrast" or option2 == "contrast":
from contrast import change_contrast
factor = args["factor"] # factor
output = change_contrast(input_image, factor)
save(option, output)
elif option == "noise" or option2 == "noise":
from noise import noise
factor = args["factor"] # factor
output = noise(input_image, factor)
save(option, output)
else:
print(
"Usage: python pmake.py o -i<input_img>.<extention> -o <option1> or -oo<option2> -f1<factor 1> -f2 <factor2>"
)
def reverb(samples):
n = noise(0.8, 0.6, 0.01, 0.0)
n[0] = 1
return list(np.convolve(samples, n))
def f():
a = noise(3)
b = 8 + noise(2)
return a, b, a + b, a /b, a*a, a*b, b*b
