def plot_organism2(x1, y1, theta, fit, ax):
# TODO: Could make color relaitive to the genetic parent to see agent clustering
# The border could be this and the inside could be fitness?
# Could also add a dotted or thin line to the food that the agent is
# currently persuing. This is like "vision"
## plt.plot(x1, y1, marker=(3, 1, theta+33), markersize=11, color='darkslateblue', linestyle='None')
## plt.plot(x1, y1, marker=(3, 1, theta+33), markersize=6, color='mediumslateblue', linestyle='None')
##
plt.plot(x1,
y1,
marker=(3, 1, theta + 33),
markersize=11,
color=cm.hot(fit / 200),
linestyle='None')
plt.plot(x1,
y1,
marker=(3, 1, theta + 33),
markersize=6,
color='grey',
linestyle='None')
dir_len = 0.075
x2 = cos(radians(theta)) * dir_len + x1
y2 = sin(radians(theta)) * dir_len + y1
ax.add_line(
lines.Line2D([x1, x2], [y1, y2],
color=cm.hot(fit / 200),
linewidth=1,
zorder=10))
pass
def get_heatmap(self, sensor_data):
rgb_array = sensor_data['CenterRGB'].copy()
image = Image.fromarray(rgb_array)
transform = get_truncated_transform()
input_image = transform(image)
input_image.putalpha(255) # add RGB + A dimension for composite
mask_array = self.pilotnet.visual_mask.detach().squeeze().numpy().copy(
)
mask_array *= 30.0 # TODO: Hack to emphasize really light activations
mask_array[mask_array < 0.2] = 0
mask_image = Image.fromarray(np.uint8(cm.hot(mask_array) * 255))
# return mask_image
# Convert black pixels into transparent pixels
datas = mask_image.getdata()
newData = []
for item in datas:
if item[0] == 10 and item[1] == 0 and item[2] == 0:
newData.append((255, 255, 255, 0))
else:
newData.append(item)
mask_image.putdata(newData)
return Image.alpha_composite(input_image, mask_image)
def plot_path(x, f, max_pts=500):
"""
x: list of ndarray.
All the points on the way to the result
f: function.
Function to be optimized
"""
fig = figure()
ax = fig.add_subplot(111, projection='3d')
xs = []
ys = []
zs = []
for k in range(len(x)):
xs.append(x[k][0, 0])
ys.append(x[k][1, 0])
zs.append(f(x[k]))
max_f = 1.0 * max(zs)
min_f = 1.0 * min(zs)
colors = []
for idx in range(len(x)):
color = cm.hot(1.0 - (zs[idx] - min_f) / (max_f - min_f))
colors.append(color)
ax.scatter(xs=xs, ys=ys, zs=zs, color=colors)
show()
return
def add_masks_viz(ax,
masks,
bboxes_norm,
labels=None,
mask_alpha=0.6,
debug=False):
if labels is None:
labels = ["coco" for _ in bboxes_norm]
detect2d.visualize_bboxes(
ax,
bboxes_norm,
labels=labels,
label_color="w",
linewidth=2,
color="c",
)
for mask in masks:
if isinstance(mask, torch.Tensor):
mask = mask.cpu().detach().numpy()
base_mask = mask[:, :, np.newaxis].astype(np.float)
base_mask = base_mask / base_mask.max()
show_mask = np.concatenate(
[
# cm.YlGn(base_mask)[:, :, 0][:, :, :3],
cm.hot(base_mask)[:, :, 0][:, :, :3],
mask_alpha * (base_mask > 0).astype(np.float),
],
2,
)
ax.imshow(show_mask)
def visualize_map(self, title, figure):
map = self.stiffness
self.stiffness[self.stiffness == 0] = self.min
map = map - self.min
map = map / np.max(map)
map = map.reshape(int(np.sqrt(map.size)), int(np.sqrt(map.size)))
im = PIL.Image.fromarray(np.uint8(cm.hot(map) * 255))
cv_im = np.array(im)
msg_frame = CvBridge().cv2_to_imgmsg(cv_im, 'rgba8')
self.pub.publish(msg_frame)
if self.visualize:
plt.figure(figure)
plt.clf()
fig = plt.imshow(map, origin='lower', cmap=cm.hot)
plt.title(title)
plt.colorbar()
if len(self.probedPoints) != 0:
plt.scatter(self.probedPoints[:, 0], self.probedPoints[:, 1])
# plt.xlim((0,int(np.sqrt(map.size))))
# plt.ylim((0,int(np.sqrt(map.size))))
# plt.tight_layout()
# plt.show()
plt.pause(0.01)
def visualize_map(self, title, figure, map=None, probed_points=None):
if map is None:
dense_grid = self.generateGrid(res=1)
ymu = self.gp.predict(self.grid, return_std=False)
ymu[ymu < 0] = 0
map = ymu
map = map.reshape(int(np.sqrt(map.size)), int(np.sqrt(map.size)))
# map = map/np.max(map)
im = PIL.Image.fromarray(np.uint8(cm.hot(map) * 255))
cv_im = np.array(im)
# cv_im = cv2.flip(cv_im,0)
# embed()
msg_frame = CvBridge().cv2_to_imgmsg(cv_im, 'rgba8')
self.pub.publish(msg_frame)
if self.visualize:
plt.figure(figure)
plt.clf()
fig = plt.imshow(map, origin='upper', cmap=cm.hot)
plt.title(title)
plt.colorbar()
if not (probed_points is None):
plt.scatter(probed_points[:, 0], probed_points[:, 1])
plt.xlim((0, int(np.sqrt(map.size))))
plt.ylim((0, int(np.sqrt(map.size))))
# plt.tight_layout()
# plt.show()
plt.pause(0.01)
return msg_frame
def display(pc, i):
# 3D plotting of point cloud
fig = plt.figure(i)
ax = fig.gca(projection='3d')
#ax.set_aspect('equal')
X = pc[0]
Y = pc[1]
Z = pc[2]
c = pc[3]
"""
radius = np.sqrt(X**2 + Y**2 + Z**2)
X = X[np.where(radius<20)]
Y = Y[np.where(radius<20)]
Z = Z[np.where(radius<20)]
c = pc.points[3][np.where(radius<20)]
print(radius)
"""
ax.scatter(X, Y, Z, s=1, c=cm.hot((c / 100)))
max_range = np.array(
[X.max() - X.min(),
Y.max() - Y.min(),
Z.max() - Z.min()]).max()
Xb = 0.5 * max_range * np.mgrid[
-1:2:2, -1:2:2, -1:2:2][0].flatten() + 0.5 * (X.max() + X.min())
Yb = 0.5 * max_range * np.mgrid[
-1:2:2, -1:2:2, -1:2:2][1].flatten() + 0.5 * (Y.max() + Y.min())
Zb = 0.5 * max_range * np.mgrid[
-1:2:2, -1:2:2, -1:2:2][2].flatten() + 0.5 * (Z.max() + Z.min())
i = 0
for xb, yb, zb in zip(Xb, Yb, Zb):
i = i + 1
ax.plot([xb], [yb], [zb], 'w')
def get_colors(self, qty):
qty = np.power(qty / qty.max(), 1.0 / CONTRAST)
if COLORMAP == 0:
rgba = cm.gray(qty, alpha=ALPHA)
elif COLORMAP == 1:
rgba = cm.afmhot(qty, alpha=ALPHA)
elif COLORMAP == 2:
rgba = cm.hot(qty, alpha=ALPHA)
elif COLORMAP == 3:
rgba = cm.gist_heat(qty, alpha=ALPHA)
elif COLORMAP == 4:
rgba = cm.copper(qty, alpha=ALPHA)
elif COLORMAP == 5:
rgba = cm.gnuplot2(qty, alpha=ALPHA)
elif COLORMAP == 6:
rgba = cm.gnuplot(qty, alpha=ALPHA)
elif COLORMAP == 7:
rgba = cm.gist_stern(qty, alpha=ALPHA)
elif COLORMAP == 8:
rgba = cm.gist_earth(qty, alpha=ALPHA)
elif COLORMAP == 9:
rgba = cm.spectral(qty, alpha=ALPHA)
return rgba
def draw_output(im,
rect_gt=None,
rect_pred=None,
hmap_pred=None,
color_pred=None,
color_gt=None):
'''Modifies original image and returns an image that may be different.'''
if color_pred is None:
color_pred = COLOR_PRED
if color_gt is None:
color_gt = COLOR_GT
draw = ImageDraw.Draw(im)
if rect_gt is not None:
rect_gt = _rect_to_int_list(_unnormalize_rect(rect_gt, im.size))
draw.rectangle(rect_gt, outline=color_gt)
if rect_pred is not None:
rect_pred = _rect_to_int_list(_unnormalize_rect(rect_pred, im.size))
draw.rectangle(rect_pred, outline=color_pred)
del draw
if hmap_pred is not None:
assert len(hmap_pred.shape) == 3
hmap_pred = Image.fromarray(cm.hot(hmap_pred[:, :, 0],
bytes=True)).convert('RGB')
# Caution: This does not resize with align_corners=True.
if hmap_pred.size != im.size: # i.e., OTB
hmap_pred = hmap_pred.resize(im.size)
im = Image.blend(im, hmap_pred, 0.5)
return im
def get_cmap(group, age, scale=1):
structure_vals, n = get_mean_value_per_structure(group, age, dat['structure_id'].unique())
rgb_vals = structure_vals.copy()
for key in structure_vals:
rgb_vals[key] = tuple([255*i for i in cm.hot(structure_vals[key]*scale)[:3]])
rgb_vals[0] = (0, 0, 0)
return rgb_vals, n
def get_colors(self, qty):
qty = np.power(qty / qty.max(), 1.0 / CONTRAST)
if COLORMAP == 0:
rgba = cm.gray(qty, alpha=ALPHA)
elif COLORMAP == 1:
rgba = cm.afmhot(qty, alpha=ALPHA)
elif COLORMAP == 2:
rgba = cm.hot(qty, alpha=ALPHA)
elif COLORMAP == 3:
rgba = cm.gist_heat(qty, alpha=ALPHA)
elif COLORMAP == 4:
rgba = cm.copper(qty, alpha=ALPHA)
elif COLORMAP == 5:
rgba = cm.gnuplot2(qty, alpha=ALPHA)
elif COLORMAP == 6:
rgba = cm.gnuplot(qty, alpha=ALPHA)
elif COLORMAP == 7:
rgba = cm.gist_stern(qty, alpha=ALPHA)
elif COLORMAP == 8:
rgba = cm.gist_earth(qty, alpha=ALPHA)
elif COLORMAP == 9:
rgba = cm.spectral(qty, alpha=ALPHA)
return rgba
def manipulate_latent(model, data, args):
print('-' * 30 + 'Begin: manipulate' + '-' * 30)
x_test, y_test = data
index = np.argmax(y_test, 1) == args.digit
number = np.random.randint(low=0, high=sum(index) - 1)
x, y = x_test[index][number], y_test[index][number]
x, y = np.expand_dims(x, 0), np.expand_dims(y, 0)
noise = np.zeros([1, 3, 16])
x_recons = []
for dim in range(16):
for r in [
-0.25, -0.2, -0.15, -0.1, -0.05, 0, 0.05, 0.1, 0.15, 0.2, 0.25
]:
tmp = np.copy(noise)
tmp[:, :, dim] = r
x_recon = model.predict([x, y, tmp])
x_recons.append(x_recon)
x_recons = np.concatenate(x_recons)
img = combine_images(x_recons, height=16)
image = cm.hot(img) * 255
Image.fromarray(image.astype(
np.uint8)).save(args.save_dir + '/manipulate-%d.png' % args.digit)
print('manipulated result saved to %s/manipulate-%d.png' %
(args.save_dir, args.digit))
print('-' * 30 + 'End: manipulate' + '-' * 30)
def visualize_map(self, title, figure, map=None, probed_points=None):
if map is None:
dense_grid=self.generateGrid(res=1)
ymu = self.gp.predict(self.grid, return_std=False)
ymu[ymu<0]=0
map = ymu
map=map.reshape(self.domain['L1'],self.domain['L2'])
normalized_map = map/np.max(map)
im = PIL.Image.fromarray(np.uint8(cm.hot(normalized_map)*255))
cv_im=np.array(im)
msg_frame = CvBridge().cv2_to_imgmsg(cv_im,'rgba8')
self.pub.publish(msg_frame)
if self.visualize:
plt.figure(figure)
plt.clf()
fig=plt.imshow(normalized_map, origin='lower',cmap=cm.hot)
plt.title(title)
plt.colorbar()
if not (probed_points is None):
plt.scatter(probed_points[:,0],probed_points[:,1])
# plt.xlim((0,100))
# plt.ylim((0,100))
# plt.tight_layout()
# plt.show()
filename = str(len(probed_points)).zfill(4)
plt.savefig(filename + '.png')
np.savetxt(filename + '.txt', self.estimated_map['mean'])
plt.pause(0.01)
return msg_frame
def main():
print("static uint32_t colormap[] = {")
for i in range(256):
l = i / 255
c = cm.hot(l, 1, True)
print(" 0x{0:02x}{1:02x}{2:02x}".format(c[0], c[1], c[2]), end='')
if (i != 255):
print(",")
print("\n};")
def test(model, data, args):
time_date = datetime.datetime.now().strftime("%y-%m-%d-%H-%M")
filename = time_date + '_capsule_network' + '_batch_size=' + str(
args.batch_size) + '_epochs=' + str(args.epochs) + '_ntr=' + str(
x_train.shape[0]) + '_nch=' + str(x_train.shape[3])
x_test, y_test = data
y_pred, x_recon = model.predict(x_test, batch_size=100)
print('-' * 30 + 'Begin: test' + '-' * 30)
img = combine_images(
np.concatenate([np.abs(x_test[:50]),
np.abs(x_recon[:50])]))
image = cm.hot(img) * 255
Image.fromarray(image.astype(np.uint8)).save(args.save_dir +
"/real_and_recon4.png")
print('Reconstructed images are saved to %s/real_and_recon4.png' %
args.save_dir)
print('-' * 30 + 'End: test' + '-' * 30)
predicted_classes = y_pred
predicted_classes = np.asarray(predicted_classes)
b = np.zeros_like(predicted_classes)
b[np.arange(len(predicted_classes)), predicted_classes.argmax(1)] = 1
predicted_classes = np.argmax(np.round(b), axis=1)
b = np.zeros_like(y_test)
b[np.arange(len(y_test)), y_test.argmax(1)] = 1
y_test = np.argmax(np.round(b), axis=1)
class_names = ['Unresolved', 'FRI', 'FRII']
from sklearn.metrics import classification_report
print(
classification_report(y_test,
predicted_classes,
target_names=class_names,
digits=4))
os.chdir(args.save_dir)
filename1 = filename + '_classification_report.txt'
f = open(filename1, 'w')
f.write(
classification_report(y_test,
predicted_classes,
target_names=class_names,
digits=4))
f = open(filename1, 'w')
f.write(
classification_report(y_test,
predicted_classes,
target_names=class_names,
digits=4))
############# Architecture
with open(filename + '_architecture.txt', 'w') as fh:
model.summary(print_fn=lambda x: fh.write(x + '\n'))
with open(filename + '_architecture.txt', 'w') as fh:
model.summary(print_fn=lambda x: fh.write(x + '\n'))
def write_images(output_dir, truth_path, color_img_path, visual_path,
maj_voted, strength_of_disagreement_img, equality_img):
img_width = equality_img.shape[1]
img_height = equality_img.shape[0]
# Write already computed images
cv2.imwrite(output_dir + '/majority_voted/' +
Path(truth_path).name, maj_voted[:, :, 0])
cv2.imwrite(output_dir + '/votes/' +
Path(truth_path).name, maj_voted[:, :, 1])
cv2.imwrite(output_dir + '/disagreement_strength/' +
Path(truth_path).name, strength_of_disagreement_img)
# Create image for direct visual inspection:
pad_top = 40
pad_middle = 10
color_map = [i * (255 // 3) for i in range(4)]
colorize = np.vectorize(lambda x: color_map[x])
imgs = [equality_img * 255,
colorize(maj_voted[:, :, 1]),
colorize(np.multiply(np.invert(equality_img), maj_voted[:, :, 1])),
colorize(maj_voted[:, :, 0])]
img_captions = ['Majority vote == ground truth?',
'Number of Votes',
'Disagreement with ground truth',
'Majority-predicted classes']
# Combine all four images on a single canvas with respective captions
canvas = np.full(shape=(img_height + pad_top,
img_width * 5 + 4 * pad_middle),
fill_value=23, dtype=np.uint8)
for j, (img, img_caption) in enumerate(zip(imgs, img_captions)):
left = (img_width + pad_middle) * j
right = (img_width + pad_middle) * j + img_width
canvas[pad_top:, left:right] = img.astype(np.uint8)
canvas = cv2.putText(img=canvas,
text=img_caption,
org=(left, pad_top - 10),
fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.4,
color=120,
thickness=1)
if type(canvas) == cv2.UMat:
canvas = cv2.UMat.get(canvas)
canvas = cm.hot(canvas, bytes=True)[:, :, :3]
canvas = cv2.cvtColor(canvas, cv2.COLOR_BGR2RGB)
# Add color image
left = (img_width + pad_middle) * 4
right = (img_width + pad_middle) * 4 + img_width
canvas[pad_top:, left:right] = cv2.imread(color_img_path)
cv2.imwrite(visual_path, canvas)
def heatmap(self, activations, unit=None, mode='bilinear'):
amin, amax = self.range_for(activations, unit)
if unit is None:
a = activations
else:
a = activations[unit]
upsampler = self.upsampler_for(a)
a = upsampler(a[None,None,...], mode=mode)[0,0].cpu()
return PIL.Image.fromarray(
(cm.hot((a - amin) / (1e-10 + amax - amin)) * 255
).astype('uint8'))
def __init__(self, parent):
wx.Panel.__init__(self, parent)
self.Bind(wx.EVT_PAINT, self.OnPaint)
self.graph = np.zeros((SIZE,SIZE),np.float32)
data = np.uint8(255*cm.hot(self.graph)[:,:,0:3])
self.bmp = wx.BitmapFromBuffer(SIZE, SIZE, data)
# create a buffer that can be used to store historical samples
self.sample_buffer=np.zeros(NUM_WINDOWS*CHUNK,np.float32)
self.sample_pos = 0 # start a first sample location
self.graph_pos = 0 # which time line we are at
self.in_setup = True
def color(self, n):
n = self.color_clip(n)
n = (n - self.color_range[0]) / float(
self.color_range[1] - self.color_range[0]) # range to 0~1
c = cm.hot(n) # rgba tuple
red = int(c[0] * 255) << 16 & 0xFF0000
blue = int(c[1] * 255) << 8 & 0x00FF00
green = int(c[2] * 255) & 0x0000FF
rgb = 0x000000 | red | blue | green
return rgb
def __init__(self, parent):
wx.Panel.__init__(self, parent)
self.Bind(wx.EVT_PAINT, self.OnPaint)
self.graph = np.zeros((SIZE,SIZE),np.float32)
data = np.uint8(255*cm.hot(self.graph)[:,:,0:3])
self.bmp = wx.BitmapFromBuffer(SIZE, SIZE, data)
# create a buffer that can be used to store historical samples
self.sample_buffer=np.zeros(NUM_WINDOWS*CHUNK,np.float32)
self.sample_pos = 0 # start a first sample location
self.sample_count = -1
self.graph_pos = 0 # which time line we are at
self.in_setup = True
def update(frame_number):
# print frame_number
x, y = gen.next()
f.set_pos(xx + x, yy + y)
# f.set_pos(frame_number,100)
src = f.next()
img.set_array(src)
a, z = ld.get_value(copy(src))
# print x,y, z
img2.set_array(a)
# print z, x, y
xs.append(xx + x)
ys.append(yy + y)
line.set_data(xs, ys)
scatter2.set_data([xx + x], [yy + y])
ts.append(time.time() - st)
# print z
# z /= 3716625.0
# z += 0.1 * random.random()
# print x, y, z
pattern.set_point(z, x, y)
zs.append(z)
# print ts
# print zs
tvint.set_data(ts, zs)
ax4.set_ylim(min(zs) * 0.9, max(zs) * 1.1)
tcs.append(f.time_constant)
rcs.append(f.cradius)
tc.set_data(ts, tcs)
rs.set_data(ts, rcs)
# >>>>>>>>>>>>>>>>>>>>>>>
# add a plot that is the current points of the triangle
# >>>>>>>>>>>>>>>>>>>>>>>
xy = pattern._tri.xys()
x, y, z = zip(*xy)
# print x, y
# cp.set_data(x, y)
cp.set_offsets(zip(x, y))
maz, miz = max(z), min(z)
z = [(zi - miz) / (maz - miz) for zi in z]
print z
cp.set_facecolors([cm.hot(zi) for zi in z])
# cp.set_edgecolors(z)
if not pattern._tri.is_equilateral():
print 'not eq'
def update(frame_number):
# print frame_number
x, y = gen.next()
f.set_pos(xx + x, yy + y)
# f.set_pos(frame_number,100)
src = f.next()
img.set_array(src)
a, z = ld.get_value(copy(src))
# print x,y, z
img2.set_array(a)
# print z, x, y
xs.append(xx + x)
ys.append(yy + y)
line.set_data(xs, ys)
scatter2.set_data([xx + x], [yy + y])
ts.append(time.time() - st)
# print z
# z /= 3716625.0
# z += 0.1 * random.random()
# print x, y, z
pattern.set_point(z, x, y)
zs.append(z)
# print ts
# print zs
tvint.set_data(ts, zs)
ax4.set_ylim(min(zs) * 0.9, max(zs) * 1.1)
tcs.append(f.time_constant)
rcs.append(f.cradius)
tc.set_data(ts, tcs)
rs.set_data(ts, rcs)
# >>>>>>>>>>>>>>>>>>>>>>>
# add a plot that is the current points of the triangle
# >>>>>>>>>>>>>>>>>>>>>>>
xy = pattern._tri.xys()
x, y, z = zip(*xy)
# print x, y
# cp.set_data(x, y)
cp.set_offsets(zip(x, y))
maz, miz = max(z), min(z)
z = [(zi - miz) / (maz - miz) for zi in z]
print z
cp.set_facecolors([cm.hot(zi) for zi in z])
# cp.set_edgecolors(z)
if not pattern._tri.is_equilateral():
print 'not eq'
def animate_behav():
with open("pickles/learnednet5.p", "rb") as p:
dic = pickle.load(p)
sim3 = dic["sim3"]
m = dic["m"]
f = dic["f"]
net = sim3.network
start_idx = 8500
end_idx = 11250
interval = 25
query_start = (9500 - 8500) / 25
query_end = (10250 - 8500) / 25
f_pl = f[net.J_place_indices,
start_idx:end_idx:interval] # 9500-10250 query
f_pl = np.flip(f_pl, axis=0)
f_ep = f[net.J_episode_indices,
start_idx:end_idx:interval] # 9500-10250 query
f_ep = np.flip(f_ep, axis=0)
fig, axs = plt.subplots(2, 1, gridspec_kw={'height_ratios': [1, 4]})
camera = Camera(fig)
f_colors = [cm.hot(i) for i in f_pl[:, 0]]
my_circle = plt.Circle((0, 0), 0.7, color='white')
axs[0].imshow(repmat(f_ep[:, 0], 5, 1), cmap='hot')
axs[0].set_yticks([])
axs[0].set_xticks([])
axs[1].pie(x=np.ones(net.N_pl), colors=f_colors)
axs[1].add_artist(my_circle)
for i in range(f_pl.shape[1]):
f_colors = [cm.hot(i) for i in f_pl[:, i]]
axs[0].imshow(repmat(f_ep[:, i], 5, 1), cmap='hot')
axs[1].pie(x=np.ones(net.N_pl), colors=f_colors)
axs[1].add_artist(my_circle)
camera.snap()
Writer = ani.writers['ffmpeg']
writer = Writer(fps=15, metadata=dict(artist='Me'), bitrate=1800)
anim = camera.animate(interval=500, repeat_delay=3000, blit=True)
anim.save('behav.mp4', writer=writer)
def simpleTest():
width = 50
height = 100
#points = np.array([
# [0.2, 0.3],
# [0.7, 0.5],
# [0.6,0.2],
# [0.3, 0.9],
# [0.8, 0.1]
#])
points = sampleRandomPoints(100, width, height) #np.random.rand(20, 2)
print("Points:")
print(points)
cells = bounded_voronoi(points, [0, width, 0, height])
for i, cell in enumerate(cells):
print("cell", i, "with point", cell.generator())
print(cell.corner())
# example density map
density = np.zeros((width, height))
for x in range(width):
for y in range(height):
density[x, y] = 1 - math.sqrt((x / width - 0.5)**2 +
(y / height - 0.5)**2)
# compute cell integrals
cell_densities = [cell.integrate(density) for cell in cells]
# plotting
plt.figure(figsize=(20, 10))
ax1 = plt.subplot(1, 2, 1)
ax1.imshow(density, cmap=cm.hot)
ax2 = plt.subplot(1, 2, 2)
for cell, (d, _) in zip(cells, cell_densities):
vertices = np.array(list(cell.corner()) + [cell.corner()[0]])
ax2.fill(vertices[:, 0], vertices[:, 1], facecolor=cm.hot(d))
ax2.plot([cell.generator()[0] for cell in cells],
[cell.generator()[1] for cell in cells], 'b.')
ax2.plot([d[1][0] for d in cell_densities],
[d[1][1] for d in cell_densities], 'r.')
for cell in cells:
ax2.plot(cell.corner()[:, 0], cell.corner()[:, 1], 'go')
for cell in cells:
vertices = np.array(list(cell.corner()) + [cell.corner()[0]])
ax2.plot(vertices[:, 0], vertices[:, 1], 'k-')
#fig.show()
plt.savefig("bounded_voronoi.png")
def draw(L):
X = range(len(L))
Y = []
for v in get_next(L, items):
Y.append(v)
offset = [0] * len(Y[0])
plt.bar(X,
Y[0],
align='center',
color=cm.hot(map(lambda c: c * 2.5, Y[0])))
for p in zip(range(1, len(Y)), range(0, len(Y) - 1)):
offset = [a + b for a, b in zip(Y[p[1]], offset)]
plt.bar(X,
Y[p[0]],
bottom=offset,
align='center',
color=cm.hot(map(lambda c: c * 2.5, Y[p[0]])))
# plt.savefig('bin_packing02.png', bbox_inches='tight')
plt.gca().margins(0.1)
plt.show()
def visualize(self, samples, costs):
marker_array = vm.MarkerArray()
for i, (sample, cost) in enumerate(zip(samples, costs)):
if sample is None:
continue
origin = np.array([0., 0., 0.])
angle = 0.
for t in xrange(len(sample.get_U())):
marker = vm.Marker()
marker.id = i * len(sample.get_U()) + t
marker.header.frame_id = '/map'
marker.type = marker.ARROW
marker.action = marker.ADD
speed = sample.get_U(t=t, sub_control='cmd_vel')[0] / 5.
steer = sample.get_U(t=t, sub_control='cmd_steer')[0]
angle += (steer - 50.) / 100. * (np.pi / 2)
new_origin = origin + [
speed * np.cos(angle), speed * np.sin(angle), 0.
]
marker.points = [
gm.Point(*origin.tolist()),
gm.Point(*new_origin.tolist())
]
origin = new_origin
marker.lifetime = rospy.Duration()
marker.scale.x = 0.05
marker.scale.y = 0.1
marker.scale.z = 0.1
if cost == min(costs):
rgba = (0., 1., 0., 1.)
marker.scale.x *= 2
marker.scale.y *= 2
marker.scale.z *= 2
else:
rgba = list(cm.hot(1 - cost))
rgba[-1] = 0.5
marker.color.r, marker.color.g, marker.color.b, marker.color.a = rgba
marker_array.markers.append(marker)
# for id in xrange(marker.id+1, int(1e4)):
# marker_array.markers.append(vm.Marker(id=id, action=marker.DELETE))
self.debug_cost_probcoll_pub.publish(marker_array)
def heatmap(self, activations, unit=None, mode='bilinear'):
'''
Produces a heatmap from the given activations. The unit,
if specified, is an index into the activations tensor,
and is also used to dereference quantile ranges.
'''
amin, amax = self.range_for(activations, unit)
if unit is None:
a = activations
else:
a = activations[unit]
upsampler = self.upsampler_for(a)
a = upsampler(a[None, None, ...], mode=mode)[0, 0].cpu()
return PIL.Image.fromarray((cm.hot(
(a - amin) / (1e-10 + amax - amin)) * 255).astype('uint8'))
def _plot_distributions(self):
prs, pgg = self._samples()
to = self.template_objects
x_axis_width = len(prs[0])
bin_size = x_axis_width / (self.generalization_generator.range * 2)
amp = 0.3
# Scale the regularization samples bins to the [0, 1] interval
prs = prs / np.amax(prs)
prs = prs / 3
# x-axis
p_x = np.linspace(-self.generalization_generator.range,
self.generalization_generator.range, x_axis_width)
f, ax = plt.subplots(1)
# Draw the template objects at their position on the x axis
for idx, mu in enumerate(to.mu):
color = cm.hot(idx / 3 + 0.15)
# Draw the regularization samples
# label_regularization_samples = 'regularization samples ' + str(idx + 1)
plt.bar(p_x, prs[idx], 0.5, color=color, alpha=0.6)
# Draw the template objects at their position on the x axis
to_x = np.zeros(np.shape(p_x))
to_offset = ((self.generalization_generator.range + mu) * bin_size)
to_x[int(to_offset)] = amp
#label_template_objects = 'template object ' + str(idx + 1)
label_template_objects = 'Object type ' + str(idx + 1)
plt.bar(p_x, to_x, 1., color=color, label=label_template_objects)
# Draw the generalization generated samples
#label_generalization_samples = 'generalization generator ' + str(idx + 1)
plt.plot(p_x, pgg[idx], color=color, linewidth=1.2)
# Draw the figure
plt.xlim([
-self.generalization_generator.range,
self.generalization_generator.range
])
plt.ylim([0, 0.5])
plt.title('Generative One-Shot Learning')
plt.xlabel('Data values')
plt.ylabel('Probability density')
plt.legend()
plt.show()
def plot4d(inputarray, outname, xlabel='', ylabel='', zlabel='', project=1):
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
import numpy as np
from pylab import cm
fig = plt.figure(figsize=(8, 6), dpi=300)
ax = fig.add_subplot(111, projection='3d')
x, y, z, _ = inputarray.nonzero()
c = inputarray.flatten()
c = c[c != 0]
# c=c/c.max()
# c*=100
s = c / c.min()
from math import log10
s = s + 0.01
for i in range(len(s)):
s[i]=log10(s[i])
#sarr=c>0.06914700
#sarr2=c<0.06914702
#sarr=sarr&sarr2
#s[sarr]=100
#print(c)
#colmap = cm.ScalarMappable(cmap=cm.hsv)
#colmap.set_array(c)
#s=s+1
ax.scatter(x, z, y, c=cm.hot(c),s=s)
ax.set_xlabel(xlabel)
ax.set_ylabel(ylabel)
ax.set_zlabel(zlabel)
#fig.axes.get_xaxis().set_visible(False)
#fig.axes.get_yaxis().set_visible(False)
#fig.axes.get_zaxis().set_visible(False)
if len(outname):
fig.savefig(outname)
return ax,plt,x,y,z,c
def draw_pnt(data,header,output):
#x = np.linspace(0,nrays,nrays)
print "drawing"
npoints = int(header["npoints"])
date_begin_str = header["date_begin"]
date_end_str = header["date_end"]
time_begin_str = header["time_begin"]
time_end_str = header["time_end"]
o = ephem.Observer()
o.lon = "37:37:48"
t_begin = datetime.strptime(date_begin_str+" "+time_begin_str,"%d.%m.%Y %H:%M:%S")
msk = pytz.timezone("Europe/Moscow")
t_begin.replace(tzinfo=msk)
x = range(0,npoints)
t_tuple = [t_begin+timedelta(seconds=i*3600*1.00/npoints) for i in x]
s_t_tuple = []
for t in t_tuple:
o.date = datetime.utcfromtimestamp(float(t.strftime('%s')))
s_t_tuple.append(datetime.strptime(str(o.sidereal_time()),"%H:%M:%S.%f"))
fig = plt.figure()
ax1 = fig.add_subplot(111)
ax2 = ax1.twiny()
plt.ylim([0,8000])
ax2.plot(s_t_tuple,[0]*len(s_t_tuple),color = 'white',marker = '.',markersize = 0,lw = 0)
for r in reversed(xrange(nrays)):
c = cm.hot(r/76.,1)
y = [(np.mean(data[i][r][0:6])+ 100*r) for i in x]
ax1.plot(t_tuple,y,color = c,marker = '.',markersize = 0.1,lw = 0)
ax2.grid(True)
ax1.grid(True)
ax1.set_xlabel("Moscow Time")
ax2.set_xlabel("Star Time")
fig.suptitle("BSA data from:" + date_begin_str)
fig.autofmt_xdate()
fig.set_size_inches(18.5,10.5)
fig.savefig(output,dpi=300)
plt.close(fig)
def plot_walkers_traj(obs, traj_index='all'):
"""Plot the trajectory of all the individual walkers
Arguments:
obs_dict {dict} -- dictionary of observable
Keyword Arguments:
traj_index {str} -- all or index of a given walker (default: {'all'})
Returns:
float -- decorelation time
"""
eloc = obs['local_energy']
eloc = np.array(eloc).squeeze(-1)
nstep, nwalkers = eloc.shape
celoc = np.cumsum(eloc, axis=0).T
celoc /= np.arange(1, nstep + 1)
var_decor = np.sqrt(np.var(np.mean(celoc, axis=1)))
print(var_decor)
var = np.sqrt(np.var(celoc, axis=1) / (nstep - 1))
print(var)
Tc = (var_decor / var)**2
if traj_index is not None:
plt.subplot(1, 2, 1)
if traj_index == 'all':
plt.plot(eloc, 'o', alpha=1 / nwalkers, c='grey')
cmap = cm.hot(np.linspace(0, 1, nwalkers))
for i in range(nwalkers):
plt.plot(celoc.T[:, i], color=cmap[i])
else:
plt.plot(eloc[traj_index, :], 'o', alpha=1 / nwalkers, c='grey')
plt.plot(celoc.T[traj_index, :])
plt.subplot(1, 2, 2)
plt.hist(Tc)
plt.show()
return Tc
def get_fused_output(ct, pet, mask, true_mask):
pet = pet.astype('float32')
pet = pet**(1/5)
pet_img = cm.hot(pet)
pet_img[..., -1] = pet
mask_outline = get_outline(mask, [1, 0, 0])
true_mask_outline = get_outline(true_mask, [0, 1, 0])
fig, sub = plt.subplots(figsize=(5, 5))
sub.axis('off')
sub.imshow(ct, cmap='gray')
sub.imshow(pet_img)
sub.imshow(mask_outline)
sub.imshow(true_mask_outline)
return fig
def bloch(ex_values, tlist, ops):
"""Plots expectation values of sx, sy and sz on the Bloch sphere.
If one of these operators is not calculated, zeros are passed for that operator.
Parameters:
-----------
ex_values : qutip.Result.expect or list
expectation values per operator
tlist : list, or numpy array
times at which expectation values are evaluated
ops : list
operators of which the expectation values are found in ex_values
Remark:
-------
Does not plot a color bar yet. The lines from the QuTiP tutorial (https://nbviewer.jupyter.org/github/qutip/qutip-notebooks/blob/master/examples/bloch_sphere_with_colorbar.ipynb), commented out down here, give an error which I have not been able to solve yet.
"""
if 'sx' in ops:
sx_exp = ex_values[ops.index('sx')]
elif 'sx' not in ops:
sx_exp = np.zeros(len(list(tlist)))
if 'sy' in ops:
sy_exp = ex_values[ops.index('sy')]
elif 'sy' not in ops:
sy_exp = np.zeros(len(list(tlist)))
if 'sz' in ops:
sz_exp = ex_values[ops.index('sz')]
elif 'sz' not in ops:
sz_exp = np.zeros(len(list(tlist)))
b = Bloch()
b.add_points([sx_exp, sy_exp, sz_exp], 'm')
nrm = Normalize(-2,10)
colors = cm.hot(nrm(tlist))
b.point_color = list(colors)
b.point_marker = ['o']
# TODO: Plot color bar
# left, bottom, width, height = [0.98, 0.05, 0.05, 0.9]
# ax2 = b.fig.add_axes([left, bottom, width, height])
# ColorbarBase(ax2, cmap=cm.hot, norm=nrm, orientation='vertical')
b.show()
def fix_opacity_and_color_map(x, opacity_thresh=0.1, max_opacity=0.8):
""""Given a 2d image, x, get rgbt values for each pixel and modify
opacity values"""
tmp = cm.hot(x)
for i in xrange(tmp.shape[0]):
for j in xrange(tmp.shape[1]):
# XXXXXXXXXXXX
tmp[i,j][0] *= 2
tmp[i,j][1] *= 2
tmp[i,j][2] *= 2
# ############
if x[i,j] > opacity_thresh:
tmp[i,j][3] = max_opacity
else:
tmp[i,j][3] = max_opacity * x[i,j] / opacity_thresh
return tmp
def plot_one_histo(data, sequence, x_maximum):
#x is a list
a = np.arange(0.0, len(data))
#pylab.hist(hist['Data'], bins=600, range=(0,600), histtype='step', align='mid', label=hist['Sequence'])
j=0
y_maximum = 0
for i in data:
c = cm.hot(a[j]/(len(data)+1), 1)
n, bins, patches = pylab.hist(i, bins=600, range=(0,600), histtype='step', label=str(sequence[j]), color=c)
if max(n) > y_maximum:
y_maximum = max(n)
#pylab.title('Read Length Histogram Barcode Sequence %s' % seq)
j=j+1
pylab.axis([0, x_maximum+100, 0, y_maximum+100])
pylab.title("Read Length Histogram for Barcodes")
pylab.legend(loc="upper right", ncol=1, labelspacing=0.01)
pylab.xlabel('Read Length')
pylab.ylabel('Count')
pylab.savefig("ion_readLenHisto.png")
pylab.grid(True)
def draw_circle(axes_handle, centers_x, centers_y, radii, scores=None):
axes_handle.hold(True)
n = 100 # number of samples for drawing circles
if scores is None:
colors = np.repeat(np.reshape([0, 0, 1], (1, 3)), 10, axis=0)
else:
(smin, smax) = (np.min(scores), np.max(scores))
# create color map
colors = cm.hot(np.arange(0, 256))
colors = colors[63 : 192, 0:3]
# use scores as indices into colormap
scores = np.int64(np.ceil(127. * ((scores - smin) / (smax - smin))))
colors = colors[scores, :]
# iterate over all circles that are passed
for x, y, radius, color in zip(centers_x, centers_y, radii, colors):
angles = np.arange(0, (2. - 2. / n) * pi, 1. / n)
(xs, ys) = (radius * np.cos(angles) + x, radius * np.sin(angles) + y)
axes_handle.plot(xs, ys, color=tuple(color), linewidth=2)
fh.canvas.draw()
def _cm_plot(data,headings,dates,ave_data,fig_num,colour_col,
y_col,run_ave_points,n_plot_row,n_plot_col,n_plot_num):
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.cm as cm
depth = np.max(data[:,colour_col]) - np.min(data[:,colour_col])
c_map = cm.hot(np.arange(256))
Z = [[0,0],[0,0]]
levels = range(int(np.min(data[:,colour_col])),
int(np.max(data[:,colour_col])),1)
mappl = plt.contourf(Z,levels,cmap=cm.hot)
plt.figure(fig_num)
plt.subplot(n_plot_row,n_plot_col,n_plot_num)
for i in range(len(data)):
plot1 = plt.plot(dates[i],data[i,y_col])
cm_point = np.floor(((np.max(data[:,colour_col]) -
data[i,colour_col]) / depth) * 255)
plt.setp(plot1,linestyle = 'none',marker = '.')
plt.setp(plot1,color = c_map[255 - cm_point,])
plt.grid(True)
plt.title('%s on %s'%(headings[colour_col],headings[y_col]))
plt.ylabel('%s'%(headings[y_col]))
plt.axis([min(dates),max(dates),min(data[:,y_col]),max(data[:,y_col])])
plt.xticks(rotation = 90)
ax = plt.gca()
ax.xaxis.label.set_color([0.75,0.75,0.75])
ax.tick_params(axis='x', colors=[0.75,0.75,0.75])
# fig.patch.set_facecolor([1,0,0])
ax.patch.set_facecolor([0.9,0.9,0.9])
fig = plt.gcf()
fig.subplots_adjust(top=0.95)
fig.subplots_adjust(bottom=0.12)
fig.subplots_adjust(left=0.08)
fig.subplots_adjust(right=0.97)
plt.colorbar(mappl)
def drawcountry(self,
ax,
base_map,
iso2,
color,
alpha = 1):
if iso2 not in self.countries:
raise ValueError, "Where is that country ?"
vertices = self.countries[iso2]
shape = []
for vertex in vertices:
longs, lats = zip(*vertex)
# conversion to plot coordinates
x,y = base_map(longs, lats)
shape.append(zip(x,y))
lines = LineCollection(shape,antialiaseds=(1,))
lines.set_facecolors(cm.hot(np.array([color,])))
lines.set_edgecolors('white')
lines.set_linewidth(0.5)
lines.set_alpha(alpha)
ax.add_collection(lines)
def add_pizza():
ax = fig.add_axes([0.025, 0.075, 0.3, 0.85], polar=True, resolution=50)
ax.axesPatch.set_alpha(axalpha)
ax.set_axisbelow(True)
N = 8
arc = 2. * np.pi
theta = np.arange(0.0, arc, arc/N)
radii = 10 * np.array([0.79, 0.81, 0.78, 0.77, 0.79, 0.78, 0.83, 0.78])
width = np.pi / 4 * np.array([1.0]*N)
theta = theta[1:]
radii = radii[1:]
width = width[1:]
bars = ax.bar(theta, radii, width=width, bottom=0.0)
for r, bar in zip(radii, bars):
bar.set_facecolor(cm.hot(r/10.))
bar.set_edgecolor('r')
bar.set_alpha(0.6)
for label in ax.get_xticklabels() + ax.get_yticklabels():
label.set_visible(False)
for line in ax.get_ygridlines() + ax.get_xgridlines():
line.set_lw(0.8)
line.set_alpha(0.9)
line.set_ls('-')
line.set_color('0.5')
# add some veggie peperoni
#theta = np.array([.08,.18,.32,.46,.54,.68,.77,.85,.96]) * np.pi * 2.0
#radii = 10*np.array([.6,.38,.58,.5,.62,.42,.58,.67,.45])
theta = np.array([.18,.32,.46,.54,.68,.77,.85,.96]) * np.pi * 2.0
radii = 10*np.array([.38,.58,.5,.62,.42,.58,.67,.45])
c = plt.scatter(theta,radii,c='r',s=7**2)
c.set_alpha(0.75)
ax.set_yticks(np.arange(1, 9, 2))
ax.set_rmax(9)
def drawConfigurationGraphs(self, graphData, variable, keys=minimalDataKeys, normalize = True):
data = [{k:v1 for k,v1 in v[1].items() if k in keys} for v in graphData]
dataConfigs = [v[0] for v in graphData]
maxVals = self.getMaxVals(data)
#is the data a float (percent) or int (absolute value)
isIntType = {k: isinstance(v, int) for k,v in data[0].items()}
if normalize:
relativeVals = [{k: (v*1.0/maxVals[k] if maxVals[k] != 0 else 0.0) if isIntType[k] else v for k,v in item.items()} for item in data]
else:
relativeVals = data
N = len(keys)
Ndata = len(data)
Fdata = float(Ndata)
ind = np.arange(N)
width = 0.35
# plot = plt.figure()
fig, ax = plt.subplots()
# fig, ax = plot.subplot()
colors = [cm.hot(i/Fdata,1) for i in np.arange(len(data))]
rects = []
legends = ['{0}: {1}, unified: {2}'.format(variable, item[variable], item['unified']) for item in dataConfigs]
i = 0
for item in relativeVals:
rects.append(ax.bar([idx*(Ndata+1)*width + i*width for idx in ind], [item[k] for k in keys], width, color=colors[i]))
i = i+1
ax.set_xticks([idx*width*(Ndata+1) + Ndata*width/2 for idx in ind])
ax.set_xticklabels(keys, rotation=70)
ax.set_ylabel('Normalized Values')
ax.set_title('Performance by {0}'.format(variable))
leg = ax.legend( (rects[i][0] for i in np.arange(Ndata)), legends, loc='center left',
bbox_to_anchor=(0.98, 0.5), ncol=1, fancybox=True, shadow=True, prop={'size':6})
# plt.savefig(variable+'_in.png', bbox_extra_artists=[leg], bbox_inches='tight', figure=fig)
return (fig, leg)
nest.Simulate(1600.0)
# plot
fig, (ax1, ax2, ax3) = plt.subplots(3, 1, sharex=True)
# get the neuron's membrane potential
for i in range(num_neurons):
dmm = nest.GetStatus(multimeter)[i]
da_voltage = dmm["events"]["V_m"]
voltage_precise[i] = da_voltage
if i == num_neurons - 1:
for j in range(num_neurons - 1):
voltage_precise[j] -= da_voltage
da_adapt = dmm["events"]["w"]
da_syn = dmm["events"]["I_ex"]
da_time = dmm["events"]["times"]
ax1.plot(da_time, da_voltage, c=cm.hot(0.8 * i / float(num_neurons)), label=models[i])
ax1.set_ylabel("Voltage (mV)")
ax2.plot(da_time, da_adapt, c=cm.hot(0.8 * i / float(num_neurons)), label=models[i])
ax2.set_ylabel("Current (pA)")
ax3.plot(da_time, da_syn, c=cm.hot(0.8 * i / float(num_neurons)), label=models[i])
ax3.set_xlabel("Time (ms)")
ax3.set_ylabel("Conductance (nS)")
plt.legend(loc=4)
# ~ lst_di_param = [ nest.GetStatus(neuron)[0] for neuron in lst_neurons ]
# ~ for di in lst_di_param:
# ~ b_equal = True
# ~ for key,val in lst_di_param[-1].iteritems():
# ~ if key not in tpl_ignore:
# ~ b_equal *= (val == di[key])
def ellipse(data): x,y,a,b,theta,color = data R,G,B, _ = cm.hot(color) params = (x, y, a, b, theta, x, y, int(R*225), int(G*225), int(B*225)) return '<ellipse cx="%f" cy="%f" rx="%f" ry="%f" transform="rotate(%f %f %f)" fill="#%02x%02x%02x"/>'%params
# Create array to calculate d(alpha)/dt to do coarse Euler integration.
alphaDerivatives = np.zeros((3,))
# Fill array with d(alpha)/dt values for each alphas.
alphaDerivatives[:] = (a2-a1)/smallTstep
# Find next alpha by an Euler step.
nextAlpha = a2 + alphaDerivatives * bigTstep
t += bigTstep
# Lift back to the fine fine (theta) space to resume the fine integration.
theta0 = np.dot(hermitePoly, nextAlpha)
## Plot the theta and alpha histories.
fig = plt.figure()
colors = cm.hot(np.linspace(0, 1, ncsteps))
ax = fig.add_subplot(2, 1, 1)
ax.set_xlabel('$\omega_i(t)$')
ax.set_ylabel(r'$\theta_i(t)$') # The r'... (for "raw") is to prevent Python from interpreting \t as an escape code for a tab character.
bx = fig.add_subplot(2, 1, 2)
bx.set_xlabel('$t$')
bx.set_ylabel(r'$\alpha_k(t)$')
for i, c in zip(range(ncsteps), colors):
ax.scatter(om, totalThetaHistory[:, i], color=c)
ax.plot(om, totalThetaHistory[:, i], color='black')
bx.scatter([times[i]]*nHermite, totalAlphaHistory[:, i], color=c)
# Draw lines through coarse points.
def povray_making_movie(position) :
X_size = 60
# position == 0 ----> dermis
# position == 1 ----> epidermis
# position == 2 ----> top of basal membrane
# position == 3 ----> bottom of basal membrane
#position = 1
if position == 0:
open_file = 'states/indaderma_state_'
open_file_pressure = 'pressure/indaderma_pressure.dat'
write_file = '3D/indaderma_3D_'
if position == 1:
open_file = 'states/indaepidermis_state_'
open_file_pressure = 'pressure/indaepidermis_pressure.dat'
write_file = '3D/indaepidermis_3D_'
if position == 2:
open_file = 'states/top_basal_membrane_state_'
open_file_pressure = 'pressure/top_basal_membrane_pressure.dat'
write_file = '3D/top_basal_membrane_3D_'
if position == 3:
open_file = 'states/bottom_basal_membrane_state_'
open_file_pressure = 'pressure/bottom_basal_membrane_pressure.dat'
write_file = '3D/bottom_basal_membrane_3D_'
filename_upload = open_file_pressure
upload = loadtxt(filename_upload)
stress = empty((len(upload), 1))
for i in range(0,len(upload)) :
stress[i] = upload[i][1]
max_number_of_cells = int(1 + upload[len(upload)-1][0])
print 'max number of cells = ', max_number_of_cells
for num_of_cell in range (0, max_number_of_cells):
print 'num_of_cell = ', num_of_cell
open_file_state = open_file + str(num_of_cell) +'.dat'
with open(open_file_state,'r') as f:
config, cells, links, ghosts, T = pickle.load(f)
stress_partial = empty((num_of_cell + 1, 1))
for i in range (0, num_of_cell+1) :
stress_partial[i] = stress[(num_of_cell+1)*num_of_cell/2+i]
print stress_partial
if len(stress_partial)>1 :
stress_partial = (stress_partial-stress_partial.min())/(stress_partial.max()-stress_partial.min())*0.9+0.1
else :
stress_partial[0] = 0.5
col_array = []
for i in range(0,len(stress_partial)) :
rgb_color = cm.hot(1-stress_partial[i],1.0)
col_array.append(rgb_color)
write_file_3D = write_file + str(num_of_cell) + '.txt'
file_povray=open(write_file_3D,'w')
numero_cancer = 0
for cell in cells:
if cell.type.name == 'tDermal':
color = 'color LimeGreen'
if cell.type.name == 'Epidermal':
color = 'color MediumBlue'
if cell.type.name == 'Basal':
color = 'color Gray20'
if cell.type.name == 'Corneum':
color = 'color MediumVioletRed'
if cell.type.name == 'Cancer':
color = 'color <' + str(col_array[numero_cancer][0][0]) + ',' + str(col_array[numero_cancer][0][1]) + ',' + str(col_array[numero_cancer][0][2]) + '>'
numero_cancer = numero_cancer + 1
s = 'sphere { <0, 0, 0>,' + str(cell.radius) + ' material {texture {pigment {' + color + '} finish { specular specularvalue roughness roughnessvalue reflection phongreflection }}}translate<' + str(cell.pos[0]) + ',' + str(cell.pos[1]) + ', 0.0 >}\n'
file_povray.write(s)
cutoff = X_size/2
for link in links:
if (link.two.pos[0]!=link.one.pos[0]) and (link.two.pos[1]!=link.one.pos[1]):
d12 = link.one.pos-link.two.pos
abs_d12 = norm(d12)
if abs_d12 < cutoff :
color = 'color White'
s = 'cylinder { <' + str(link.one.pos[0]) + ',' + str(link.one.pos[1]) + ', 0.0 >,<' + str(link.two.pos[0]) + ',' + str(link.two.pos[1]) + ', 0.0 >,' + str(0.1) + ' material {texture {pigment {' + color + '} finish { phong phongvalue_cyl phong_size phongsize_cyl reflection phongreflection_cyl}}}}\n'
file_povray.write(s)
file_povray.close
clusters = clusterator.predict(data)
result_df["cluster"] = clusters
# Color code by cluster, marker code by PHA flag
categories = np.unique(result_df["cluster"])
colors = np.linspace(0, 1, len(categories))
colordict = dict(zip(categories, colors))
result_df["color"] = result_df["cluster"].apply(lambda x: colordict[x])
fig, ax = plt.subplots()
xv = 0
yv = 2
for c in colors:
pha_df = result_df[(result_df["pha"] == True) & (result_df["color"] == c)]
if not pha_df.empty:
ax.scatter(pha_df[xv], pha_df[yv], marker='s', s=200, c=cm.hot(c))
# uncomment to get the label for the PHA asteroid
# for i, row in pha_df.iterrows():
# ax.annotate(df['neo_reference_id'][i], xy=row[[xv, yv]], textcoords='data')
# pha_df.plot(kind='scatter', x=xv, y=yv, marker='x', markersize=20, color=cm.hot(c), ax=ax)
non_pha_df = result_df[(result_df["pha"] == False) & (result_df["color"] == c)]
if not non_pha_df.empty:
ax.scatter(non_pha_df[xv], non_pha_df[yv], marker='o', c=cm.hot(c))
# non_pha_df.plot(kind='scatter', x=xv, y=yv, marker='o', color=cm.hot(c), ax=ax)
# 2D
# fig, ax = plt.subplots()
# result_df[(result_df["pha"] == True)].plot(kind='scatter', x=1, y=2, marker='x', color='r', ax=ax)
# result_df[(result_df["pha"] == False)].plot(kind='scatter', x=1, y=2, marker='o', color='b', ax=ax)
def draw_top_image(top):
top_image = np.moveaxis(top, -1, 0)
top_image = cm.hot(np.concatenate(top_image, axis=1))
return cv2.cvtColor(top_image.astype(np.float32), cv2.COLOR_RGB2BGR)
nest.Connect(pn,(neuron[0],))
nest.Simulate(1600.0)
# plot
plt.close("all")
fig, (ax1, ax2) = plt.subplots(2,1,sharex=True)
# get the neuron's membrane potential
for i in range(num_models):
dmm = nest.GetStatus(multimeter)[i]
da_voltage = dmm["events"]["V_m"]
da_adapt = dmm["events"][lst_record[i]]
da_time = dmm["events"]["times"]
ax1.plot(da_time,da_voltage,c=cm.hot(i/float(num_models)), label=models[i])
ax1.set_ylabel('Voltage (mV)')
ax2.plot(da_time,da_adapt,c=cm.hot(i/float(num_models)), label=models[i])
ax2.set_xlabel('Time (ms)')
ax2.set_ylabel(lst_ylabel[i])
plt.legend(loc=4)
#-----------------------------------------------------------------------------#
# Compare network dynamics
#------------------------
#
if network:
# time the simulations for each neural model and network size
w_main = QtGui.QMainWindow()
w_cent = QtGui.QSplitter() #QtGui.QWidget()
w_main.setCentralWidget(w_cent)
#lay = QtGui.QHBoxLayout()
#w_cent.setLayout(lay)
# surface plot
w_surf = gl.GLViewWidget()
w_cent.addWidget(w_surf)
w_surf.setCameraPosition(distance=120)
verts = np.load('/home/duke/tvb-vert.npy')
faces = np.load('/home/duke/tvb-tri.npy')
nc = verts.shape[0]
colors = np.zeros((nc, 4))
colors = cm.hot(np.random.rand(nc))
m_data = gl.MeshData(vertexes=verts, faces=faces, vertexColors=colors)
m_surf = gl.GLMeshItem(meshdata=m_data)
#m_surf.setGLOptions('additive')
w_surf.addItem(m_surf)
# fake data
import h5py
h5 = h5py.File('/home/duke/surfspec.h5')
P = h5['all'][int(sys.argv[1])]
fs = np.r_[:P.shape[-1]*1.0]
lfs = np.log10(fs)
# line plots
split_lines = QtGui.QSplitter(QtCore.Qt.Vertical)
def plot(data, graph_type, file_prefix, label_param): fig = plt.figure() ax = fig.add_subplot(111, projection='3d') if(graph_type == 1):#heatmap alfa beta vs accuracy rto vs rtt title="Aproximacion Rtt vs Rto" ax.set_title(title) ancho_barra_x=([0.1]*len(data.alfas)) ancho_barra_y=([0.1]*len(data.alfas)) pos_inicial_z=([0]*len(data.alfas)) #con el calculo loco de aca abajo, entre 0 y 1 los valores normalizados indican, mientras mas grande el valor, mas cerca RTT y RTO normalized_values = [1-(elem/float(max(data.diffsRttVsRTOProm))) for elem in data.diffsRttVsRTOProm] bar_colors_heat = [cm.hot(heat) for heat in normalized_values] ejex = "Alpha" ejey = "Beta" ejez = "Precision" ax.set_xlabel(ejex) ax.set_ylabel(ejey) ax.set_zlabel(ejez) #parametros: #pos_inicial_x, pos_inicial_y, pos_inicial_z, ancho_barra_x, ancho_barra_y, altura_barra_z, color=color_barras, alpha= transparencia entre 0 y 1 ax.bar3d(data.alfas, data.betas, pos_inicial_z, ancho_barra_x, ancho_barra_y, normalized_values, color=bar_colors_heat, alpha=1) plt.suptitle(label_param) plt.savefig(file_prefix + "_rtt_vs_rto.png") #plt.show() plt.close() if(graph_type == 2): title="Cant. Retransmisiones" ax.set_title(title) ancho_barra_x=([0.1]*len(data.alfas)) ancho_barra_y=([0.1]*len(data.alfas)) pos_inicial_z=([0]*len(data.alfas)) #con el calculo loco de aca abajo, entre 0 y 1 los valores normalizados indican, mientras mas grande el valor, mas cerca RTT y RTO normalized_values = [(elem/float(max(data.retransmissions))) for elem in data.retransmissions] bar_colors_heat = [cm.hot(heat) for heat in normalized_values] ejex = "Alpha" ejey = "Beta" ejez = "Cant. Retransmisiones" ax.set_xlabel(ejex) ax.set_ylabel(ejey) ax.set_zlabel(ejez) #parametros: #pos_inicial_x, pos_inicial_y, pos_inicial_z, ancho_barra_x, ancho_barra_y, altura_barra_z, color=color_barras, alpha= transparencia entre 0 y 1 ax.bar3d(data.alfas, data.betas, pos_inicial_z, ancho_barra_x, ancho_barra_y, data.retransmissions, color=bar_colors_heat, alpha=1) plt.suptitle(label_param) plt.savefig(file_prefix + "_retransmisiones.png") #plt.show() plt.close() if(graph_type == 3): title="Rtx Timeout(RTO)" ax.set_title(title) ancho_barra_x=([0.1]*len(data.alfas)) ancho_barra_y=([0.1]*len(data.alfas)) pos_inicial_z=([0]*len(data.alfas)) #con el calculo loco de aca abajo, entre 0 y 1 los valores normalizados indican, mientras mas grande el valor, mas cerca RTT y RTO normalized_values = [(elem/float(max(data.rtosProm))) for elem in data.rtosProm] bar_colors_heat = [cm.hot(heat) for heat in normalized_values] ejex = "Alpha" ejey = "Beta" ejez = "RTO" ax.set_xlabel(ejex) ax.set_ylabel(ejey) ax.set_zlabel(ejez) #parametros: #pos_inicial_x, pos_inicial_y, pos_inicial_z, ancho_barra_x, ancho_barra_y, altura_barra_z, color=color_barras, alpha= transparencia entre 0 y 1 ax.bar3d(data.alfas, data.betas, pos_inicial_z, ancho_barra_x, ancho_barra_y, data.rtosProm, color=bar_colors_heat, alpha=1) plt.suptitle(label_param) plt.savefig(file_prefix + "_rto.png") #plt.show() plt.close() if(graph_type == 4):#heatmap alfa beta vs accuracy rto vs rtt title="Aproximacion Rtt vs Rto" ax.set_title(title) ancho_barra_x=([0.1]*len(data.delays)) ancho_barra_y=([0.01]*len(data.pErrors)) pos_inicial_z=([0]*len(data.delays)) #con el calculo loco de aca abajo, entre 0 y 1 los valores normalizados indican, mientras mas grande el valor, mas cerca RTT y RTO normalized_values = [1-(elem/float(max(data.diffsRttVsRTOProm))) for elem in data.diffsRttVsRTOProm] bar_colors_heat = [cm.hot(heat) for heat in normalized_values] ejex = "Delay" ejey = "Probabilidad de Error" ejez = "Precision" ax.set_xlabel(ejex) ax.set_ylabel(ejey) ax.set_zlabel(ejez) #parametros: #pos_inicial_x, pos_inicial_y, pos_inicial_z, ancho_barra_x, ancho_barra_y, altura_barra_z, color=color_barras, alpha= transparencia entre 0 y 1 ax.bar3d(data.delays, data.pErrors, pos_inicial_z, ancho_barra_x, ancho_barra_y, normalized_values, color=bar_colors_heat, alpha=1) plt.suptitle(label_param) plt.savefig(file_prefix + "dp_rtt_vs_rto.png") #plt.show() plt.close() if(graph_type == 5): title="Cant. Retransmisiones" ax.set_title(title) ancho_barra_x=([0.1]*len(data.delays)) ancho_barra_y=([0.01]*len(data.pErrors)) pos_inicial_z=([0]*len(data.delays)) #con el calculo loco de aca abajo, entre 0 y 1 los valores normalizados indican, mientras mas grande el valor, mas cerca RTT y RTO normalized_values = [(elem/float(max(data.retransmissions))) for elem in data.retransmissions] bar_colors_heat = [cm.hot(heat) for heat in normalized_values] ejex = "Delay" ejey = "Probabildad de Error" ejez = "Cant. Retransmisiones" ax.set_xlabel(ejex) ax.set_ylabel(ejey) ax.set_zlabel(ejez) #parametros: #pos_inicial_x, pos_inicial_y, pos_inicial_z, ancho_barra_x, ancho_barra_y, altura_barra_z, color=color_barras, alpha= transparencia entre 0 y 1 ax.bar3d(data.delays, data.pErrors, pos_inicial_z, ancho_barra_x, ancho_barra_y, data.retransmissions, color=bar_colors_heat, alpha=1) plt.suptitle(label_param) plt.savefig(file_prefix + "dp_retransmisiones.png") #plt.show() plt.close() if(graph_type == 6): title="Rtx Timeout(RTO)" ax.set_title(title) ancho_barra_x=([0.1]*len(data.delays)) ancho_barra_y=([0.01]*len(data.pErrors)) pos_inicial_z=([0]*len(data.delays)) #con el calculo loco de aca abajo, entre 0 y 1 los valores normalizados indican, mientras mas grande el valor, mas cerca RTT y RTO normalized_values = [(elem/float(max(data.rtosProm))) for elem in data.rtosProm] bar_colors_heat = [cm.hot(heat) for heat in normalized_values] ejex = "Delay" ejey = "Probabilidad de Error" ejez = "RTO" ax.set_xlabel(ejex) ax.set_ylabel(ejey) ax.set_zlabel(ejez) #parametros: #pos_inicial_x, pos_inicial_y, pos_inicial_z, ancho_barra_x, ancho_barra_y, altura_barra_z, color=color_barras, alpha= transparencia entre 0 y 1 ax.bar3d(data.delays, data.pErrors, pos_inicial_z, ancho_barra_x, ancho_barra_y, data.rtosProm, color=bar_colors_heat, alpha=1) plt.suptitle(label_param) plt.savefig(file_prefix + "_rto.png") #plt.show() plt.close()
def plot_vector(ax,uv,x=0,y=0,color=0):
return ax.quiver(x,y,uv[0],uv[1],cm.hot(color,1),angles='xy',scale_units='xy',scale=1)
all_frequencies = np.array(all_frequencies_list) k = np.array(k) r = np.array(r) y_n = np.array(get_yn(all_frequencies)) vars = [1.09, -1.02] #popt, pcov = curve_fit(function, x, y_n) #popt = leastsq(residual, vars, args = (y_n, k, r)) solution = myleastsq(residual, vars, args = (y_n, k, r)) rcParams['font.size'] = 16 intercept = solution[0] slope = solution[1] print intercept print slope fig = figure(count) ax = plt.axes([0.12, 0.10, 0.85, 0.85]) p1 = plot(k_lists[0], freq_bin_1, marker= 'o', linestyle = 'None', color = cm.hot(.1)) p2 = plot(k_lists[1], freq_bin_2, marker= 'o', linestyle = 'None', color = cm.hot(.2)) p3 = plot(k_lists[2], freq_bin_3, marker= 'o', linestyle = 'None', color = cm.hot(.3)) p4 = plot(k_lists[3], freq_bin_4, marker= 'o', linestyle = 'None', color = cm.hot(.4)) p5 = plot(k_lists[4], freq_bin_5, marker= 'o', linestyle = 'None', color = cm.hot(.5)) p6 = plot(k_lists[5], freq_bin_6, marker= 'o', linestyle = 'None', color = cm.hot(.6)) p7 = plot(k_lists[6], freq_bin_7, marker= 'o', linestyle = 'None', color = cm.hot(.7)) p8 = plot(k_lists[7], freq_bin_8, marker= 'o', linestyle = 'None', color = cm.hot(.8)) p9 = plot(k_lists[8], freq_bin_9, marker= 'o', linestyle = 'None', color = cm.hot(.9)) p10 = plot(k_lists[9], freq_bin_10, marker= 'o', linestyle = 'None', color = cm.hot(1)) knew = arange(0,20,.1) r1 = RSA_midpoints[0] r2 = RSA_midpoints[1] r3 = RSA_midpoints[2]
def SpectroPlot(data,nx=100,ny=100,ylabel='incidence angle',zlabel='Amplification',yscale='lin',cmap='rainbow'):
import scipy.interpolate
from matplotlib.colors import LogNorm
from matplotlib.ticker import MaxNLocator
if data['sensitivity']==False:
raise IOError("Sensitivity calculation hasn't been performed! Please do so.")
x = np.asarray(data['x'])
y = np.asarray(data['y'])
z = np.asarray(data['z'])
xi,yi = np.logspace(np.log10(x.min()),np.log10(x.max()),nx), np.linspace(y.min(),y.max(),ny)
xi,yi = np.meshgrid(xi,yi)
# interpolation
zi = scipy.interpolate.griddata((x,y),z,(xi,yi),method='linear')
# plot the data
f = plt.figure(figsize=(10.,5.),dpi=300)
# plot velocity profile
a2= f.add_subplot(1,5,5)
if type(data['hl'][0])==float:
xmin = []
xmax = []
# plot vp
if data['modeID']>=5:
if type(data['vp'][0])==list:
lvp = len(data['vp'][0])
clistvp = cm.cool(np.arange(lvp))
for i in range(len(data['vp'][0])):
vptemp = np.concatenate(([data['vp'][0][i]],data['vp'][1:]))
depthtemp = np.concatenate(([0.],np.cumsum(data['hl'])))
depth = [0.]
vp = [vptemp[0]/1000.]
for j in range(1,len(depthtemp)-1):
depth.append(depthtemp[j])
vp.append(vptemp[j-1]/1000.)
depth.append(depthtemp[j])
vp.append(vptemp[j]/1000.)
depth.append(depth[-1]+0.1*depth[-1])
vp.append(vp[-1])
a2.plot(vp,depth,color=clistvp[i])
xmax.append(np.max(vp))
else:
vptemp = data['vp']
depthtemp = np.concatenate(([0.],np.cumsum(data['hl'])))
depth = [0.]
vp = [vptemp[0]/1000.]
for j in range(1,len(depthtemp)-1):
depth.append(depthtemp[j])
vp.append(vptemp[j-1]/1000.)
depth.append(depthtemp[j])
vp.append(vptemp[j]/1000.)
depth.append(depth[-1]+0.1*depth[-1])
vp.append(vp[-1])
a2.plot(vp,depth,color='b')
xmax.append(np.max(vp))
# plot vs
if type(data['vs'][0])==list:
lvs = len(data['vs'][0])
clistvs = cm.hot(np.arange(lvs))
for i in range(len(data['vs'][0])):
vstemp = np.concatenate(([data['vs'][0][i]],data['vs'][1:]))
depthtemp = np.concatenate(([0.],np.cumsum(data['hl'])))
depth = [0.]
vs = [vstemp[0]/1000.]
for j in range(1,len(depthtemp)-1):
depth.append(depthtemp[j])
vs.append(vstemp[j-1]/1000.)
depth.append(depthtemp[j])
vs.append(vstemp[j]/1000.)
depth.append(depth[-1]+0.1*depth[-1])
vs.append(vs[-1])
a2.plot(vs,depth,color=clistvs[i])
xmin.append(np.min(vs))
xmax.append(np.max(vs))
else:
vstemp = data['vs']
depthtemp = np.concatenate(([0.],np.cumsum(data['hl'])))
depth = [0.]
vs = [vstemp[0]/1000.]
for j in range(1,len(depthtemp)-1):
depth.append(depthtemp[j])
vs.append(vstemp[j-1]/1000.)
depth.append(depthtemp[j])
vs.append(vstemp[j]/1000.)
depth.append(depth[-1]+0.1*depth[-1])
vs.append(vs[-1])
a2.plot(vs,depth,color='r')
xmin.append(np.min(vs))
xmax.append(np.max(vs))
a2.set_xlim(np.min(xmin)-0.05,np.max(xmax)+0.05)
a2.set_ylim(np.min(depth),np.max(depth))
else:
if data['modeID']>=5:
ld = len(data['hl'][0])
clistvp = cm.cool(np.arange(ld))
clistvs = cm.hot(np.arange(ld))
for i in range(len(data['hl'][0])):
hl = np.concatenate(([data['hl'][0][i]],data['hl'][1:]))
depthtemp = np.concatenate(([0.],np.cumsum(hl)))
vptemp = data['vp']
vstemp = data['vs']
depth = [0.]
vs = [vstemp[0]/1000.]
vp = [vptemp[0]/1000.]
for j in range(1,len(hl)):
depth.append(depthtemp[j])
vs.append(vstemp[j-1]/1000.)
vp.append(vptemp[j-1]/1000.)
depth.append(depthtemp[j])
vs.append(vstemp[j]/1000.)
vp.append(vptemp[j]/1000.)
depth.append(depth[-1]+0.1*depth[-1])
vs.append(vs[-1])
vp.append(vp[-1])
a2.plot(vp,depth,color=clistvp[i])
a2.plot(vs,depth,color=clistvs[i])
a2.set_xlim(np.min(vs)-0.05,np.max(vp)+0.05)
a2.set_ylim(np.min(depth),np.max(depth))
else:
ld = len(data['hl'][0])
clistvs = cm.hot(np.arange(ld))
for i in range(len(data['hl'][0])):
hl = np.concatenate(([data['hl'][0][i]],data['hl'][1:]))
vstemp = data['vs']
depthtemp = np.concatenate(([0.],np.cumsum(hl)))
depth = [0.]
vs =[vstemp[0]/1000.]
for j in range(1,len(hl)):
depth.append(depthtemp[j])
vs.append(vstemp[j-1]/1000.)
depth.append(depthtemp[j])
vs.append(vstemp[j]/1000.)
depth.append(depth[-1]+0.1*depth[-1])
vs.append(vs[-1])
a2.plot(vs,depth,color=clistvs[i])
a2.set_xlim(np.min(vs)-0.05,np.max(vs)+0.05)
a2.set_ylim(np.min(depth),np.max(depth))
a2.invert_yaxis()
a2.set_xlabel('Velocity (km/s)')
a2.set_ylabel('Depth (m)')
a2.set_title('Velocity profile')
plt.xticks(rotation='vertical')
# plot data
a = f.add_subplot(1,5,(1,4))
#zi = np.log10(zi)
#z = np.log10(z)
am = a.imshow(zi, vmin=0.1, vmax=z.max(), origin='lower', extent=[x.min(), x.max(), y.min(), y.max()],
aspect = 'auto',norm=LogNorm())
a.set_xlabel('Frequency (Hz)')
a.set_ylabel(ylabel)
a.set_xscale('log')
if yscale=='log':
print yscale
a.set_yscale('log')
a.minorticks_on()
a.tick_params(axis='x', which='major', labelsize=11, labelcolor='k')
a.tick_params(axis='x', which='minor', labelsize=10, labelcolor='grey')
for axis in [a.xaxis]:
axis.set_major_formatter(FuncFormatter(majortickformat))
axis.set_minor_formatter(FuncFormatter(minortickformat))
cb = plt.colorbar(am,label=zlabel)
cb.locator = MaxNLocator(10)
# cb.formatter = ScalarFormatter()
cb.formatter = FuncFormatter(cbtickformat)
cb.update_ticks()
f.tight_layout()