def plot_linreg_internal(alpha, E_batch, mu_P, S_P, n, sigma):
"""
Plots the points of linreg and the line too.
"""
S_P1 = S_P(alpha, E_batch) # cov matrix
mu_P1 = mu_P(alpha, E_batch, S_P1) # mean vector
X, Y = E_batch # unwrap
fig, ax = plt.subplots(figsize=graph_size)
ax.ticklabel_format(useOffset=False, style='plain')
# Hide the right and top spines
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.spines['bottom'].set_visible(False)
ax.spines['left'].set_visible(False)
# Only show ticks on the left and bottom spines
ax.yaxis.set_ticks_position('left')
ax.xaxis.set_ticks_position('bottom')
a = .5
x_grid = np.linspace(np.min(X), np.max(X), num=100)[:, None]
Phi_X = np.hstack((np.ones((x_grid.shape[0], 1)), x_grid))
stdev_pred = np.sqrt(np.sum(np.dot(Phi_X, S_P1) * Phi_X, 1)[:, None])
upper_bound = Phi_X @ mu_P1 + stdev_pred
lower_bound = Phi_X @ mu_P1 - stdev_pred
stdev_pred1 = np.sqrt(
np.sum(np.dot(Phi_X, S_P1) * Phi_X, 1)[:, None] + sigma)
upper_bound1 = Phi_X @ mu_P1 + stdev_pred1
lower_bound1 = Phi_X @ mu_P1 - stdev_pred1
a = .2
plt.fill_between(np.squeeze(x_grid[:, 0]),
np.squeeze(lower_bound1),
np.squeeze(upper_bound1),
color=cm.autumn(.5),
alpha=a,
label='Std dev of predictions')
plt.plot(x_grid,
Phi_X @ mu_P1,
'-',
color=cm.autumn(1),
label='Mean of predictions')
plt.plot(X[:, 1], Y, 'xk', label='Observations')
#plt.ylim(2 * np.min(Y), 2 * np.max(Y))
plt.legend()
def region2rgba(rg, region_id, alpha=0.5):
"""
Paints out the region outline selected via region_ID and converts the resulting 2D array slice to RGBA.
Other than overlay2RGBA() and bg2RGBA(), this method takes a 2D array
as input instead of a 3D array. That is because it is not useful to preprocess the
entire 3D region outline layer, which would have to be redone every time the cross hair
is pointing at another region.
:param rg: The 2D region relevance map.
:param region_id: the region to plot
:param alpha: the transparency
:return: The resulting RGBA array.
:rtype: numpy.ndarray
"""
rgcopy = np.copy(rg)
if region_id != 0:
rgcopy[rgcopy != region_id] = 0
rgcopy[rgcopy == region_id] = 1
else: # crosshair on background
rgcopy[True] = 0
alpha_mask = np.copy(rgcopy)
alpha_mask[alpha_mask == 1] = alpha
img = np.uint8(cm.autumn(rgcopy) * 255)
img[:, :, 3] = np.uint8(alpha_mask * 255)
ret = img.view("uint32").reshape(
img.shape[:2]) # convert to 2D array of uint32
return ret
def _counter_post_per_thread(Thread_list, Post_list, group_n):
fig = plt.figure()
ax = fig.add_subplot(111)
for th_i in range(1, len(Thread_list) + 1):
thread = Thread_list[th_i]
print(" title", thread.title, th_i)
x_times, y_nums = _extract_time(thread.pi_list, Post_list)
if len(x_times) == 0:
print(" this thread has not post from users.")
continue
if group_n == "ALPHA":
ax.plot(x_times, y_nums, label=th_i, color=cm.spring((th_i - 1) / len(Thread_list)))
elif group_n == "BRAVO":
ax.plot(x_times, y_nums, label=th_i, color=cm.summer((th_i - 1) / len(Thread_list)))
elif group_n == "CHARLIE":
ax.plot(x_times, y_nums, label=th_i, color=cm.autumn((th_i - 1) / len(Thread_list)))
else:
ax.plot(x_times, y_nums, label=th_i, color=cm.winter((th_i - 1) / len(Thread_list)))
minutes = mdates.MinuteLocator(interval=5) # 5分間隔で描画
timeFmt = mdates.DateFormatter('%H:%M') # x軸の時刻表示フォーマットの設定
ax.xaxis.set_major_locator(minutes) # 上記の条件をグラフに設定
ax.xaxis.set_major_formatter(timeFmt) # 上記の条件をグラフに設定
plt.xlim(start_t, finish_t) # x軸の範囲を設定
plt.ylim(0, 70) # y軸の範囲を設定
plt.title(group_n)
# plt.legend(loc='upper left') #データの名前を表示
fig.autofmt_xdate() # いい感じにx軸の時刻表示を調節
plt.savefig(fn_path + group_n[0] + "_per_counter_post_no_legend" + ".png")
plt.show()
def __init__(self, I1, I2, positive=None, negative=None, **kwargs):
"""A visualizer for labelling positive and negative regions
Args:
I1, I2: The images to label
Keyword Args:
positive: The initial set of positive correspondences
negative: The initial set of background examples
"""
super(PNLabeller, self).__init__(I1, I2, **kwargs)
self.positive = []
self.negative = []
self.mode = 'P'
# initialize the colormaps
self.pmap = cycle([cm.cool(i) for i in range(0, cm.cool.N, 16)])
self.nmap = cycle([cm.autumn(i) for i in range(0, cm.autumn.N, 16)])
self.cmap = self.pmap
self.color = self.pmap.next()
# redraw the initial set
if positive is not None:
for left, right in positive:
left, right = tuple(left), tuple(right)
self.positive.append((left, right))
self.plot(left, right)
if negative is not None:
for left, self.color in zip(negative, self.nmap):
left = tuple(left)
self.negative.append(left)
self.plot(left)
def get_color(nr: int, all: int):
# some magic to crate a redish hue mixed with some color
fact = float(nr) / float(all)
## a bit fo magic to create a 2-d vector
## needed for numpy colors
res = np.array(cm.autumn(fact))[None, :]
return res
def print_and_plot_results(count, results, verbose, plot_file_name):
print("RPS calculated as 95% confidence interval")
rps_mean_ar = []
low_ar = []
high_ar = []
test_name_ar = []
for test_name in sorted(results):
data = results[test_name]
rps = count / array(data)
rps_mean = tmean(rps)
rps_var = tvar(rps)
low, high = norm.interval(0.95, loc=rps_mean, scale=rps_var**0.5)
times = array(data) * 1000000 / count
times_mean = tmean(times)
times_stdev = tstd(times)
print('Results for', test_name)
print('RPS: {:d}: [{:d}, {:d}],\tmean: {:.3f} μs,'
'\tstandard deviation {:.3f} μs'
.format(int(rps_mean),
int(low),
int(high),
times_mean,
times_stdev))
test_name_ar.append(test_name)
rps_mean_ar.append(rps_mean)
low_ar.append(low)
high_ar.append(high)
if verbose:
print(' from', times)
print()
if plot_file_name is not None:
import matplotlib.pyplot as plt
from matplotlib import cm
fig = plt.figure()
ax = fig.add_subplot(111)
L = len(rps_mean_ar)
color = [cm.autumn(float(c) / (L - 1)) for c in arange(L)]
bars = ax.bar(
arange(L), rps_mean_ar,
color=color, yerr=(low_ar, high_ar), ecolor='k')
# order of legend is reversed for visual appeal
ax.legend(
reversed(bars), reversed(test_name_ar),
loc='upper left')
ax.get_xaxis().set_visible(False)
plt.ylabel('Requets per Second', fontsize=16)
print(plot_file_name)
plt.savefig(plot_file_name, dpi=96)
print("Plot is saved to {}".format(plot_file_name))
if verbose:
plt.show()
def clusterColor(i, norm, permutation, autumn=False):
if i == -2:
return (1, 0, 0, 1) # Waypoints
if i == -1:
return (0, 0, 0, .1) # Outlier
if i == 0:
return (0, 0, 0, 1)
if autumn:
return cm.autumn(norm(permutation[i]))
return cm.rainbow(norm(permutation[i]))
def colorize(self, labelmap):
#print(labelmap.shape)
# Assign a unique color to each label
labelmap = labelmap.astype(np.float32) / self.CONFIG.DATASET.N_CLASSES
if self.autumn:
colormap = cm.autumn(labelmap)[..., :-1] * 255.0
else:
colormap = cm.jet_r(labelmap)[..., :-1] * 255.0
return np.uint8(colormap)
def set_color(c,n):
if c == "rainbow":
color = cm.rainbow(n/8.)
elif c == "gnuplot":
color = cm.gnuplot(n/8.)
elif c == "autumn":
color = cm.autumn(n/8.)
elif c == "cool":
color = cm.cool(n/8.)
return color
def colorfor(v_0_1, testname):
if True:
if is_c_code(testname):
return cm.spectral(v_0_1)
else:
return cm.spectral(v_0_1)
else:
if is_c_code(testname):
return cm.winter(v_0_1)
else:
return cm.autumn(v_0_1)
def main_lidar():
#dset = SimpleGeoDataset('/data/pointCloudStuff/img/DC/2013.utm.tif')
dset = SimpleGeoDataset('/data/dc_tiffs/dc.tif')
dir_ = '/data/pointCloudStuff/pc/dc'
ext = '.las'
#app = PointRenderer(1024//2,1024//2)
app = PointRenderer(1024, 1024)
#app = PointRenderer(1024*2,2*1024)
app.init(True)
for fname in [f for f in os.listdir(dir_) if f.endswith(ext)]:
fname = os.path.join(dir_, fname)
print(' - Opening', fname)
with pylas.open(fname) as fp:
las = fp.read()
stride = 1
'''
pts0 = np.stack((las.x[::stride], las.y[::stride], las.z[::stride]),-1).astype(np.float32)
lo,hi = np.quantile(pts0, [.003,.997], 0)
pts0 = pts0[(pts0[:,0] > lo[0]) & (pts0[:,1] > lo[1]) & (pts0[:,2] > lo[2]) & \
(pts0[:,0] < hi[0]) & (pts0[:,1] < hi[1]) & (pts0[:,2] < hi[2]) ]
pts1 = ((pts0 - lo) / max(hi - lo)) * 2 - 1
#pts = torch.from_numpy(pts1).cuda()
app.pts = pts1
app.colors = (pts0 - lo) / (hi - lo)
'''
pts, colors, img = process_pc_get_pts(las, dset, stride)
pts, density = voxelize.filterOutliers(torch.from_numpy(pts), 12)
density = density.cpu().numpy()
pts = pts.cpu().numpy()
T = 2
pts = pts[density > T]
colors = autumn(density[density > T] / 5)[..., :3].astype(
np.float32)
#colors[density<1.7] = (0,1,0)
#app.pts,app.colors = pts,colors
app.setPts(pts, colors, None)
print(' - Setting', app.pts.shape[0], 'pts')
while not app.q_pressed:
app.startFrame()
app.render()
app.endFrame()
#if app.endFrame(): break
app.q_pressed = False
def view_map(map_data, stores=None, paths=None, points=None, grid=True):
"""
View map with entities like paths, stores, and points. Each
entity argument is a list and the z-order (visibility) increases
with index in the list. Among entities z-order varies as
stores < paths < points.
The map shows passable as yellow and impassable as purple. Each
entity list cycles through an indepedent color map, i.e. each
path in paths will have a different color. All points are
displayed in black color.
Args:
map_data: Boolean numpy array. True for passable, False
for impassable.
paths: List of paths to be shown.
stores: Point stores to be shown.
points: Special points to be displayed.
grid: Display grid. Defaults to True.
Notes:
coordinates: click anywhere on the map to view coordinates
of that point.
"""
if stores:
colors = cm.autumn(np.linspace(0, 1, len(stores)))
for c, store in zip(colors, stores):
x, y = get_x_and_y_from_itr(store)
plt.scatter(x, y, color=c)
if paths:
colors = cm.winter(np.linspace(0, 1, len(paths)))
for c, path in zip(colors, paths):
start, end = path[0], path[-1]
x, y = get_x_and_y_from_itr(path)
plt.scatter(x, y, color=c)
plt.plot(start.x, start.y, marker='x')
plt.plot(end.x, end.y, marker='x')
if points:
x = [point.x for point in points]
y = [point.y for point in points]
plt.scatter(x, y, color='black', marker='o')
m, n = map_data.shape
plt.imshow(map_data)
plt.xticks(np.arange(0.5, n, 1.0), [])
plt.yticks(np.arange(-0.5, m, 1.0), [])
plt.grid(grid)
plt.connect('button_press_event', mouse_move)
plt.show()
return
def main_linear_basis():
parser = argparse.ArgumentParser()
parser.add_argument('model', type=str)
parser.add_argument('initial_geometry', type=str)
parser.add_argument('--use_linear_transform',
action='store_true',
default=False)
parser.add_argument('axis', type=str)
parser.add_argument('rotations', type=str)
parser.add_argument('--in_radians', action='store_true', default=False)
args = parser.parse_args()
args.axis = eval(args.axis)
args.rotations = eval(args.rotations)
vis = VisualiseMesh()
T = load_input_mesh(args.model)
V = load_input_geometry(args.initial_geometry, args.use_linear_transform)
V -= np.mean(V, axis=0)
vis.add_mesh(V,
T,
actor_name='base',
color=(31, 120, 180),
compute_normals=True)
axis = np.asarray(args.axis)
assert axis.ndim == 1 and axis.shape[0] == 3
axis /= norm(axis)
rotations = (args.rotations if args.in_radians else map(
np.deg2rad, args.rotations))
q = map(lambda r: quaternion.quat(axis * r), rotations)
R = map(quaternion.rotationMatrix, q)
V1 = map(lambda r: np.dot(V, np.transpose(r)), R)
n = len(V1)
colors = cm.autumn(np.linspace(0., 1., n, endpoint=True), bytes=True)
for i in xrange(n):
vis.add_mesh(V1[i],
T,
actor_name='rotated_%d' % i,
color=colors[i],
compute_normals=True)
vis.camera_actions(('SetParallelProjection', (True, )))
for actor in vis.actors.keys():
vis.actor_properties(actor, ('SetRepresentation', (3, )))
vis.execute()
def main_compositional_basis():
parser = argparse.ArgumentParser()
parser.add_argument('model', type=str)
parser.add_argument('initial_geometry', type=str)
parser.add_argument('--use_linear_transform',
action='store_true',
default=False)
parser.add_argument('axes', type=str)
parser.add_argument('rotations', type=str)
parser.add_argument('--in_radians', action='store_true', default=False)
args = parser.parse_args()
for key in ['axes', 'rotations']:
setattr(args, key, eval(getattr(args, key)))
vis = VisualiseMesh()
T = load_input_mesh(args.model)
V = load_input_geometry(args.initial_geometry, args.use_linear_transform)
V -= np.mean(V, axis=0)
vis.add_mesh(V,
T,
actor_name='base',
color=(31, 120, 180),
compute_normals=True)
axes = np.asarray(args.axes)
assert axes.shape == (2, 3)
axes = normalise(axes)
rotations = np.atleast_2d(args.rotations)
assert rotations.shape[1] == 2
if args.in_radians:
rotations = np.rad2deg(rotations)
n = len(rotations)
colors = cm.autumn(np.linspace(0., 1., n, endpoint=True), bytes=True)
for i in xrange(n):
x = reduce(axis_angle.axAdd, axes * rotations[i][:, np.newaxis])
q = quaternion.quat(x)
R = quaternion.rotationMatrix(q)
V1 = np.dot(V, np.transpose(R))
vis.add_mesh(V1,
T,
actor_name='rotation_%d' % i,
color=colors[i],
compute_normals=True)
vis.execute()
def __init__(self, video, df_PSFs, median_filter_flag=False, step=1, color='gray', imgSizex=5, imgSizey=5, value=0):
"""
This class displays video in the Jupyter notebook interactively while highlighting PSFs with trajectories.
Parameters
----------
video: NDArray
Input video.
df_PSFs: panda data_frame
Data Frames that contains the location of PSFs.
median_filter_flag: bool
In case it defines as True, a median filter is applied with size 3 to remove hot pixel effect.
color: str
It defines the colormap for visualization.
imgSizex: int
Image length size.
imgSizey: int
Image width size.
IntSlider_width: str
Size of slider.
step: int
Stride between visualization frames.
value: int
Initial frame value for visualization
"""
self.video = video
self.df_PSFs = df_PSFs
self.imgSizex = imgSizex
self.imgSizey = imgSizey
self.df_PSFs = df_PSFs
self.median_filter_flag = median_filter_flag
self.color = color
self.median_filter_flag = median_filter_flag
if self.df_PSFs.particle.isnull().any():
self.df_PSFs['particle'] = 0
self.list_particles_idx = self.df_PSFs.particle.unique()
else:
self.list_particles_idx = self.df_PSFs.particle.unique()
colors_ = cm.autumn(np.linspace(0, 1, len(self.list_particles_idx)))
self.colors = colors_[0:len(self.list_particles_idx), :]
interact(self.show_psf, frame_number=widgets.IntSlider(min=0, max=self.video.shape[0] - 1, step=step, value=value,
readout_format='1', continuous_update=False, description='Frame:'))
def print_and_plot_results(count, results, verbose, plot_file_name):
print("RPS calculated as 95% confidence interval")
rps_mean_ar = []
low_ar = []
high_ar = []
test_name_ar = []
for test_name in sorted(results):
data = results[test_name]
rps = count / array(data)
rps_mean = tmean(rps)
rps_var = tvar(rps)
low, high = norm.interval(0.95, loc=rps_mean, scale=rps_var**0.5)
times = array(data) * 1000000 / count
times_mean = tmean(times)
times_stdev = tstd(times)
print('Results for', test_name)
print('RPS: {:d}: [{:d}, {:d}],\tmean: {:.3f} μs,'
'\tstandard deviation {:.3f} μs'.format(int(rps_mean), int(low),
int(high), times_mean,
times_stdev))
test_name_ar.append(test_name)
rps_mean_ar.append(rps_mean)
low_ar.append(low)
high_ar.append(high)
if verbose:
print(' from', times)
print()
if plot_file_name is not None:
import matplotlib.pyplot as plt
from matplotlib import cm
fig = plt.figure()
ax = fig.add_subplot(111)
L = len(rps_mean_ar)
color = [cm.autumn(float(c) / (L - 1)) for c in arange(L)]
bars = ax.bar(arange(L),
rps_mean_ar,
color=color,
yerr=(low_ar, high_ar),
ecolor='k')
# order of legend is reversed for visual appeal
ax.legend(reversed(bars), reversed(test_name_ar), loc='upper left')
ax.get_xaxis().set_visible(False)
plt.ylabel('Requets per Second', fontsize=16)
print(plot_file_name)
plt.savefig(plot_file_name, dpi=96)
print("Plot is saved to {}".format(plot_file_name))
if verbose:
plt.show()
def showStressMap():
#load data
coords, values, time = loadStressData()
#normalize distribution
values = np.array(values)
values /= (np.mean(values)/0.5)
values = np.clip(values,0,1)
#display data
for i in range(len(coords)-1):
zipped = zip(coords[i], coords[i+1])
#choose appropriate colormap
col = cm.autumn(values[i])
plt.plot(zipped[0],zipped[1],lw=3,color=col)
def _counter_post_per_user(User_list, Post_list, group_n):
fig = plt.figure()
ax = fig.add_subplot(111)
sort_uilist = sorted(User_list.keys())
for u_i in sort_uilist:
if u_i in except_usr_id:
continue
user = User_list[u_i]
print(" user", user.name, u_i)
x_times, y_nums = _extract_time(user.pi_list, Post_list)
if len(x_times) == 0:
print(" this thread has not post from users.")
continue
if group_n == "ALPHA":
ax.plot(x_times, y_nums, label=user.name, color=cm.spring((u_i - 1) / len(User_list)))
ax.plot(x_times[-1], y_nums[-1], marker='o', color=cm.spring((u_i - 1) / len(User_list)))
elif group_n == "BRAVO":
ax.plot(x_times, y_nums, label=user.name, color=cm.summer((u_i - 1) / len(User_list)))
ax.plot(x_times[-1], y_nums[-1], marker='o', color=cm.summer((u_i - 1) / len(User_list)))
elif group_n == "CHARLIE":
ax.plot(x_times, y_nums, label=user.name, color=cm.autumn((u_i - 1) / len(User_list)))
ax.plot(x_times[-1], y_nums[-1], marker='o', color=cm.autumn((u_i - 1) / len(User_list)))
else:
ax.plot(x_times, y_nums, label=user.name, color=cm.winter((u_i - 1) / len(User_list)))
ax.plot(x_times[-1], y_nums[-1], marker='o', color=cm.winter((u_i - 1) / len(User_list)))
days = mdates.MinuteLocator(interval=5)
daysFmt = mdates.DateFormatter('%H:%M')
ax.xaxis.set_major_locator(days)
ax.xaxis.set_major_formatter(daysFmt)
plt.xlim(start_t, finish_t)
plt.ylim(0, 40)
plt.title(group_n)
plt.legend(loc='upper left')
fig.autofmt_xdate()
plt.savefig(fn_path + group_n[0] + "_per_user_post_counter" + ".png")
plt.show()
def makePlots(arr):
clips = []
clipi = []
smooths = []
smoothi = []
for t in list(set(arr)):
subset = stars[arr == t]
clispec, clivar = sigclip_coadd(subset, sigLim=3.0)
clips.append(clispec)
clipi.append(clivar)
for s, i in zip(clips, clipi):
flux = Spectrum(lbin, s, ivar=i)
flux.smooth(width=5, filtertype="gaussian", replace=True)
smooths.append(flux.flux)
smoothi.append(flux.ivar)
uniqarr = list(set(arr))
sortInd = np.argsort(uniqarr)
n = len(uniqarr)
fig = py.figure(1, (9, 14))
ax = fig.add_subplot(111)
offset = 0
weight = np.ones(len(lbin))
weight[(7550 < lbin) & (lbin < 7700)] = np.nan
for i in sortInd[1:]:
color = cmp.autumn(i / np.float(n))
ax.plot(lbin, weight * smooths[i] + offset, lw=2.0, color=color)
offset += 1
ax.set_xlim(5500, 9500)
ax.set_ylim(0, 21)
ax.set_xlabel("$\lambda$")
ax.axes.get_yaxis().set_visible(False)
fig = py.figure(1, (15, 5))
ax = fig.add_subplot(111)
for i in sortInd:
color = cmp.cool(i / float(n))
ax.plot(lbin, weight * smooths[i], color=color)
ax.set_xlim(5500, 9500)
def __init__(self,
video,
df_PSFs,
time_delay=0.1,
GUI_progressbar=False,
*args,
**kwargs):
"""
This class displays video while highlighting PSFs.
Parameters
----------
video: NDArray
Input video.
df_PSFs: panda data_frame
Data Frames that contains the location of PSFs.
time_delay: float
Delay between frames in milliseconds.
GUI_progressbar: bool
This actives GUI progress bar
"""
self.cpu = CPUConfigurations()
super(DisplayDataFramePSFsLocalization, self).__init__()
self.video = video
self.time_delay = time_delay
self.df_PSFs = df_PSFs
self.pressed_key = {}
self.list_line = []
if 'particle' in self.df_PSFs.keys():
self.list_particles_idx = self.df_PSFs.particle.unique()
else:
self.df_PSFs['particle'] = 0
self.list_particles_idx = self.df_PSFs.particle.unique()
self.GUI_progressbar = GUI_progressbar
self.args = args
self.kwargs = kwargs
colors_ = cm.autumn(np.linspace(0, 1, len(self.list_particles_idx)))
self.colors = colors_[0:len(self.list_particles_idx), :]
self.obj_connection = SignalConnection()
def getLightColors(_log, LightPer):
try:
_log.info("lightColors(LightPer) =" + str(LightPer))
#normalize item number values to colormap
norm = matplotlib.colors.Normalize(vmin=0, vmax=100)
#colormap possible values = viridis, jet, spectral
#rgba_color = cm.RdBu(norm(LightPer),bytes=True)
rgba_color = cm.autumn(norm(int(LightPer)), bytes=True)
_log.info("rgba_color =" + str(rgba_color))
c = Color(rgb=(rgba_color[0] / 255, rgba_color[1] / 255,
rgba_color[2] / 255))
_log.info("getLightColors =" + str(c))
except:
c = '#00FF00'
return c
def plot_star_3D():
#THIS IS NOT WORKING, ITS A WIP.
import numpy as np
import matplotlib.pyplot as plt
from matplotlib import cm, colors
from mpl_toolkits.mplot3d import Axes3D
from scipy.special import sph_harm #import package to calculate spherical harmonics
import pdb
theta = np.linspace(0, 2 * np.pi, 100) #setting range for theta
phi = np.linspace(0, np.pi, 100) #setting range for phi
phi, theta = np.meshgrid(phi, theta) #setting the grid for phi and theta
#Setting the cartesian coordinates of the unit sphere
#Converting phi, theta, z to cartesian coordinates
x = np.sin(phi) * np.cos(theta)
y = np.sin(phi) * np.sin(theta)
z = np.cos(phi)
#Setting the aspect ratio to 1 which makes the sphere look spherical and not elongated
fig = plt.figure(figsize=plt.figaspect(1.)) #aspect ratio
axes = fig.add_subplot(111, projection='3d') #sets figure to 3d
fig.suptitle('m=4 l=4', fontsize=18, x=0.52, y=.85)
m, l = 4, 4 #m and l control the mode of pulsation and overall appearance of the figure
#Calculating the spherical harmonic Y(l,m) and normalizing it
axes.view_init(30, 45)
plt.ion()
plt.show()
for idx, angle in enumerate(np.linspace(0, 360, 20)):
figcolors = sph_harm(m, l, theta + angle, phi).real
figmax, figmin = figcolors.max(), figcolors.min()
figcolors = (figcolors - figmin) / (figmax - figmin)
#Sets the plot surface and colors of the figure where seismic is the color scheme
axes.plot_surface(x,
y,
z,
rstride=1,
cstride=1,
facecolors=cm.autumn(figcolors))
fig.canvas.draw_idle()
pdb.set_trace()
def animate(i):
global res_lcs, com_lcs, ax, scat, fig
print i
fig.suptitle('Timestep: '+str((i+1)*36))
res_data = res_lcsnorm[:,(i+1)*36-36:(i+1)*36+36,:].mean(1)
com_data = com_lcsnorm[:,(i+1)*36-36:(i+1)*36+36,:].mean(1)
res_newdata = ternaryPlot(res_data)
com_newdata = ternaryPlot(com_data)
res_scat.set_offsets(res_newdata)
com_scat.set_offsets(com_newdata)
res_colors = cm.autumn(res_data.sum(1))
res_colors[:,3] = res_data.sum(1) #set opacity equal to intensity (currently not working)
res_scat.set_facecolors(res_colors)
com_colors = cm.winter(com_data.sum(1))
com_colors[:,3] = com_data.sum(1)
com_scat.set_facecolors(com_colors)
return res_scat, com_scat
plt.figure()
ax=plt.gca()
zbarr=0.5*(z0bins+zfbins)
ax.imshow([[0.,1.],[0.,1.]],extent=[2.0,2.24,-0.35,-0.30],interpolation='bicubic',cmap=cm.winter,aspect='auto')
ax.imshow([[0.,1.],[0.,1.]],extent=[2.0,2.24,-0.42,-0.37],interpolation='bicubic',cmap=cm.autumn,aspect='auto')
plt.text(1.7,-0.34,'$1^{\\rm st}\\,\\,{\\rm mode}$',{'fontsize':16})
plt.text(1.7,-0.41,'$2^{\\rm nd}\\,\\,{\\rm mode}$',{'fontsize':16})
for i in (1+np.arange(199))*10 :
ax.plot(zbarr,e_o[ i,:,0]*np.sqrt(ndens)/np.sqrt(np.sum(e_o[ i,:,0]**2*ndens)),'o-',markeredgewidth=0,
color=cm.winter((i+0.5)/2001))
if e_o[i,2,1]>0 :
sign=-1
else :
sign=1
ax.plot(zbarr,sign*e_o[ i,:,1]*np.sqrt(ndens)/np.sqrt(np.sum(e_o[ i,:,1]**2*ndens)),'o-',
markeredgewidth=0,color=cm.autumn((i+0.5)/2001))
plt.xlabel('$z_\\alpha$',fontsize=18)
plt.ylabel('$\\sqrt{\\bar{n}^\\alpha}\\,({\\sf E}_\\ell)^1_\\alpha$',fontsize=18)
plt.xlim([0.5,2.3])
plt.savefig("../Draft/Figs/kl_modes_wl.pdf",bbox_inches='tight')
fisher=(larr+0.5)[:,None]*(c_p_dw0/c_p_fid)**2
fish_permode=np.sum(fisher,axis=0)
fish_cum=np.cumsum(fish_permode)
print fish_cum
print fish_permode
plt.figure();
imodes=np.arange(nbins)+1
plt.plot(imodes,fish_permode/np.sum(fish_permode),'go-',lw=2,label='${\\rm Information\\,\\,in\\,\\,mode}\\,\\,p_{\\rm KL}$',markeredgewidth=0)
plt.plot(imodes[:-1],1-fish_cum[:-1]/fish_cum[-1],'ro-',lw=2,label='${\\rm Information\\,\\,in\\,\\,modes}\\,\\,>p_{\\rm KL}$',markeredgewidth=0)
def ikkKO(AA, AB, AC, kv, nftot, **kwargs):
TR = kwargs.get('TR', 1)
start = kwargs.get('start', 2)
ikk = kwargs.get('ikk')
###INITIATION#######
sol = init(AA, AB, 1, kv, nftot, 0)
###SIMULATION#######
Y0 = sol.y[:, -1]
sol1 = sim(AA, AB, 1, kv, TR, Y0, 60 * 60 * 24)
###PROCESSING#######
SOL = prc.fuse(sol.t, sol.y, sol1.t, sol1.y)
pt = np.stack(prc.hour(SOL[0]))
py = np.stack(SOL[1:SOL.shape[0]])
KO = False
time = None
for i, j in enumerate(sol1.t):
if j / 3600 >= start:
KO = i
time = j / 3600
break
py0 = sol1.y[1, KO] ###.00128 1.8; .00125 24###
py1 = sol1.y[6, KO] * 1000
pt1 = sol1.t[KO] / 3600
y0 = sol1.y[:, KO]
y0[1] = 0
sol3 = sim(AA, AB, AC, kv, TR, y0, (24 - start) * 60 * 60, ikk=True)
y0[1] = py0 / 2
sol4 = sim(AA, AB, AC, kv, TR, y0, (24 - start) * 60 * 60, ikk=True)
y0[1] = py0 / 4
sol5 = sim(AA, AB, AC, kv, TR, y0, (24 - start) * 60 * 60, ikk=True)
y0[1] = py0
sol6 = sim(AA, AB, AC, kv, TR, y0, (24 - start) * 60 * 60, ikk=True)
y0[1] = 2 * py0
sol7 = sim(AA, AB, AC, kv, TR, y0, (24 - start) * 60 * 60, ikk=True)
plt.style.use('seaborn-paper')
MAP = np.linspace(0, 1, 6)
fig = plt.figure()
gs = GridSpec(9, 1, hspace=0.1)
ax = fig.add_subplot(gs[:8, :])
ax3 = fig.add_subplot(gs[-1, :])
colours1 = cm.winter(MAP)
ax.plot(pt - 101, 1000 * py[6], c=(0, 0, 128 / 255))
ax.plot(time + sol7.t / 3600,
1000 * sol7.y[6],
linestyle=(0, (5, 1)),
c=colours1[0])
ax.plot(time + sol6.t / 3600,
1000 * sol6.y[6],
linestyle=(0, (5, 5)),
c=colours1[1])
ax.plot(time + sol4.t / 3600, 1000 * sol4.y[6], '-.', c=colours1[2])
ax.plot(time + sol5.t / 3600,
1000 * sol5.y[6],
linestyle=(0, (3, 1, 1, 1, 1, 1)),
c=colours1[3])
ax.plot(time + sol3.t / 3600,
1000 * sol3.y[6],
c=colours1[4],
linestyle=(0, (1, 1)))
handles = [
Line2D([0], [0], color='black', lw=1, linestyle='-'),
Line2D([0], [0], color='black', lw=1, linestyle=(0, (5, 1))),
Line2D([0], [0], color='black', lw=1, linestyle=(0, (5, 5))),
Line2D([0], [0], color='black', lw=1, linestyle='-.'),
Line2D([0], [0],
color='black',
lw=1,
linestyle=(0, (3, 1, 1, 1, 1, 1))),
Line2D([0], [0], color='black', lw=1, linestyle=(0, (1, 1)))
]
PY0 = 1000 * py0
labels = [
'wild-type', 'IKK={} nM'.format(np.around(2 * PY0, decimals=1)),
'IKK={} nM'.format(np.around(PY0, decimals=1)),
'IKK={} nM'.format(np.around(PY0 / 2, decimals=1)),
'IKK={} nM'.format(np.around(PY0 / 4, decimals=1)), 'IKK=0 nM'
]
ax.legend(handles,
labels,
loc=0,
fontsize=10,
framealpha=0,
prop={'size': 10},
edgecolor=None)
ax.set_xlim(-1, 8)
ax.set_xticks([])
ax.set_ylabel('$c$ in nM')
ax2 = ax.twinx()
ax2.set_yticks([])
ax2.set_ylabel('NF$\kappa$B')
colours2 = cm.autumn(MAP)
ax3.plot(pt - 101, 1000 * py[1], c=(170 / 255, 0, 0))
ax3.plot(time + sol7.t / 3600,
1000 * sol7.y[1],
linestyle=(0, (5, 1)),
c=colours2[0])
ax3.plot(time + sol6.t / 3600,
1000 * sol6.y[1],
linestyle=(0, (5, 5)),
c=colours2[1])
ax3.plot(time + sol4.t / 3600, 1000 * sol4.y[1], '-.', c=colours2[2])
ax3.plot(time + sol5.t / 3600,
1000 * sol5.y[1],
linestyle=(0, (3, 1, 1, 1, 1, 1)),
c=colours2[3])
ax3.plot(time + sol3.t / 3600,
1000 * sol3.y[1],
c=colours2[4],
linestyle=(0, (1, 1)))
ax3.set_ylim(-.1, 3)
ax3.set_yticks([0, 1000 * py0, 2000 * py0])
ax3.yaxis.set_major_formatter(FormatStrFormatter('%.1f'))
ax3.set_xlabel('$t$ in h')
ax4 = ax3.twinx()
ax4.set_yticks([])
ax4.set_ylabel('IKK')
trans = ax3.transAxes + ax.transData.inverted()
((xmin, _), (xmax, _)) = trans.transform([[0, 1], [1, 1]])
ax3.set_xlim(xmin, xmax)
ax3.axvline(pt1, linewidth=0.5, c='gray')
ax.axvline(pt1, linewidth=0.5, c='gray')
fig.tight_layout()
plt.savefig('../../graphics/IKKvar.png', dpi=500)
plt.close()
return [pt, py, sol1, sol.t.shape[0], sol3, time]
data = glob.glob('/backup/yuliya/v30/graphs_largedomain/fixed/*.gpickle')
data.sort()
d1.append(data1)
d.append(data)
data1 = glob.glob('/backup/yuliya/v35/breakups_con/fixed/*.npy')
data1.sort()
data = glob.glob(
'/backup/yuliya/v35/graphs_largedomain/fixed/*_no_term_br.gpickle')
data.sort()
d1.append(data1)
d.append(data)
colors = cm.autumn(np.linspace(0, 1, 114))
colors1 = cm.winter(np.linspace(0, 1, 12))
t = np.load('/backup/yuliya/vsi01/vsi01_TimeInSec.npy')
u = 0
ell = np.load('/backup/yuliya/vsi05/dataellg_ellp.npy')[8:-10]
ellv34 = np.load('/backup/yuliya/v34/dataellg_ellp.npy')[13:-1][::1]
ellv34[7:-1] = ellv34[8:]
ellv34[10:-1] = ellv34[11:]
ellv34[45:-1] = ellv34[46:]
ellvsi05 = np.load('/backup/yuliya/vsi05/dataellg_ellp.npy')[8:-10]
ellvsi05[19:-1] = ellvsi05[20:]
ellv30 = np.load('/backup/yuliya/v30/dataellg_ellp.npy')[10:-1]
ellv30[61:-1] = ellv30[62:]
ellv30[63:-1] = ellv30[64:]
ellv35 = np.load('/backup/yuliya/v35/dataellg_ellp.npy')[22:-1]
ellv35[4:-1] = ellv35[5:]
# ======
# this figure shows the time levels in each run
figTimes = plt.figure(30, facecolor='w')
axTimes = figTimes.add_subplot(1, 1, 1)
plt.xlabel('Year')
axTimesYlabels = []
# =========
print "Done setting up figure axes."
# =========
# --- Define colors for lines ---
#colors = ['tab:blue', 'tab:orange', 'tab:green', 'tab:red', 'tab:purple', 'tab:brown', 'tab:pink', 'tab:olive', 'tab:cyan']
n150 = sum("amp150" in r for r in runs)
colors = [ cm.autumn(x) for x in np.linspace(0.0, 1.0, n150) ]
n300 = sum("amp300" in r for r in runs)
colors.extend( [ cm.winter(x) for x in np.linspace(0.0, 1.0, n300) ] )
color_index = 0
# ================
# Loop over runs and plot data
# ================
runNumber = 0
for run in runs:
print "Plotting run: " + run
thisRun = runData[run]
# Pull needed data for plotting from this run
# TODO: replace local variables below in plotting commands with reference to object variables
def plotkSpaceTraj(self, contdraw, view3d):
if not self.active:
return
fgc = 'white' # Foreground color
bgc = 'black' # Background color
self.fig.clear()
plt.figure(self.fig.number) # Workaround to set current figure
gridspec = plt.GridSpec(1, 50) # Ratio Figure Width / Colorbar Width
if not view3d:
ax = plt.subplot(gridspec[0, :-1])
else:
ax = plt.subplot(gridspec[0, :-1], projection='3d')
C = cm.autumn(np.linspace(0, 1, np.size(self.J)))[::-1, :]
# Draw Actual Trajectory
for j in range(np.size(self.J)):
if j < np.size(self.J) - 1:
n_end = self.J[j + 1] - 2 + contdraw
else:
n_end = np.size(self.KX) - 1
n_all = np.arange(start=self.J[j] + 1 - contdraw,
stop=n_end,
step=1)
if not view3d:
ax.plot(self.KX[n_all],
self.KY[n_all],
c=C[j, :],
linewidth=0.5)
else:
ax.plot(self.KX[n_all],
self.KY[n_all],
self.KZ[n_all],
c=C[j, :],
linewidth=0.5)
ax.view_init(elev=90, azim=-90)
if (self.KX.min() < self.KX.max()) and (self.KY.min() < self.KY.max()):
ax.axis([
1.1 * self.KX.min(), 1.1 * self.KX.max(), 1.1 * self.KY.min(),
1.1 * self.KY.max()
])
ax.grid(True, color=fgc)
ax.set_aspect('equal')
# Set Labels
labels = ['%s [rad / mm]' % s for s in ['Kx', 'Ky', 'Kz']]
ax.set_xlabel(labels[0], fontsize=15, fontweight='bold', color=fgc)
ax.set_ylabel(labels[1], fontsize=15, fontweight='bold', color=fgc)
if view3d:
ax.set_zlabel(labels[2], fontsize=15, fontweight='bold', color=fgc)
# Add colorbar
if np.size(self.J) > 2:
cax = plt.subplot(gridspec[0, -1])
sm = plt.cm.ScalarMappable(cmap=cm.autumn)
sm._A = [] # fake up the array of the scalar mappable.
h = plt.colorbar(sm, orientation='vertical', cax=cax, ax=ax)
h.set_ticks([])
h.set_label('Late \t\t\t\t Early'.expandtabs(),
fontsize=15,
color=fgc)
# Color elements in foreground color
for child in ax.get_children():
if isinstance(child, spines.Spine):
child.set_color(fgc)
ax.tick_params(axis='x', colors=fgc)
ax.tick_params(axis='y', colors=fgc)
ax.tick_params(axis='z', colors=fgc)
# Color background
self.fig.set_facecolor(bgc)
ax.set_facecolor(bgc)
# Transparent background
if view3d:
ax.w_xaxis.set_pane_color((0.0, 0.0, 0.0, 0.0))
ax.w_yaxis.set_pane_color((0.0, 0.0, 0.0, 0.0))
ax.w_zaxis.set_pane_color((0.0, 0.0, 0.0, 0.0))
self.canvas.draw()
self.cP = None
self.ADCs = None
n = 0
while n <= len(df["x"]) - 1:
function = df["x"][n]**2 + df["y"][n]**2 + df["z"][n]**2
if function < 0.:
function = 0.
color_list.append(function)
elif function > 1.:
function = 1.
color_list.append(function)
else:
color_list.append(function)
n += 1
df["color"] = color_list
# Import data to draw lines (Could be used to draw molecules later on)
data_df = pd.read_excel(r"data.xlsx", sheet_name=0)
# Create the figure
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
norm = mpl.colors.Normalize(vmin=0, vmax=1)
ax.scatter(df['x'], df['y'], df['z'], color=cm.autumn(df["color"]), s=1)
ax.plot(data_df['x'], data_df['y'], data_df['z'])
ax.set_title('3D Color Function')
ax.set_xlabel('X axis')
ax.set_ylabel('Y axis')
ax.set_zlabel('Z axis')
plt.show()
plt.savefig("colorfunction.png")
def plot(self):
""" Generates plots using different calc functions. """
flu_dist = self.calc_FLU()
pam_dist = self.calc_PAM()
ham_dist = self.calc_Ham()
print("Hamming distance:", ham_dist)
# distance calculations that only reference sub matrix if two characters don't match
hyb_flu_dist = self.calc_FLU(hybrid=True)
hyb_pam_dist = self.calc_PAM(hybrid=True)
print("FLU hybrid", hyb_flu_dist)
print("PAM hybrid", hyb_pam_dist)
plt.figure(figsize=(
8, 5)) # set up figure for plotting, width and height in inches
cmap_1 = cm.autumn(np.linspace(0, 1, self.size))
cmap_2 = cm.winter(np.linspace(0, 1, self.size))
# plot standard form of FLU matrix
slope, intercept, r_value, p_value, std_err = stats.linregress(
self.eff, np.log(flu_dist))
plt.scatter(self.eff,
np.log(flu_dist),
c=cmap_1,
alpha=0.8,
label=r"$\mathrm{standard: R^{2} = }$" +
str(round(r_value**2, 3)))
abline(slope, intercept, plt, cmap_1[0])
# plot hybrid form of FLU matrix
slope, intercept, r_value, p_value, std_err = stats.linregress(
self.eff, np.log(hyb_flu_dist))
plt.scatter(self.eff,
np.log(hyb_flu_dist),
marker='*',
c=cmap_2,
alpha=0.8,
label=r"$\mathrm{hybrid: R^{2} = }$" +
str(round(r_value**2, 3)))
abline(slope, intercept, plt, cmap_2[0])
plt.title("FLU Distance Prediction Performance")
plt.xlabel("Observed Efficacy")
plt.ylabel("Relative Distance (log scaled)")
plt.legend()
plt.savefig('Figures/FLU_validation.png')
plt.figure(figsize=(
8, 5)) # set up figure for plotting, width and height in inches
# plot standard form of PAM250 matrix
slope, intercept, r_value, p_value, std_err = stats.linregress(
self.eff, np.log(pam_dist))
plt.scatter(self.eff,
np.log(pam_dist),
c=cmap_1,
alpha=0.8,
label=r"$\mathrm{standard: R^{2} = }$" +
str(round(r_value**2, 3)))
abline(slope, intercept, plt, cmap_1[0])
# plot hybrid form of PAM250 matrix
slope, intercept, r_value, p_value, std_err = stats.linregress(
self.eff, np.log(hyb_pam_dist))
plt.scatter(self.eff,
np.log(hyb_pam_dist),
marker='*',
c=cmap_2,
alpha=0.8,
label=r"$\mathrm{hybrid: R^{2} = }$" +
str(round(r_value**2, 3)))
abline(slope, intercept, plt, cmap_2[0])
plt.title("PAM250 Distance Prediction Performance")
plt.xlabel("Observed Efficacy")
plt.ylabel("Relative Distance (log scaled)")
plt.legend()
plt.savefig('Figures/PAM_validation.png')
plt.figure(figsize=(
8, 5)) # set up figure for plotting, width and height in inches
slope, intercept, r_value, p_value, std_err = stats.linregress(
self.eff, ham_dist)
plt.scatter(self.eff,
ham_dist,
c=cmap_1,
alpha=0.8,
label=r"$\mathrm{R^{2} = }$" + str(round(r_value**2, 3)))
abline(slope, intercept, plt, cmap_1[0])
plt.title("Hamming Distance Prediction Performance")
plt.xlabel("Observed Efficacy")
plt.ylabel("Relative Distance")
plt.legend()
plt.savefig('Figures/Hamming_validation.png')
plt.grid()
ax4VAFrate = fig4.add_subplot(nrow, ncol, 4, sharex=ax4timelag)
plt.xlabel('SLR (mm)')
plt.ylabel('VAF rate (Gt yr$^{-1}$)')
plt.grid()
# =========
print("Done setting up figure axes.")
# =========
# --- Define colors for lines ---
colors = []
#colors = ['tab:blue', 'tab:orange', 'tab:green', 'tab:red', 'tab:purple', 'tab:brown', 'tab:pink', 'tab:olive', 'tab:cyan']
n150 = sum("amp150" in r for r in runs)
colors.extend([cm.autumn(x) for x in np.linspace(0.0, 1.0, n150)])
n300 = sum("amp300" in r for r in runs)
colors.extend([cm.winter(x) for x in np.linspace(0.0, 1.0, n300)])
n05 = sum("amp0." in r for r in runs)
colors.extend([cm.spring(x) for x in np.linspace(0.0, 1.0, n05)])
color_index = 0
# ================
# Loop over runs and plot data
# ================
runNumber = 0
for run in runs:
print("Plotting run: " + run)
thisRun = runData[run]
# Pull needed data for plotting from this run
data1.sort()
data = glob.glob('/backup/yuliya/v30/graphs_largedomain/fixed/*.gpickle')
data.sort()
d1.append(data1)
d.append(data)
data1 = glob.glob('/backup/yuliya/v35/breakups_con/fixed/*.npy')
data1.sort()
data = glob.glob('/backup/yuliya/v35/graphs_largedomain/fixed/*_no_term_br.gpickle')
data.sort()
d1.append(data1)
d.append(data)
colors = cm.autumn(np.linspace(0, 1, 114))
colors1 = cm.winter(np.linspace(0, 1, 12))
t = np.load('/backup/yuliya/vsi01/vsi01_TimeInSec.npy')
u = 0
ell = np.load('/backup/yuliya/vsi05/dataellg_ellp.npy')[8:-10]
ellv34 = np.load('/backup/yuliya/v34/dataellg_ellp.npy')[13:-1][::1]
ellv34[7:-1] = ellv34[8:]
ellv34[10:-1] = ellv34[11:]
ellv34[45:-1] = ellv34[46:]
ellvsi05 = np.load('/backup/yuliya/vsi05/dataellg_ellp.npy')[8:-10]
ellvsi05[19:-1] = ellvsi05[20:]
ellv30 = np.load('/backup/yuliya/v30/dataellg_ellp.npy')[10:-1]
ellv30[61:-1] = ellv30[62:]
ellv30[63:-1] = ellv30[64:]
ellv35 = np.load('/backup/yuliya/v35/dataellg_ellp.npy')[22:-1]
ellv35[4:-1] = ellv35[5:]
def main(
args,
cfg=None,
min_allowed_score=None):
# Setup config
if cfg is None:
cfg = setup_config(args, random_seed=args.random_seed, is_testing=True)
cfg.defrost()
cfg.ACTUAL_TEST_DATASET = args.test_dataset
# Build path to gt instances and inference output
inference_output_dir = get_inference_output_dir(
cfg['OUTPUT_DIR'],
args.test_dataset,
args.inference_config,
args.image_corruption_level)
# Get thresholds to perform evaluation on
if min_allowed_score is None:
# Check if F-1 Score has been previously computed.
try:
with open(os.path.join(inference_output_dir, "mAP_res.txt"), "r") as f:
min_allowed_score = f.read().strip('][\n').split(', ')[-1]
min_allowed_score = round(float(min_allowed_score), 4)
except FileNotFoundError:
# If not, process all detections. Not recommended as the results might be influenced by very low scoring
# detections that would normally be removed in robotics/vision
# applications.
min_allowed_score = 0.0
# get preprocessed instances
preprocessed_predicted_instances, preprocessed_gt_instances = evaluation_utils.get_per_frame_preprocessed_instances(
cfg, inference_output_dir, min_allowed_score)
# get metacatalog and image infos
meta_catalog = MetadataCatalog.get(args.test_dataset)
images_info = json.load(open(meta_catalog.json_file, 'r'))['images']
# Loop over all images and visualize errors
for image_info in images_info:
image_id = image_info['id']
image = cv2.imread(
os.path.join(
meta_catalog.image_root,
image_info['file_name']))
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
v = ProbabilisticVisualizer(
image,
meta_catalog,
scale=1.5)
class_list = v.metadata.as_dict()['thing_classes']
predicted_box_means = preprocessed_predicted_instances['predicted_boxes'][image_id].cpu(
).numpy()
gt_box_means = preprocessed_gt_instances['gt_boxes'][image_id].cpu(
).numpy()
predicted_box_covariances = preprocessed_predicted_instances[
'predicted_covar_mats'][image_id].cpu(
).numpy()
predicted_cls_probs = preprocessed_predicted_instances['predicted_cls_probs'][image_id]
if predicted_cls_probs.shape[0] > 0:
if cfg.MODEL.META_ARCHITECTURE == "ProbabilisticGeneralizedRCNN" or cfg.MODEL.META_ARCHITECTURE == "ProbabilisticDetr":
predicted_scores, predicted_classes = predicted_cls_probs[:, :-1].max(
1)
predicted_entropies = entropy(
predicted_cls_probs.cpu().numpy(), base=2)
else:
predicted_scores, predicted_classes = predicted_cls_probs.max(
1)
predicted_entropies = entropy(
np.stack(
(predicted_scores.cpu().numpy(),
1 - predicted_scores.cpu().numpy())),
base=2)
predicted_classes = predicted_classes.cpu(
).numpy()
predicted_classes = [class_list[p_class]
for p_class in predicted_classes]
assigned_colors = cm.autumn(predicted_entropies)
predicted_scores = predicted_scores.cpu().numpy()
else:
predicted_scores=np.array([])
predicted_classes = np.array([])
assigned_colors = []
gt_cat_idxs = preprocessed_gt_instances['gt_cat_idxs'][image_id].cpu(
).numpy()
thing_dataset_id_to_contiguous_id = meta_catalog.thing_dataset_id_to_contiguous_id
if gt_cat_idxs.shape[0] > 0:
gt_labels = [class_list[thing_dataset_id_to_contiguous_id[gt_class]]
for gt_class in gt_cat_idxs[:, 0]]
else:
gt_labels = []
# noinspection PyTypeChecker
_ = v.overlay_covariance_instances(
boxes=gt_box_means,
assigned_colors=[
'lightgreen' for _ in gt_box_means],
labels=gt_labels,
alpha=1.0)
plotted_detections = v.overlay_covariance_instances(
boxes=predicted_box_means,
covariance_matrices=predicted_box_covariances,
assigned_colors=assigned_colors,
alpha=1.0,
labels=predicted_classes)
cv2.imshow(
'Detected Instances.',
cv2.cvtColor(
plotted_detections.get_image(),
cv2.COLOR_RGB2BGR))
cv2.waitKey()
np.sqrt(np.sum(e_o[i, :, 0]**2 * ndens**2)),
'o-',
markeredgewidth=0,
color=cm.winter((i + 0.5) / 2001))
if e_o[i, 2, 1] > 0:
sign = -1
else:
sign = 1
# ax.plot(zbarr,sign*e_o[ i,:,1]*np.sqrt(ndens)/np.sqrt(np.sum(e_o[ i,:,1]**2*ndens)),'o-',
# markeredgewidth=0,color=cm.autumn((i+0.5)/2001))
ax.plot(zbarr,
sign * e_o[i, :, 1] * ndens /
np.sqrt(np.sum(e_o[i, :, 1]**2 * ndens**2)),
'o-',
markeredgewidth=0,
color=cm.autumn((i + 0.5) / 2001))
ax.plot(zbarr,
f_tm1[0, :, 0] / np.sqrt(np.sum(f_tm1[0, :, 0]**2)),
'ko-',
markeredgewidth=0)
plt.xlabel('$z_\\alpha$', fontsize=18)
plt.ylabel('$\\sqrt{\\bar{n}^\\alpha}\\,({\\sf F}_\\ell)^p_\\alpha$',
fontsize=18)
plt.xlim([0.5, 2.3])
plt.savefig('../Draft/Figs/kl_modes_wl.pdf', bbox_inches='tight')
fisher = (larr + 0.5)[:, None] * (c_p_dfn / c_p_fid)**2
fish_permode = np.sum(fisher, axis=0)
fish_cum = np.cumsum(fish_permode)
if plot_stuff:
# ======
# this figure shows the time levels in each run
figTimes = plt.figure(30, facecolor='w')
axTimes = figTimes.add_subplot(1, 1, 1)
plt.xlabel('Year')
axTimesYlabels = []
# =========
print "Done setting up figure axes."
# =========
# --- Define colors for lines ---
#colors = ['tab:blue', 'tab:orange', 'tab:green', 'tab:red', 'tab:purple', 'tab:brown', 'tab:pink', 'tab:olive', 'tab:cyan']
n150 = sum("amp150" in r for r in runs)
colors = [cm.autumn(x) for x in np.linspace(0.0, 1.0, n150)]
n300 = sum("amp300" in r for r in runs)
colors.extend([cm.winter(x) for x in np.linspace(0.0, 1.0, n300)])
color_index = 0
# ================
# Loop over runs and plot data
# ================
runNumber = 0
for run in runs:
print "Plotting run: " + run
thisRun = runData[run]
# Pull needed data for plotting from this run
# TODO: replace local variables below in plotting commands with reference to object variables
yrs = thisRun.yrs
def Astar(start, goal):
#The A Star function
#Translate to planner coordinates
start = translator(start)
goal = translator(goal)
#Set up the graphs
start_node = LLNode(start)
goal_node = LLNode(goal)
closed_list = []
fringe = [start_node]
#Plot all walls
wall_plotter()
#Plot
plt.scatter(start_node.x, start_node.y, c=cm.autumn(0.0))
found = False #Flag set when search is complete
resign = False #Flag set when nothing left in fringe
while (found != True and resign != True):
cur_node = fringe.pop() #POP the least cost node from fringe
#print('current',cur_node.x, cur_node.y)
#check if we just popped goal if so end now
if at_goal([cur_node.x, cur_node.y], [goal_node.x, goal_node.y]):
print('Found path')
found = True
continue
#Add the popped node into closed_list
closed_list.append(cur_node)
#Check for allowed actions
denied_actions = [temp.obs_actionlist for temp in params.obs_list if [cur_node.x, cur_node.y] == [temp.x, temp.y]]
if len(denied_actions) > 0:
denied_actions = denied_actions[0]
#print(denied_actions, )
for i,action in enumerate(params.delta):
if i in denied_actions: #apply allowed actions on current state
#print('bad action', i)
continue
new = expander([cur_node.x, cur_node.y], action) #get new node
if any(new == [temp.x, temp.y] for temp in closed_list): #check if in closed list
continue
#calculate cost to get to the node
new_orient = np.arctan2((new[1] - cur_node.y),(new[0] - cur_node.x))
if cur_node.parent == None:
old_orient = 0.0
else:
old_orient = np.arctan2((cur_node.y - cur_node.parent.y),(cur_node.x - cur_node.parent.x))
#Penalise turns to avoid zig zag motion
if np.abs(old_orient - new_orient) > 3.2:
cost = 2.58*heuristic([cur_node.x, cur_node.y], new)
else:
cost = (1.0 + np.abs(old_orient - new_orient))*heuristic([cur_node.x, cur_node.y], new)
#Check if we have already visited the node
visited = False
for temp in fringe:
if (new == [temp.x, temp.y]):
visited = True
if (temp.g > cur_node.g + cost):
temp.g = cur_node.g + cost
temp.parent = cur_node
temp.parent_action = i
#If not add to fringe
if visited == False:
new_g = cur_node.g + cost
new_h = heuristic(new, [goal_node.x, goal_node.y])
new_node = LLNode(new, new_g, new_h, cur_node, i)
fringe.append(new_node)
if len(fringe) == 0: #Check if any left to expand in fringe if not end now
print('Failed to find path')
resign = True
else:
fringe = sorted(fringe, key=lambda llnode: llnode.g+(params.heuristic_weight*llnode.h), reverse = True) #sort fringe by g+heusristic
#Debug
for temp in fringe:
#print(temp.x, temp.y, temp.g,temp.h )
plt.scatter(temp.x, temp.y, c = cm.autumn((temp.g)/(temp.g+temp.h)))
#time.sleep(0.1)
#input('h')
if found:
path = []
action_list = []
path.append([cur_node.x, cur_node.y])
action_list.append(cur_node.parent_action)
while cur_node.parent != None:
#print(cur_node.x, cur_node.y)
plt.plot([cur_node.x, cur_node.parent.x], [cur_node.y, cur_node.parent.y], c = [0.0, 0.5, 0.0])
cur_node = cur_node.parent
path.append([cur_node.x, cur_node.y])
action_list.append(cur_node.parent_action)
#print(cur_node.x, cur_node.y)
plt.draw()
plt.pause(0.15)
#return path
return path, action_list
else:
return None, None