def plot_x_query(lwr_theta):
print 'Welcome to the plot demo!\n================================\n Locally weighted regression: lwr_theta for an x_query:\n%s' % lwr_theta
# Read data to make a scatter plot and LR fit
x, y = read_data('/nosql/input/x_data.txt')
x_2dim = x[:, [0, COLUMN_NUMBER]] # because we can't plot more than two dimensions
theta = batch_linear_regression(x_2dim, y)
print 'Simple liner regression produce the following theta for the regression line:\n%s' % theta
# construct a line from intercept and coefficient
x_pred = [min(x[:, 1])[0, 0], max(x[:, 1])[0, 0]]
y_pred = [theta[0, 0] + v * theta[1, 0] for v in x_pred]
plt.figure(figsize=(14,10))
plt.plot(x_2dim[:, 1], y, 'bo', label='Training Set')
plt.plot(x_pred, y_pred, 'r-', label='Linear Regression')
# Predict outcome for x_query
y_query = x_query * lwr_theta
print 'Given the x_query: %s\nLWR predicts target: %s' % (x_query, y_query)
plt.plot(x_query[:, COLUMN_NUMBER], y_query, 'go', markersize=10,
label='Locally Wheighted Linear Regression x_query Prediction')
# Circle the prediction and fine tune the plot
circle = plt.Circle((x_query[:, COLUMN_NUMBER], y_query), 2, color='y', fill=False)
plt.gca().add_artist(circle)
plt.grid()
plt.legend(loc=2)
plt.tight_layout()
plt.show()
def map_along_line(x, y, q, ax=None, cmap=None, norm=None,
time=None, max_step=1., missing=np.nan,
new_timebase=None,
**kwargs):
"""Map some quantity q along x,y as a coloured line.
With time set, perform linear interpolation of x,y,q onto new_timebase
filling with missing, and with max_step."""
if ax is None:
ax = plt.gca()
if x.shape != y.shape:
raise ValueError('Shape mismatch')
if x.shape != q.shape:
raise ValueError('Shape mismatch')
if time is not None:
if new_timebase is None:
new_timebase = np.arange(time[0], time[-1], np.min(np.diff(time)))
# Bit redundant
x = interp_safe(new_timebase, time, x, max_step=max_step, missing=missing)
y = interp_safe(new_timebase, time, y, max_step=max_step, missing=missing)
q = interp_safe(new_timebase, time, q, max_step=max_step, missing=missing)
points = np.array([x, y]).T.reshape(-1, 1, 2)
segments = np.concatenate([points[:-1], points[1:]], axis=1)
lc = LineCollection(segments, cmap=cmap, norm=norm, **kwargs)
lc.set_array(q)
plt.gca().add_collection(lc)
return lc
def plot_svc(X, y, mysvc, bounds=None, grid=50):
if bounds is None:
xmin = np.min(X[:, 0], 0)
xmax = np.max(X[:, 0], 0)
ymin = np.min(X[:, 1], 0)
ymax = np.max(X[:, 1], 0)
else:
xmin, ymin = bounds[0], bounds[0]
xmax, ymax = bounds[1], bounds[1]
aspect_ratio = (xmax - xmin) / (ymax - ymin)
xgrid, ygrid = np.meshgrid(np.linspace(xmin, xmax, grid),
np.linspace(ymin, ymax, grid))
plt.gca(aspect=aspect_ratio)
plt.xlim(xmin, xmax)
plt.ylim(ymin, ymax)
plt.xticks([])
plt.yticks([])
plt.hold(True)
plt.plot(X[y == 1, 0], X[y == 1, 1], 'bo')
plt.plot(X[y == -1, 0], X[y == -1, 1], 'ro')
box_xy = np.append(xgrid.reshape(xgrid.size, 1), ygrid.reshape(ygrid.size, 1), 1)
if mysvc is not None:
scores = mysvc.decision_function(box_xy)
else:
print 'You must have a valid SVC object.'
return None;
CS=plt.contourf(xgrid, ygrid, scores.reshape(xgrid.shape), alpha=0.5, cmap='jet_r')
plt.contour(xgrid, ygrid, scores.reshape(xgrid.shape), levels=[0], colors='k', linestyles='solid', linewidths=1.5)
plt.contour(xgrid, ygrid, scores.reshape(xgrid.shape), levels=[-1,1], colors='k', linestyles='dashed', linewidths=1)
plt.plot(mysvc.support_vectors_[:,0], mysvc.support_vectors_[:,1], 'ko', markerfacecolor='none', markersize=10)
CB = plt.colorbar(CS)
def on_circle_click(self, x, y):
if not self.scale["length_of_scale_in_mu_meter"]:
self.on_scala_click(x, y)
return
plt.scatter(x, y, color="white", s=60, edgecolor="blue", marker="x",
lw=2.5)
if len(self.current_circle_points) == 3:
self.current_circle_points[:] = []
self.current_circle_points.append((x, y))
if len(self.current_circle_points) == 3:
p = self.current_circle_points
point, radius = cirle_from_three_points(p[0][0], p[0][1], p[1][0],
p[1][1], p[2][0], p[2][1])
color = self.colors[self.color_index % len(self.colors)]
self.color_index += 1
circle = matplotlib.patches.Circle(point, radius, lw=4,
facecolor="none", edgecolor=color)
factor = self.scale["length_of_scale_in_mu_meter"] / \
self.scale["length_of_scale_in_px"]
plt.text(point[0], point[1], "Radius: %.4fE-6 m" % (radius *
factor), horizontalalignment='center', color="black",
verticalalignment='center', fontsize=11,
bbox=dict(facecolor=color, alpha=0.85, edgecolor="0.7"))
plt.gca().add_patch(circle)
self.redraw()
def _gaussian_test():
import matplotlib.pyplot as plt
n = 10000
mu_x = 0.0
mu_y = 0.0
#sig_x, sig_y = 1.5, 1.5
tau = 0.0
seeing = 1.5
sigma = seeing / (2. * np.sqrt(2. * np.e))
slit_width = 0.2
slit_height = 10.0
slit_x = np.empty(n, dtype=np.float64)
slit_y = np.empty(n, dtype=np.float64)
slit_x, slit_y = slit_gaussian_psf(n, mu_x, mu_y, sigma, sigma, tau, slit_width, slit_height)
log.info("x range: [%s, %s]", slit_x.min(), slit_x.max())
log.info("y range: [%s, %s]", slit_y.min(), slit_y.max())
plt.scatter(slit_x, slit_y, alpha=0.8)
plt.fill([-slit_width/2, slit_width/2, slit_width/2, -slit_width/2],
[-slit_height/2, -slit_height/2, slit_height/2, slit_height/2],
'r',
alpha=0.10,
edgecolor='k')
plt.gca().set_aspect("equal")
plt.title("Gaussian distribution")
plt.xlim([-slit_height/2., slit_height/2])
plt.show()
def check_models(self):
temp = np.logspace(0, np.log10(600))
num = len(self.available_models())
fig, ax = plt.subplots(1)
self.plotting_colours(num, fig, ax, repeats=2)
for author in self.available_models():
Nc, Nv = self.update(temp=temp, author=author)
# print Nc.shape, Nv.shape, temp.shape
ax.plot(temp, Nc, '--')
ax.plot(temp, Nv, '.', label=author)
ax.loglog()
leg1 = ax.legend(loc=0, title='colour legend')
Nc, = ax.plot(np.inf, np.inf, 'k--', label='Nc')
Nv, = ax.plot(np.inf, np.inf, 'k.', label='Nv')
plt.legend([Nc, Nv], ['Nc', 'Nv'], loc=4, title='Line legend')
plt.gca().add_artist(leg1)
ax.set_xlabel('Temperature (K)')
ax.set_ylabel('Density of states (cm$^{-3}$)')
plt.show()
def plot_std_meshlines(self, step=0.1):
'''
plot mesh circles for stdv
'''
color = self.std_color
nstdmax = self.stdmax
if self.negative:
axmin = -np.pi / 2.
else:
axmin = 0.
th = np.arange(axmin, np.pi / 2, 0.01)
for ra in np.arange(0, nstdmax + 0.1 * step, step):
self.ax.plot(ra * np.sin(th), ra * np.cos(th), ':', color=color)
if self.normalize:
self.ax.set_ylabel('$\sigma / \sigma_{obs}$', color=color)
self.ax.set_xlabel('$\sigma / \sigma_{obs}$', color=color)
else:
self.ax.set_ylabel('Standard Deviation', color=color)
self.ax.set_xlabel('Standard Deviation', color=color)
xticklabels = plt.getp(plt.gca(), 'xticklabels')
plt.setp(xticklabels, color=color)
yticklabels = plt.getp(plt.gca(), 'yticklabels')
plt.setp(yticklabels, color=color)
def __init__(self, history_length=100):
self.history_length = history_length
self.root = Tk.Tk()
self.root.wm_title("GPU MEMORY DISPLAY")
self.root.protocol('WM_DELETE_WINDOW', self.quit_button_cb)
fig = plt.figure(figsize=(8, 3))
self.subplot = plt.subplot(211)
plt.gca().invert_xaxis()
self.canvas = FigureCanvasTkAgg(fig, master=self.root)
self.max_gpu_mem = None
self.current_gpu_mem = None
self.mem_data = [0] * self.history_length
self.mem_range = list(reversed(range(self.history_length)))
self.canvas.get_tk_widget().pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)
self.update()
Tk.mainloop()
def plot_timeseries(self, ax=None, vmin=None, vmax=None,
colorbar=False, label=True):
if vmin is None:
vmin = self.vmin
if vmax is None:
vmax = self.vmax
if ax is None:
ax = plt.gca()
plt.sca(ax)
plt.cla()
plt.imshow(self.tser_arr[::-1,:], vmin=vmin, vmax=vmax,
interpolation='Nearest', extent=self.extent, origin='upper',aspect='auto')
plt.xlim(self.extent[0], self.extent[1])
plt.ylim(self.extent[2], self.extent[3])
# plt.vlines(self.ionogram_list[0].time, self.extent[2], self.extent[3], 'r')
if label:
celsius.ylabel('f / MHz')
if colorbar:
old_ax = plt.gca()
plt.colorbar(
cax = celsius.make_colorbar_cax(), ticks=self.cbar_ticks
).set_label(r"$Log_{10} V^2 m^{-2} Hz^{-1}$")
plt.sca(old_ax)
def AR_predict(d,coeffs, names=None, plot=False): """ Plot the auto-regression predictions and the actual data. """ predictions, ax1, ax2 = [], None, None i = 0 for c in coeffs: p = len(c) y_predict = np.convolve(d,c[::-1],mode='valid') y_predict = y_predict[:-1] ## discard the last value because its outside our domain predictions.append(y_predict) if plot: series_name = names[i] if names!=None else "" y_gt = d[p:] N = len(y_gt) plt.subplot(2,1,1) if ax1== None: ax1 = plt.gca() ax1.plot(np.arange(N), y_gt, label="actual") ax1.plot(np.arange(N), y_predict, label="prediction %s (p=%d)"%(series_name,p)) ax1.legend() plt.subplot(2,1,2) if ax2==None: ax2 = plt.gca() ax2.plot(np.arange(p), c[::-1], label= series_name+' coefficients') ax2.legend() i += 1 if plot: plt.show() return predictions
def showladderpca(self):
import mdp
from matplotlib import pylab as plt
import pprint
import math
cerr('calculating PCA & plotting')
peak_sizes = sorted(list([ x.rtime for x in self.alleles ]))
#peak_sizes = sorted( peak_sizes )[:-5]
#pprint.pprint(peak_sizes)
#comps = algo.simple_pca( peak_sizes )
#algo.plot_pca(comps, peak_sizes)
from fatools.lib import const
std_sizes = const.ladders['LIZ600']['sizes']
x = std_sizes
y = [ x * 0.1 for x in peak_sizes ]
D = np.zeros( (len(y), len(x)) )
for i in range(len(y)):
for j in range(len(x)):
D[i,j] = math.exp( ((x[j] - y[i]) * 0.001) ** 2 )
pprint.pprint(D)
im = plt.imshow(D, interpolation='nearest', cmap='Reds')
plt.gca().invert_yaxis()
plt.xlabel("STD")
plt.ylabel("PEAK")
plt.grid()
plt.colorbar()
plt.show()
def mwpc_last_spills(df, filename, n_spills):
spacing = 6.0 # mm
L = spacing*len(df.T)
y_max = df.max().max()
fig = plt.figure()
ax = fig.add_subplot(111)
x = ((np.array(df.sum().index))*spacing)-(L/2.0)+(spacing/2)
colormap = plt.cm.gist_ncar
plt.gca().set_color_cycle([colormap(i) for i in np.linspace(0, 0.9, n_spills)])
for i, index in enumerate(df.index):
y = df.ix[i]
index = str(index)
stime = index.split(' ')[1].split('.')[0]
sdate = index.split(' ')[0]
ax.plot(x,y, ls='steps-mid', label=stime, linewidth=2)
ax.grid()
ax.set_xlabel('x [mm]')
ax.set_ylabel('y')
ax.set_title('MWPC ({}) {}'.format(sdate, filename))
ax.legend(fancybox=True, framealpha=0.5)
ax.set_ylim(top=y_max*1.2)
figname = filename.split('_')[1].split('.')[0]
filename_figure = '{}{}_1.png'.format(directory, figname)
try:
fig.savefig(filename_figure)
except IOError:
print('{} access denied!'.format(filename_figure))
pass
plt.close(fig)
return df
def graphical_test(satisfactory=0):
from matplotlib import cm, pylab
def cons():
return np.random.random(2)*4-2
def foo(x,y,a,b):
"banana function"
tmp=a-x
tmp*=tmp
out=-x*x
out+=y
out*=out
out*=b
out+=tmp
return out*(abs(np.cos((x-1)**2+(y-1)**2))+10.0/b)
def f(params):
return foo(params[0], params[1],1,100)
optimizer=optimize(f, cons, verbose=False,its=1, hillWalks=0, satisfactory=satisfactory, finalWalk=0)
bgx,bgy=np.mgrid[-2:2:1000j,-2:2:1000j]
bg=foo(bgx,bgy, 1,100)
for i in xrange(20):
pylab.clf()
pylab.imshow(bg, cmap=cm.RdBu,vmax=bg.mean()/10)
for x in optimizer.pool: pylab.plot((x[2]+2)/4*1000,(x[1]+2)/4*1000, ('gx'))
print optimizer.pool[0],optimizer.muterate
pylab.gca().set_xbound(0,1000)
pylab.gca().set_ybound(0,1000)
pylab.draw()
pylab.colorbar()
optimizer.run()
raw_input('enter to advance')
return optimizer
def mass_flux_plot(*args,**kwargs):
fltm = idls.read(args[0])
injm = idls.read(args[1])
f1 = plt.figure()
ax1 = f1.add_subplot(211)
plt.plot(injm.nt_sc,injm.nmf_rscale,'r')
plt.plot(injm.nt_sc,injm.nmf_zscale,'b')
plt.plot(injm.nt_sc,injm.nmf_z0scale,'k')
plt.plot(injm.nt_sc,(injm.nmf_rscale+injm.nmf_zscale),'g')
plt.axis([0.0,160.0,0.0,3.5e-5])
plt.minorticks_on()
locs,labels = plt.yticks()
plt.yticks(locs, map(lambda x: "%.1f" % x, locs*1e5))
plt.text(0.0, 1.03, r'$10^{-5}$', transform = plt.gca().transAxes)
plt.xlabel(r'Time [yr]',labelpad=6)
plt.ylabel(r'$\dot{\rm M}_{\rm out} [ \rm{M}_{\odot} \rm{yr}^{-1} ]$',labelpad=15)
ax2 = f1.add_subplot(212)
plt.plot(fltm.nt_sc,fltm.nmf_rscale,'r')
plt.plot(fltm.nt_sc,fltm.nmf_zscale,'b')
plt.plot(fltm.nt_sc,fltm.nmf_z0scale,'k')
plt.plot(fltm.nt_sc,(fltm.nmf_rscale+fltm.nmf_zscale),'g')
plt.axis([0.0,160.0,0.0,4.0e-5])
plt.minorticks_on()
locs,labels = plt.yticks()
plt.yticks(locs, map(lambda x: "%.1f" % x, locs*1e5))
plt.text(0.0, 1.03, r'$10^{-5}$', transform = plt.gca().transAxes)
plt.xlabel(r'Time [yr]',labelpad=6)
plt.ylabel(r'$\dot{\rm M}_{\rm out} [ \rm {M}_{\odot} \rm{yr}^{-1} ]$',labelpad=15)
def dateticks(fmt='%Y-%m', **kwargs):
'''setup the date ticks'''
dateticker = ticker.FuncFormatter(lambda numdate, _: num2date(numdate).strftime(fmt))
pylab.gca().xaxis.set_major_formatter(dateticker)
# pylab.gcf().autofmt_xdate()
tmp = dict(rotation=30, ha='right')
tmp.update(kwargs)
pylab.setp(pylab.xticks()[1], **tmp)
def hideaxis(pos=None):
# hide x y axis
if pos:
df = pd.DataFrame(pos.values(), columns=['x', 'y'])
plt.xlim([df['x'].min()-5, df['x'].max()+5])
plt.ylim([df['y'].min()-5, df['y'].max()+5])
plt.gca().xaxis.set_major_locator(plt.NullLocator())
plt.gca().yaxis.set_major_locator(plt.NullLocator())
def continuous_calibration():
utc = Calendar()
t_start = utc.time(YMDhms(2011, 9, 1))
t_fc_start = utc.time(YMDhms(2015, 10, 1))
dt = deltahours(1)
n_obs = int(round((t_fc_start - t_start)/dt))
obs_time_axis = Timeaxis(t_start, dt, n_obs + 1)
q_obs_m3s_ts = observed_tistel_discharge(obs_time_axis.total_period())
ptgsk = create_tistel_simulator(PTGSKOptModel, tistel.geo_ts_repository(tistel.grid_spec.epsg()))
initial_state = burn_in_state(ptgsk, t_start, utc.time(YMDhms(2012, 9, 1)), q_obs_m3s_ts)
num_opt_days = 30
# Step forward num_opt_days days and store the state for each day:
recal_start = t_start + deltahours(num_opt_days*24)
t = t_start
state = initial_state
opt_states = {t: state}
while t < recal_start:
ptgsk.run(Timeaxis(t, dt, 24), state)
t += deltahours(24)
state = ptgsk.reg_model_state
opt_states[t] = state
recal_stop = utc.time(YMDhms(2011, 10, 30))
recal_stop = utc.time(YMDhms(2012, 5, 30))
curr_time = recal_start
q_obs_avg = TsTransform().to_average(t_start, dt, n_obs + 1, q_obs_m3s_ts)
target_spec = TargetSpecificationPts(q_obs_avg, IntVector([0]), 1.0, KLING_GUPTA)
target_spec_vec = TargetSpecificationVector([target_spec])
i = 0
times = []
values = []
p, p_min, p_max = construct_calibration_parameters(ptgsk)
while curr_time < recal_stop:
print(i)
i += 1
opt_start = curr_time - deltahours(24*num_opt_days)
opt_state = opt_states.pop(opt_start)
p = ptgsk.region_model.get_region_parameter()
p_opt = ptgsk.optimize(Timeaxis(opt_start, dt, 24*num_opt_days), opt_state, target_spec_vec,
p, p_min, p_max, tr_stop=1.0e-5)
ptgsk.region_model.set_region_parameter(p_opt)
corr_state = adjust_simulator_state(ptgsk, curr_time, q_obs_m3s_ts)
ptgsk.run(Timeaxis(curr_time, dt, 24), corr_state)
curr_time += deltahours(24)
opt_states[curr_time] = ptgsk.reg_model_state
discharge = ptgsk.region_model.statistics.discharge([0])
times.extend(discharge.time(i) for i in range(discharge.size()))
values.extend(list(np.array(discharge.v)))
plt.plot(utc_to_greg(times), values)
plot_results(None, q_obs=observed_tistel_discharge(UtcPeriod(recal_start, recal_stop)))
set_calendar_formatter(Calendar())
#plt.interactive(1)
plt.title("Continuously recalibrated discharge vs observed")
plt.xlabel("Time in UTC")
plt.ylabel(r"Discharge in $\mathbf{m^3s^{-1}}$", verticalalignment="top", rotation="horizontal")
plt.gca().yaxis.set_label_coords(0, 1.1)
def build_normalized_histogram(digit_freq, labels):
fig = plt.figure()
plt.hist(digit_freq.keys(), weights=digit_freq.values())
plt.title('Normalized Histogram')
plt.xlabel('Digit #')
plt.ylabel('Frequency')
plt.gca().set_xlim([0, 9])
fig.savefig('norm_hist.png')
return
def plot_exon(row):
start = int(row["exon_start"])
stop = int(row["exon_stop"])
size = stop - start
#print start, stop
rectangle = plt.Rectangle((start, -20), size, 10, fc='red')
plt.gca().add_patch(rectangle)
def panel(no, xn, yn, bs=True):
plt.subplot(no, aspect="equal")
if bs:
plt.plot(x, y, "k--")
plt.xlabel(xn + r"$ / R_M$")
celsius.ylabel(yn + r"$ / R_M$", offset=off)
plt.xlim(*lims)
plt.ylim(*lims)
plt.gca().add_patch(plt.Circle((0.0, 0.0), 1.0, fill=False))
def plot_column_elem_degree_distribution(self,logscale = True,bins = None):
column_elem_degrees = self.degrees_of_column_elems()
if bins is None:
bins = int(numpy.max(column_elem_degrees))
pylab.hist(column_elem_degrees,bins=bins)
if logscale:
pylab.gca().set_yscale("log")
pylab.gca().set_xscale("log")
pylab.show()
def plot_contours(obj, top_bottom=True):
'''A function that plots the BRF as an azimuthal projection
with contours over the TOC and soil.
Input: rt_layers object, top_bottom - True if only TOC plot, False
if both TOC and soil.
Output: contour plot of brf.
'''
sun = ((np.pi - obj.sun0[0]) * np.cos(obj.sun0[1] + np.pi), \
(np.pi - obj.sun0[0]) * np.sin(obj.sun0[1] + np.pi))
theta = obj.views[:,0]
x = np.cos(obj.views[:,1]) * theta
y = np.sin(obj.views[:,1]) * theta
z = obj.I_top_bottom # * -obj.mu_s
if top_bottom == True:
if np.max > 1.:
maxz = np.max(z)
else:
maxz = 1.
else:
maxz = np.max(z[:obj.n/2])
minz = 0. #np.min(z)
space = np.linspace(minz, maxz, 11)
x = x[:obj.n/2]
y = y[:obj.n/2]
zt = z[:obj.n/2]
zb = z[obj.n/2:]
fig = plt.figure()
if top_bottom == True:
plt.subplot(121)
plt.plot(sun[0], sun[1], 'ro')
triang = tri.Triangulation(x, y)
plt.gca().set_aspect('equal')
plt.tricontourf(triang, zt, space, vmax=maxz, vmin=minz)
plt.title('TOC BRF')
plt.ylabel('Y')
plt.xlabel('X')
if top_bottom == True:
plt.subplot(122)
plt.plot(sun[0], sun[1], 'ro')
plt.gca().set_aspect('equal')
plt.tricontourf(triang, zb, space, vmax=maxz, vmin=minz)
plt.title('Soil Absorption')
plt.ylabel('Y')
plt.xlabel('X')
s = obj.__repr__()
if top_bottom == True:
cbaxes = fig.add_axes([0.11,0.1,0.85,0.05])
plt.suptitle(s,x=0.5,y=0.93)
plt.colorbar(orientation='horizontal', ticks=space,\
cax = cbaxes, format='%.3f')
else:
plt.suptitle(s,x=0.5,y=0.13)
plt.colorbar(orientation='horizontal', ticks=space,\
format='%.3f')
#plt.tight_layout()
plt.show()
def bifurcate(x0, ntransient, nplot, r0, rmax, dr):
global x
count = int((rmax - r0) / dr)
fig = [_Plot(r0 + dr * n, x0, ntransient, nplot) for n in range(count + 1)]
plt.gca().set_xlim(r0, rmax)
plt.gca().set_ylim(np.min(x), np.max(x))
plt.xlabel('r')
plt.ylabel('x')
plt.title('Bifurcation Diagram')
plt.show()
def pltAgainst(a, label, clr):
X = a['X']
low_25 = a['low_25']
high_25 = a['high_25']
medians = a['medians']
plt.plot(X[:,1], low_25, ':', color=clr)
plt.plot(X[:,1], high_25, ':', color=clr)
plt.plot(X[:,1], medians, color=clr, label=label)
plt.gca().yaxis.set_major_formatter(plt.FuncFormatter(lambda x, pos: '%0.2f%%' %(x*100)))
def plotK(JDict, xscale='linear', n='', **kwargs):
paths = JDict.values()
names = JDict.keys()
bnpy.viz.PlotTrace.plotJobs(MakePaths(paths,n), names, MakeStyles(names),
yvar='K', tickfontsize=tickfontsize,
density=1, **kwargs)
set_xscale(xscale)
pylab.ylim(Klims); pylab.yticks(Kticks);
pylab.gca().yaxis.grid() # horizontal lines
pylab.gcf().set_size_inches(W, H);
def plotter_deturing(lam,AB,lam_meas,conv_meas,name):
fig = plt.figure(figsize=(20.0, 10.0))
plt.gca().get_xaxis().get_major_formatter().set_useOffset(False)
plt.plot(lam*1e9,AB)
plt.plot(lam_meas,conv_meas,'o')
plt.xlabel(r'$\left|\lambda_p - \lambda_i \right| (nm)$')
plt.ylabel('Conv_eff (dB)')
plt.savefig(name+'.png',bbox_inches='tight')
plt.close(fig)
return 0
def Plot(func, x0, ntransient, nplot, r0, rmax, dr):
global x
count = int((rmax-r0)/dr)
fig = [_Plot(func, r0+dr*n, x0, ntransient, nplot) for n in range(count+1)]
plt.gca().set_xlim(r0, rmax)
plt.gca().set_ylim(np.min(x), np.max(x))
plt.xlabel(r'$r$', fontsize=16)
plt.ylabel(r'$x$', fontsize=16)
plt.title('Bifurcation Diagram')
plt.show()
def plot_daily_means(time_series):
ext_ts = time_series.copy()
ext_ts['weekday'] = ext_ts.index.weekday
ext_ts.boxplot(column=['Internet traffic data (in GB)'], by='weekday')
plt.title('')
plt.ylabel('data [GB]')
plt.xlabel('time')
plt.gca().xaxis.set_major_formatter(mdates.DateFormatter('%a'))
plt.show()
def plotCorrEnh(self, parsList=None, **plotArgs):
"""
Produces enhanced correlation plots.
Parameters
----------
parsList : list of string, optional
If not given, all available traces are used.
Otherwise a list of at least two parameters
has to be specified.
plotArgs : dict, optional
Keyword arguments handed to plot procedures of
pylab. The following keywords are available:
contour,bins,cmap,origin,interpolation,colors
"""
if not ic.check["matplotlib"]:
PE.warn(PE.PyARequiredImport("To use 'plotCorr', matplotlib has to be installed.", \
solution="Install matplotlib."))
return
tracesDic = {}
if parsList is not None:
for parm in parsList:
self._parmCheck(parm)
tracesDic[parm] = self[parm]
if len(tracesDic) < 2:
raise(PE.PyAValError("For plotting correlations, at least two valid parameters are needed.", \
where="TraceAnalysis::plotCorr"))
else:
# Use all available traces
for parm in self.availableParameters():
tracesDic[parm] = self[parm]
pars = tracesDic.keys()
traces = tracesDic.values()
fontmap = {1:10, 2:9, 3:8, 4:8, 5:8}
if not len(tracesDic)-1 in fontmap:
fontmap[len(tracesDic)-1] = 8
k = 1
for j in range(len(tracesDic)):
for i in range(len(tracesDic)):
if i>j:
plt.subplot(len(tracesDic)-1,len(tracesDic)-1,k)
# plt.title("Pearson's R: %1.5f" % self.pearsonr(pars[j],pars[i])[0], fontsize='x-small')
plt.xlabel(pars[j], fontsize=fontmap[len(tracesDic)-1])
plt.ylabel(pars[i], fontsize=fontmap[len(tracesDic)-1])
tlabels = plt.gca().get_xticklabels()
plt.setp(tlabels, 'fontsize', fontmap[len(tracesDic)-1])
tlabels = plt.gca().get_yticklabels()
plt.setp(tlabels, 'fontsize', fontmap[len(tracesDic)-1])
# plt.plot(traces[j],traces[i],'.',**plotArgs)
self.__hist2d(traces[j],traces[i],**plotArgs)
if i!=j:
k+=1
def plot_S(S, path=None):
from matplotlib import pylab as plt
im1 = plt.imshow(S, interpolation='nearest', cmap='Reds')
#im2 = plt.plot(path[0], path[1], 'k')
plt.gca().invert_yaxis()
plt.xlabel("STD")
plt.ylabel("PEAK")
plt.grid()
plt.colorbar()
plt.show()
def makeplot(tsim, X1, label, pf, *var, **kwargs):
"""
SUMMARY:
It constructs the plot where tsim is on the x-axis,
X1,X2,X3 on the y-axis, and label is the label of the y-axis
SYNTAX:
makeplot(tsim,X1,label,*var):
ARGUMENTS:
+ tsim - x-axis vector (time of the simulation (min))
+ X1,X2,X3 - y-axis vectors.
X1 represent the actual value
X2 the target (eventual)
X3 the setpoint (eventual)
+ label - label for the y-axis
+ pf - path where the plots are saved
+ var - positional variables to include another vector/s X2 and X3 to plot together with X1
+ kwargs - plot options including linestyle and changing the default legend values
"""
linetype = '-' #defaul value for linetype
lableg = 'Target' #defaul value for legend label
for kwkey in kwargs:
if kwkey == 'pltopt':
linetype = kwargs['pltopt']
if kwkey == 'lableg':
lableg = kwargs['lableg']
nt = int(tsim.size)
X1 = np.array(X1)
sz = old_div(X1.size, nt)
Xout1 = np.zeros((nt, sz))
Xout2 = np.zeros((nt, sz))
Xout3 = np.zeros((nt, sz))
for k in range(sz):
x1 = X1[k::sz]
plt.figure()
plt.plot(tsim, x1)
plt.xlabel('Time ')
plt.ylabel(label + str(k + 1))
plt.gca().set_xlim(left=0, right=tsim[-1])
Xout1[:, k] = np.reshape(x1, (nt, ))
for i_var in range(len(var)):
# extract dimension of var
var_i = var[i_var]
Xi = np.array(var_i)
xi = Xi[k::sz]
plt.plot(tsim, xi, ls=linetype)
if i_var == 0:
plt.legend(('Actual', lableg))
plt.gca().set_xlim(left=0, right=tsim[-1])
Xout2[:, k] = np.reshape(xi, (nt, ))
elif i_var == 1:
plt.legend(('Actual', 'Target', 'Set-Point'))
plt.gca().set_xlim(left=0, right=tsim[-1])
Xout3[:, k] = np.reshape(xi, (nt, ))
plt.grid(True)
plt.savefig(pf + label + str(k + 1) + '.pdf',
format='pdf',
transparent=True,
bbox_inches='tight')
return [Xout1, Xout2, Xout3]
pylab.imshow(Z,
interpolation='nearest',
extent=(mx.min(), mx.max(), my.min(), my.max()),
cmap=pylab.cm.Blues,
aspect='auto',
origin='lower')
c2a, c2b, c2c = km.cluster_centers_
pylab.scatter(km.cluster_centers_[:, 0],
km.cluster_centers_[:, 1],
marker='x',
linewidth=2,
s=100,
color='black')
pylab.gca().add_patch(
pylab.Arrow(c1a[0], c1a[1], c2a[0] - c1a[0], c2a[1] - c1a[1], width=0.1))
pylab.gca().add_patch(
pylab.Arrow(c1b[0], c1b[1], c2b[0] - c1b[0], c2b[1] - c1b[1], width=0.1))
pylab.gca().add_patch(
pylab.Arrow(c1c[0], c1c[1], c2c[0] - c1c[0], c2c[1] - c1c[1], width=0.1))
pylab.savefig(os.path.join(CHART_DIR, "1400_03_0%i.png" % i))
pylab.clf()
i += 1
# 3 iterations ####################
km = KMeans(init='random',
n_clusters=num_clusters,
verbose=1,
n_init=1,
# In[110]:
get_ipython().run_line_magic('matplotlib', 'inline')
plt.rcParams['figure.figsize'] = [9.5, 13]
plt.rcParams['figure.subplot.left'] = plt.rcParams['figure.subplot.bottom'] = .1
plt.rcParams['figure.subplot.right'] = plt.rcParams['figure.subplot.top'] = .9
# In[111]:
startTime = time.time()
plt.figure(figsize=(20,4)); ax = plt.gca()
visualize_model(model, ax)
plt.axis('tight'); plt.axis('off'); plt.show()
print('elapsed time: {}'.format(time.time()-startTime))
# # Determine Target Device for Training
# In[112]:
targetDeviceCPU = torch.device('cpu')
targetDeviceGPU = torch.device('cuda:0')
def clean_ticks():
ax = plt.gca()
ax.xaxis.set_ticks_position('bottom')
ax.yaxis.set_ticks_position('left')
def plot_results(ptxsk, q_obs=None):
h_obs = None
n_temp = 1
temp=[]
precip = None
discharge = None
plt.figure() # to start a new plot window
if ptxsk is not None:
plt.subplot(8, 1, 1) # dimension 8 rows of plots
discharge = ptxsk.region_model.statistics.discharge([]) # get the sum discharge all areas
n_temp=ptxsk.region_model.size()
temp = [ptxsk.region_model.statistics.temperature([i]) for i in range(n_temp)]
precip = [ptxsk.region_model.statistics.precipitation([i]) for i in range(n_temp)]
radiation = [ptxsk.region_model.statistics.radiation([i]) for i in range(n_temp)]
rel_hum = [ptxsk.region_model.statistics.rel_hum([i]) for i in range(n_temp)]
wind_speed = [ptxsk.region_model.statistics.wind_speed([i]) for i in range(n_temp)]
snow_sca = [ptxsk.region_model.hbv_snow_state.sca([i]) for i in range(n_temp)]
snow_swe = [ptxsk.region_model.hbv_snow_state.swe([i]) for i in range(n_temp)]
# Results on same time axis, so we only need one
times = utc_to_greg([discharge.time(i) for i in range(discharge.size())])
plt.plot(times, np.array(discharge.v),color='blue')
ax=plt.gca()
ax.set_xlim(times[0], times[-1])
plt.ylabel(r"Discharge in $\mathbf{m^3/s}$")
set_calendar_formatter(Calendar())
if q_obs is not None:
obs_times = utc_to_greg([q_obs.time(i) for i in range(q_obs.size())])
ovs = [q_obs.value(i) for i in range(q_obs.size())]
h_obs, = plt.plot(obs_times, ovs, linewidth=2,color='green')
if ptxsk is not None:
plt.subplot(8, 1, 2,sharex=ax)
for i in range(n_temp):
plt.plot(times, np.array(temp[i].v))
#set_calendar_formatter(Calendar())
plt.ylabel(r"Temperature in C")
plt.subplot(8, 1, 3,sharex=ax)
for i in range(n_temp):
plt.plot(times, np.array(precip[i].v))
plt.ylabel(r"Precipitation in mm")
plt.subplot(8,1,4,sharex=ax)
for i in range(n_temp):
plt.plot(times,np.array(radiation[i].v))
plt.ylabel(r"Radiation w/m2")
plt.subplot(8, 1, 5,sharex=ax)
for i in range(n_temp):
plt.plot(times, np.array(rel_hum[i].v))
plt.ylabel(r"Rel.hum %")
plt.subplot(8, 1, 6,sharex=ax)
for i in range(n_temp):
plt.plot(times, np.array(wind_speed[i].v))
plt.ylabel(r"Wind speed m/s")
plt.subplot(8, 1, 7,sharex=ax)
for i in range(n_temp):
plt.plot(times, np.asarray(snow_swe[i].v)[:-1])
plt.ylabel(r"SWE mm")
plt.subplot(8, 1, 8,sharex=ax)
for i in range(n_temp):
plt.plot(times, np.asarray(snow_sca[i].v)[:-1])
plt.ylabel(r"SCA %")
return h_obs
def add_yaxis(fsp=None, position='right', yscale=None, basey=10, subsy=None):
"""
Adds a second y-axis to a :class:`Subplot`.
This function can also be used as a method.
Parameters
----------
fsp : {None, Subplot}
Subplot to which the secondary y-axis is added.
If None, the current subplot is selected: in that case, it should be
a valid :class:`Subplot`.
When used as a :class:`Subplot` method, this parameter points
automatically to the calling subplot.
position : {string}
Position of the new axis, as either ``'left'`` or ``'right'``.
yscale : {string}
Scale of the new axis, as either ``'log'``, ``'linear'`` or ``None``.
If None, uses the same scale as the first y axis.
basey : {integer}
Base of the logarithm for the new axis (if needed).
subsy : {sequence}
Sequence of the location of the minor ticks;
None defaults to autosubs, which depend on the number of decades in
the plot.
Eg for base 10, ``subsy=(1,2,5)`` will put minor ticks on 1, 2, 5, 11,
12,15, 21, ....
To turn off minor ticking, set ``subsy=[]``.
Raises
------
TypeError
If the selected subplot is not a valid :class:`Subplot` object.
"""
if fsp is None:
fsp = pylab.gca()
fig = fsp.figure
axisini = fsp.axis()
fsp_alt_args = (fsp._rows, fsp._cols, fsp._num + 1)
fsp_alt = fig.add_subplot(frameon=False, position=fsp.get_position(),
sharex=fsp, *fsp_alt_args)
# Set position ....................
if position.lower() == 'right':
(inipos, newpos) = ('left', 'right')
else:
(inipos, newpos) = ('right', 'left')
# Force scales tics to one side ...
fsp.yaxis.set_ticks_position(inipos)
fsp.yaxis.set_label_position(inipos)
# Force 2nd ticks to the other side..
fsp_alt.yaxis.set_ticks_position(newpos)
fsp_alt.yaxis.set_label_position(newpos)
# Force period axis scale..........
if yscale is None:
yscale = fsp.get_yscale()
try:
basey = fsp.yaxis.get_major_locator()._base
except AttributeError:
basey = 10.
fsp_alt.set_yscale(yscale, basey=basey, subsy=subsy)
pylab.draw_if_interactive()
return fsp_alt
def run_test_3(self):
"""
adding an anomaly at a certain point (constant additional demand)
comparing the evolution of clusters (with partial clustering) to the normal evolution of clusters
observing the velocity of change that gives the time until steady state error
"""
day_start_deviation = 10
day_end = 20
deviation = 200
# iterations = day_end - day_start_deviation
iterations = day_end
deviation_element_vect = [None] * iterations
deviation_total_vect = [None] * iterations
res_partial_whole_vect = [None] * iterations
res_partial_whole_anomaly_vect = [None] * iterations
data = copy.deepcopy(self.data[0]["series"])
data_with_constant_anomaly = copy.deepcopy(self.data[0]["series"])
# add constant deviation (anomaly) to the second data set
# starting with day_start_deviation (index for the day of the start of the anomaly)
for i, d in enumerate(range(day_start_deviation, day_end)):
for j in range(len(data_with_constant_anomaly[d])):
data_with_constant_anomaly[d][j] += deviation
centroids_init = dcluster.reinit(data[0:day_start_deviation - 1], 2, 5)
res_partial_whole, a = dcluster.k_means_clust_dynamic()
# for i, d in enumerate(range(day_start_deviation, day_end)):
for i, d in enumerate(range(day_end)):
print(str(i) + "," + str(d))
res_partial_whole_vect[i] = copy.deepcopy(
dcluster.k_means_clust_dynamic_partial_update_whole(data[d]))
# run clustering update for second data set with anomalies
dcluster.reinit(data_with_constant_anomaly[0:day_start_deviation - 1],
2, 5, centroids_init)
res_partial_whole_anomaly, a = dcluster.k_means_clust_dynamic()
# for i,d in enumerate(range(day_start_deviation, day_end)):
for i, d in enumerate(range(day_end)):
print(str(i) + "," + str(d))
res_partial_whole_anomaly_vect[i] = copy.deepcopy(
dcluster.k_means_clust_dynamic_partial_update_whole(
data_with_constant_anomaly[d]))
# plot the results (deviation between the 2 data sets)
for i, d in enumerate(range(day_end)):
total_deviation, average_deviation, deviation = get_comp(
res_partial_whole_anomaly_vect[i], res_partial_whole_vect[i])
print(average_deviation)
deviation_total_vect[i] = average_deviation
deviation_element_vect[i] = deviation
plt.clf()
for ts in res_partial_whole_vect[i]:
plt.plot(ts)
for ts in res_partial_whole_anomaly_vect[i]:
plt.plot(ts)
plt.gca().set_title("deviation " + str(d) +
", with anomaly from day " +
str(day_start_deviation))
plt.pause(0.2)
print(deviation_total_vect)
plt.clf()
# plt.subplot(212)
plt.plot(deviation_total_vect)
# plt.ylim([-100, 100])
plt.gca().set_title("anomaly transient effect on cluster centroids")
plt.xlabel("time (days)")
plt.ylabel("standard deviation")
plt.show()
def percent_ylabel():
plt.gca().set_yticklabels(
['%d%%' % (x * 100) for x in plt.gca().get_yticks()])
print '\ndepletion rates = \n', beta
#for j in range (0,m/2):
# aux = (radius[j])/5
# for i in range(0,360):
# x[i][j] = (radius[j]*np.cos(np.deg2rad(i))) + a[j] #polar coordinates#
# y[i][j] = (radius[j]*np.sin(np.deg2rad(i)) + e[j]) + aux
#plotting#
plt.figure(1)
plt.subplot(2, 1, 1)
plt.xlabel('Semimajor Axis (AU)')
plt.ylabel('Eccentricity')
plt.title('Initial Disk')
axes = plt.gca()
axes.set_xlim([0, 5])
axes.set_ylim([0, 0.2])
plt.scatter(a, b, s=400, edgecolors='none', color=(0.2, 0.1, 0.8)) #embryos#
plt.scatter(plan, c, s=30, color=(0.6, 0.9, 0), edgecolors='none',
alpha=0.8) #planetesimals#
plt.subplot(2, 2, 3)
plt.xlabel('Semimajor Axis (AU)')
plt.ylabel('Earth Mass')
plt.title('Initial Disk')
axes = plt.gca()
axes.set_xlim([0, 5])
axes.set_yscale('log')
axes.set_ylim([0.001, 1])
plt.scatter(a, mass_old, s=50, edgecolors='none',
# split data into train and test sets
pred_train, pred_test, tar_train, tar_test = train_test_split(predictors,
target,
test_size=.3,
random_state=123)
# specify the lasso regression model
model = LassoLarsCV(cv=10, precompute=False).fit(pred_train, tar_train)
# print variable names and regression coefficients
dict(zip(predictors.columns, model.coef_))
# plot coefficient progression
m_log_alphas = -np.log10(model.alphas_)
ax = plt.gca()
plt.plot(m_log_alphas, model.coef_path_.T, label=model.coef_path_.T)
plt.axvline(-np.log10(model.alpha_),
linestyle='--',
color='k',
label='alpha CV')
plt.ylabel('Regression Coefficients')
plt.xlabel('-log(alpha)')
plt.title('Regression Coefficients Progression for Lasso Paths')
# plot mean square error for each fold
m_log_alphascv = -np.log10(model.cv_alphas_)
plt.figure()
plt.plot(m_log_alphascv, model.cv_mse_path_, ':')
plt.plot(m_log_alphascv,
model.cv_mse_path_.mean(axis=-1),
def plot_options_greedy(self, sess, coord, saver):
eigenvectors_path = os.path.join(
os.path.join(self.config.stage_logdir, "models"),
"eigenvectors.npy")
eigenvalues_path = os.path.join(
os.path.join(self.config.stage_logdir, "models"),
"eigenvalues.npy")
eigenvectors = np.load(eigenvectors_path)
eigenvalues = np.load(eigenvalues_path)
for k in ["poz", "neg"]:
for option in range(len(eigenvalues)):
# eigenvalue = eigenvalues[option]
eigenvector = eigenvectors[
option] if k == "poz" else -eigenvectors[option]
prefix = str(option) + '_' + k + "_"
plt.clf()
with sess.as_default(), sess.graph.as_default():
for idx in range(self.nb_states):
dx = 0
dy = 0
d = False
s, i, j = self.env.get_state(idx)
if not self.env.not_wall(i, j):
plt.gca().add_patch(
patches.Rectangle(
(j, self.config.input_size[0] - i -
1), # (x,y)
1.0, # width
1.0, # height
facecolor="gray"))
continue
# Image.fromarray(np.asarray(scipy.misc.imresize(s, [512, 512], interp='nearest'), np.uint8)).show()
feed_dict = {self.orig_net.observation: np.stack([s])}
fi = sess.run(self.orig_net.fi, feed_dict=feed_dict)[0]
transitions = []
terminations = []
for a in range(self.action_size):
s1, r, d, _ = self.env.fake_step(a)
feed_dict = {
self.orig_net.observation: np.stack([s1])
}
fi1 = sess.run(self.orig_net.fi,
feed_dict=feed_dict)[0]
transitions.append(
self.cosine_similarity((fi1 - fi),
eigenvector))
terminations.append(d)
transitions.append(
self.cosine_similarity(np.zeros_like(fi),
eigenvector))
terminations.append(True)
a = np.argmax(transitions)
# if a == 4:
# d = True
if a == 0: # up
dy = 0.35
elif a == 1: # right
dx = 0.35
elif a == 2: # down
dy = -0.35
elif a == 3: # left
dx = -0.35
if terminations[a] or np.all(
transitions[a] == np.zeros_like(
fi)): # termination
circle = plt.Circle(
(j + 0.5,
self.config.input_size[0] - i + 0.5 - 1),
0.025,
color='k')
plt.gca().add_artist(circle)
continue
plt.arrow(j + 0.5,
self.config.input_size[0] - i + 0.5 - 1,
dx,
dy,
head_width=0.05,
head_length=0.05,
fc='k',
ec='k')
plt.xlim([0, self.config.input_size[1]])
plt.ylim([0, self.config.input_size[0]])
for i in range(self.config.input_size[1]):
plt.axvline(i, color='k', linestyle=':')
plt.axvline(self.config.input_size[1],
color='k',
linestyle=':')
for j in range(self.config.input_size[0]):
plt.axhline(j, color='k', linestyle=':')
plt.axhline(self.config.input_size[0],
color='k',
linestyle=':')
plt.savefig(
os.path.join(
self.summary_path,
"SuccessorFeatures_" + prefix + 'policy.png'))
plt.close()
plt.savefig(path, format="png", dpi=300) # Declare datasets X1D = np.linspace(-4, 4, 9).reshape(-1, 1) X2D = np.c_[X1D, X1D**2] y = np.array([0, 0, 1, 1, 1, 1, 1, 0, 0]) plt.figure(figsize=(11, 4)) plt.subplot(121) plt.grid(True, which='both') plt.axhline(y=0, color='k') plt.plot(X1D[:, 0][y == 0], np.zeros(4), "bs") plt.plot(X1D[:, 0][y == 1], np.zeros(5), "g^") plt.gca().get_yaxis().set_ticks([]) plt.xlabel(r"$x_1$", fontsize=20) plt.axis([-4.5, 4.5, -0.2, 0.2]) plt.subplot(122) plt.grid(True, which='both') plt.axhline(y=0, color='k') plt.axvline(x=0, color='k') plt.plot(X2D[:, 0][y == 0], X2D[:, 1][y == 0], "bs") plt.plot(X2D[:, 0][y == 1], X2D[:, 1][y == 1], "g^") plt.xlabel(r"$x_1$", fontsize=20) plt.ylabel(r"$x_2$", fontsize=20, rotation=0) plt.gca().get_yaxis().set_ticks([0, 4, 8, 12, 16]) plt.plot([-4.5, 4.5], [6.5, 6.5], "r--", linewidth=3) plt.axis([-4.5, 4.5, -1, 17])
dataframe = read_csv('../file/final_data/lstm/purchase/all_purchase230.csv',
index_col='report_date',
parse_dates=[0])
# dataset = dataframe.values
sub_total_purchase_residual_tmp = dataframe['residual2']
sub_total_purchase_residual = sub_total_purchase_residual_tmp[:]
print(sub_total_purchase_residual.describe())
stationarity_test(sub_total_purchase_residual)
sub_total_purchase_residual.describe()
sub_total_purchase_residual.plot()
plt.title('Purchase Residual By Lstm')
plt.show()
# 直方图 是否正态分布
sub_total_purchase_residual.hist()
plt.title('Purchase Residual Histogram By Lstm')
plt.show()
# autocorrelation
plot_acf(sub_total_purchase_residual, ax=plt.gca(), lags=60)
plt.title('Purchase Residual ACF By Lstm')
plt.show()
# LBQ 检验
from statsmodels.stats import diagnostic
print(
diagnostic.acorr_ljungbox(sub_total_purchase_residual,
lags=None,
boxpierce=True))
ax1.grid(axis="x", linestyle="--", color='black', linewidth=0.25, alpha=0.5)
ax1.grid(axis="y", linestyle="--", color='black', linewidth=0.25, alpha=0.5)
# Show the minor grid lines with very faint and almost transparent grey lines
plt.minorticks_on()
plt.grid(b=True, which='minor', color='#999999', linestyle='-', alpha=0.2)
#xposition = [1970, 2000, 2010, 2040, 2070]
#for xc in xposition:
# plt.axvline(x=xc, color='k', linestyle='--')
#for label in ax1.get_yticklabels():
# label.set_fontsize(20)
#for tick in ax1.xaxis.get_major_ticks():
# tick.label.set_fontsize(20)
plt.setp(plt.gca().get_xticklabels(), rotation=0, ha="right")
plt.xlabel('Année', fontsize=20, color='black', weight='semibold')
plt.ylabel('', fontsize=20, color='black', weight='semibold')
plt.title(
'Variabilités interannuelles des anomalies standardisées de l\'épaisseur de neige au 31 mars par rapport à la normale 1981-2010 \n',
fontsize=20,
color='black',
weight='semibold')
for label in ax1.get_yticklabels():
label.set_fontsize(20)
plt.savefig(
'K:/PROJETS/PROJET_CLIMHUNOR/Atlas/Atlas_figures//VI_Ano_Stand_SD_31mars_vs_1981-2010.png',
bbox_inches='tight',
format='png',
plt.ylabel('Distancia (km)')
plt.title('Distancia al centro de la tierra del satélite vs. el tiempo')
plt.show()
plt.figure(dpi=300)
num_segmentos = 1000
rad = r+80000
cx = 0
cy = 0
angulo = np.linspace(0, 2*np.pi, num_segmentos+1)
x1 = rad * np.cos(angulo) + cx
y1 = rad * np.sin(angulo) + cy
plt.plot(x1, y1, color='g', label= 'Superficie terrestre')
plt.xlabel('X (km)')
plt.ylabel('Y (km)')
plt.yticks((-r-80000,r+80000),('7 km','7 km'))
plt.xticks((-r-80000,r+80000),('7 km','7 km'))
plt.ylim(top=8000000, bottom = -8000000)
plt.gca().set_aspect('equal')
plt.grid()
plt.plot(x,y, c='lightblue', label='Trayectoria satelite',alpha=0.7)
plt.legend()
plt.tight_layout()
plt.show()
V = zeros((L, L), float)#make initial grid+BC
rho = zeros(V.shape,float)#Make rho
rho[25:76,37] = linear(rho[25:76,37], A)#set plate1
rho[25:76,63] = -rho[25:76,37]#set plate2
poisson(V,rho)
plot(V,"""Parallel Plate Capacitor w/
Triangular Charge Density""", "cs4_6", elev=el,azim=ang)
#Make plot of charge densities
cs = ones(51,float)*100#Normal parallel plate
fg = gaussian(zeros(51,float), A, 0, s)#fitted gaus
lg = gaussian(zeros(51,float), A, 0, 2*s)#large gaus
pb = parab(zeros(51,float), A)#parab
tg = linear(zeros(51,float), A)#triang
fig = p.figure()#make fig
ax = p.gca()
p.plot(arange(25,76),cs,arange(25,76),fg,arange(25,76),\
lg,arange(25,76),pb,arange(25,76),tg)
ax.set_xlabel('X'); ax.set_ylabel('Y')
ax.set_xlim(25,75); ax.set_ylim(0,100.5)
p.xticks(arange(25, 76, 5)); p.yticks(arange(0, 101, 10))
ax.set_title("Charge Distributions")
p.grid(True)
p.legend(["Constant","Fitted Gaussian", "Large Gaussian", \
"Parabolic", "Triangular"])
p.savefig('figures/cs4_cd.png', bbox_inches='tight')
#Tried sep of variables just to see soln
L = 40
sC = pi/L
C1 = 100/(1-exp(-2*sC*L))
def plot_SeaIce_argo_QC_temp_sal(plevIntp, temp2d, psal2d, platformDf_plev,
tdelta, seaice_val, platform_number):
temp2d_new = np.empty((len(plevIntp), 1))
psal2d_new = np.empty((len(plevIntp), 1))
temp2d_new[:, 0] = temp2d[:, 0]
psal2d_new[:, 0] = psal2d[:, 0]
time_new = pd.Series(pd.to_datetime(platformDf_plev['date'][0]))
for i in np.arange(1, len(platformDf_plev['date'])):
nnans_days = (pd.to_datetime(platformDf_plev['date'][i]) -
pd.to_datetime(platformDf_plev['date'][i - 1])).days
# in the following we add 1 profile of nan in between profiles that are too far apart in time
# (if more profiles of nan are needed, then you need to edit np.empty and the number of iterations in j-loop)
if nnans_days > tdelta:
temp2d_new = np.append(temp2d_new,
np.nan * np.empty((len(plevIntp), 1)),
axis=1)
psal2d_new = np.append(psal2d_new,
np.nan * np.empty((len(plevIntp), 1)),
axis=1)
bfr_date = pd.to_datetime(platformDf_plev['date'][i - 1])
# this loop is now only looping once as we only need to add one profile of nans
for j in np.arange(1, 2, 1):
bfr_date += timedelta(days=1)
time_new = time_new.append(pd.Series(bfr_date))
temp2d_new = np.append(temp2d_new, temp2d[:, np.newaxis, i], axis=1)
psal2d_new = np.append(psal2d_new, psal2d[:, np.newaxis, i], axis=1)
time_new = time_new.append(
pd.Series(pd.to_datetime(platformDf_plev['date'][i])))
plt.figure(figsize=(30, 20))
for i in [1, 2, 3]:
if i == 1:
ax = plt.subplot(311)
color = 'tab:red'
plt.plot(pd.to_datetime(platformDf_plev['date']),
platformDf_plev['position_qc'],
'sr',
markersize=14)
ax.set_ylabel('Position QC flag', size=24, labelpad=0)
ax.yaxis.label.set_color('red')
ax.tick_params(axis='y', colors='red')
ax2 = ax.twinx()
color = 'tab:blue'
ax2.plot(pd.to_datetime(platformDf_plev['date']),
np.stack(seaice_val),
'ob',
markersize=12)
ax2.set_ylabel('SOSE sea-ice fraction', size=24, labelpad=0)
ax2.yaxis.label.set_color('blue')
ax2.tick_params(axis='y', colors='blue')
plt.title('Float #' + platform_number, size=24)
plt.grid()
if i == 2:
ax = plt.subplot(312)
mt = np.ma.masked_where(np.isnan(temp2d_new), temp2d_new)
plt.pcolor(pd.to_datetime(time_new), plevIntp, mt, cmap='plasma')
#plt.pcolormesh(pd.to_datetime(platformDf_plev['date']),plevIntp,temp2d) # pd.Series(np.arange(0,79,1))
#plt.title('Temperature, degC',size=24)
cbar = plt.colorbar(orientation="horizontal", pad=0.2)
cbar.ax.tick_params(labelsize=24)
cbar.set_label('Temperature, degC', fontsize=24)
if i == 3:
ax = plt.subplot(313)
ms = np.ma.masked_where(np.isnan(psal2d_new), psal2d_new)
plt.pcolor(pd.to_datetime(time_new),
plevIntp,
ms,
cmap='viridis_r')
# plt.pcolor(pd.to_datetime(platformDf_plev['date']),plevIntp,psal2d) # pd.Series(np.arange(0,79,1))
#plt.title('Salinity, psu',size=24)
cbar = plt.colorbar(orientation="horizontal", pad=0.2)
cbar.ax.tick_params(labelsize=24)
cbar.set_label('Salinity, PSU', fontsize=24)
if i == 2 or i == 3:
plt.gca().invert_yaxis()
ax.set_ylabel('Pressure, dbar', size=24, labelpad=0)
# change font
for tick in ax.xaxis.get_majorticklabels(): # example for xaxis
tick.set_fontsize(24)
for tick in ax.yaxis.get_majorticklabels(): # example for xaxis
tick.set_fontsize(24)
for tick in ax2.yaxis.get_majorticklabels(): # example for xaxis
tick.set_fontsize(24)
plt.xlim([
min(pd.to_datetime(platformDf_plev['date'])),
max(pd.to_datetime(platformDf_plev['date']))
])
plt.subplots_adjust(left=0.1,
bottom=0.1,
right=0.9,
top=0.9,
wspace=0.4,
hspace=0.4)
plt.show()
def shap_deep_explainer(self,
model_no,
num_reference,
img_input,
norm_reverse=True,
blend_original_image=False,
gif_fps=1,
ranked_outputs=1,
base_dir_save='/tmp/DeepExplain'):
# region mini-batch because of GPU memory limitation
list_shap_values = []
batch_size = self.dicts_models[model_no]['batch_size']
split_times = math.ceil(num_reference / batch_size)
for i in range(split_times):
#shap 0.26
#shap 0.4, check_additivity=False
# shap_values_tmp1 = self.list_e[model_no][i].shap_values(img_input, ranked_outputs=ranked_outputs,
# check_additivity=check_additivity)
shap_values_tmp1 = self.list_e[model_no][i].shap_values(
img_input,
ranked_outputs=ranked_outputs,
)
# shap_values ranked_outputs
# [0] [0] (1,299,299,3)
# [1] predict_class array
shap_values_copy = copy.deepcopy(shap_values_tmp1)
list_shap_values.append(shap_values_copy)
for i in range(ranked_outputs):
for j in range(len(list_shap_values)):
if j == 0:
shap_values_tmp2 = list_shap_values[0][0][i]
else:
shap_values_tmp2 += list_shap_values[j][0][i]
shap_values_results = copy.deepcopy(list_shap_values[0])
shap_values_results[0][i] = shap_values_tmp2 / split_times
# endregion
# region save files
str_uuid = str(uuid.uuid1())
list_classes = []
list_images = []
for i in range(ranked_outputs):
predict_class = int(
shap_values_results[1][0][i]) # numpy int 64 - int
list_classes.append(predict_class)
save_filename = os.path.join(
base_dir_save, str_uuid,
'Shap_Deep_Explainer{}.jpg'.format(predict_class))
os.makedirs(os.path.dirname(save_filename), exist_ok=True)
list_images.append(save_filename)
pred_class_num = len(shap_values_results[0])
if blend_original_image:
if norm_reverse:
img_original = np.uint8(input_norm_reverse(img_input[0]))
else:
img_original = np.uint8(img_input[0])
img_original_file = os.path.join(os.path.dirname(list_images[0]),
'deepshap_original.jpg')
cv2.imwrite(img_original_file, img_original)
for i in range(pred_class_num):
# predict_max_class = attributions[1][0][i]
attribution1 = shap_values_results[0][i]
# attributions.shape: (1, 299, 299, 3)
data = attribution1[0]
data = np.mean(data, -1)
abs_max = np.percentile(np.abs(data), 100)
abs_min = abs_max
# dx, dy = 0.05, 0.05
# xx = np.arange(0.0, data1.shape[1], dx)
# yy = np.arange(0.0, data1.shape[0], dy)
# xmin, xmax, ymin, ymax = np.amin(xx), np.amax(xx), np.amin(yy), np.amax(yy)
# extent = xmin, xmax, ymin, ymax
# cmap = 'RdBu_r'
# cmap = 'gray'
cmap = 'seismic'
plt.axis('off')
# plt.imshow(data1, extent=extent, interpolation='none', cmap=cmap, vmin=-abs_min, vmax=abs_max)
# plt.imshow(data, interpolation='none', cmap=cmap, vmin=-abs_min, vmax=abs_max)
# fig = plt.gcf()
# fig.set_size_inches(2.99 / 3, 2.99 / 3) # dpi = 300, output = 700*700 pixels
plt.gca().xaxis.set_major_locator(plt.NullLocator())
plt.gca().yaxis.set_major_locator(plt.NullLocator())
plt.subplots_adjust(top=1,
bottom=0,
right=1,
left=0,
hspace=0,
wspace=0)
plt.margins(0, 0)
if blend_original_image:
plt.imshow(data,
interpolation='none',
cmap=cmap,
vmin=-abs_min,
vmax=abs_max)
save_filename1 = list_images[i]
plt.savefig(save_filename1, bbox_inches='tight', pad_inches=0)
plt.close()
img_heatmap = cv2.imread(list_images[i])
(tmp_height, tmp_width) = img_original.shape[:-1]
img_heatmap = cv2.resize(img_heatmap, (tmp_width, tmp_height))
img_heatmap_file = os.path.join(
os.path.dirname(list_images[i]),
'deepshap_{0}.jpg'.format(i))
cv2.imwrite(img_heatmap_file, img_heatmap)
dst = cv2.addWeighted(img_original, 0.65, img_heatmap, 0.35, 0)
img_blend_file = os.path.join(
os.path.dirname(list_images[i]),
'deepshap_blend_{0}.jpg'.format(i))
cv2.imwrite(img_blend_file, dst)
# region create gif
import imageio
mg_paths = [
img_original_file, img_heatmap_file, img_blend_file
]
gif_images = []
for path in mg_paths:
gif_images.append(imageio.imread(path))
img_file_gif = os.path.join(os.path.dirname(list_images[i]),
'deepshap_{0}.gif'.format(i))
imageio.mimsave(img_file_gif, gif_images, fps=gif_fps)
list_images[i] = img_file_gif
# endregion
else:
plt.imshow(data,
interpolation='none',
cmap=cmap,
vmin=-abs_min,
vmax=abs_max)
save_filename1 = list_images[i]
plt.savefig(save_filename1, bbox_inches='tight', pad_inches=0)
plt.close()
# endregion
return list_classes, list_images
def plot_many(bins,
linestyles,
bin_labels,
age_then=None,
z_func=None,
colors=None,
**kwargs):
import numpy as np
import matplotlib.pylab as plt
from seren3.array import SimArray
from seren3.utils.plot_utils import ncols
ax1 = plt.gca()
legendArtists = []
for i in range(len(bins)):
ls = linestyles[i]
plot_mean_integrated_fesc(bins[i], linestyle=ls, colors=colors,\
label=False, legend=False, **kwargs)
legendArtists.append(
plt.Line2D((0, 1), (0, 0), color='k', linestyle=ls))
handles, labels = ax1.get_legend_handles_labels()
display = tuple(range(len(bins[0])))
ax1.set_xlim(0.2, 0.7)
# Shrink current axis's height by 10% on the bottom
box = ax1.get_position()
ax1.set_position(
[box.x0, box.y0 + box.height * 0.1, box.width, box.height * 0.9])
# Put a legend below current axis
legend_kwargs = {"title" : r"log$_{10}$(Mvir)", "loc" : "upper center", \
"bbox_to_anchor" : (0.5, -0.1), "fancybox" : True, "shadow" : True, "ncol" : 4}
ax1.legend([handle for i,handle in enumerate(handles) if i in display]+legendArtists,\
[label for i,label in enumerate(labels) if i in display]+bin_labels, **legend_kwargs)
# Redshift axis
if (z_func is not None) and (age_then is not None):
ax2 = ax1.twiny()
ax2.set_position(
[box.x0, box.y0 + box.height * 0.1, box.width, box.height * 0.9])
def tick_function(X):
# return ["%1.1f" % i for i in z_func( age_now - X )]
# We've scaled the x-axis by the age of halos, so need
# to account for zero-point in age
return ["%1.1f" % i for i in z_func(SimArray(X + age_then, "Gyr"))]
xtickpos = ax1.get_xticks()
print xtickpos
new_tick_locations = np.linspace(0.215, 0.685, 5)
print new_tick_locations
# new_tick_locations = np.array(ax1.get_xticks())
ax2.set_xlim(ax1.get_xlim())
ax2.set_xticks(new_tick_locations)
ax2.set_xticklabels(tick_function(new_tick_locations))
ax2.set_xlabel(r"Redshift")
ax1.set_xlabel(r"t$_{\mathrm{H}}$ [Gyr]")
ax1.set_ylabel(
r"$\langle \mathrm{f}_{\mathrm{esc}} \rangle$ (<t$_{\mathrm{H}}$ [%])")
def save_plot(title, out_dir=os.path.join('/', 'tmp')):
plt.title(title, fontsize=11)
min_y = plt.gca().get_ylim()[0]
if min_y > 0:
plt.gca().set_ylim([0, plt.gca().get_ylim()[1]])
plt.gca().spines['left'].set_visible(False)
plt.gca().spines['right'].set_visible(False)
plt.gca().spines['top'].set_visible(False)
plt.gca().spines['bottom'].set_visible(False)
plt.gca().tick_params('x', length=0)
plt.gca().tick_params('y', length=0)
plt.tight_layout()
plt.savefig(os.path.join(out_dir, title.replace(' ', '_') + '.png'))
plt.cla()
def plot_waveforms(waveforms,
sensors,
evt,
*,
wf_type="PMT",
range=(None, ),
overlay=False,
sum=False,
zoomx=False,
zoomy=False,
dual=False):
range = slice(*range)
wfsize = waveforms.shape[2]
time = np.arange(wfsize).astype(float)
if wf_type == "PMT": time /= 40
elif wf_type == "BLR": time /= 40
elif wf_type == "SiPM": pass
else:
raise ValueError("Unrecognized wf type {}. ".format(wf_type) +
"Valid options: are 'PMT', 'BLR' and 'SiPM'")
if sum: wf_type += " SUM"
gmin, gmax = float("inf"), -float("inf")
plt.ion()
ax1 = plt.gca()
if sum:
sum_wf = np.zeros(waveforms.shape[2])
if dual:
for wf, wf_dual, ID, color in zip(waveforms[0][range],
waveforms[1][range], sensors[range],
_colors):
ymin, ymax = min(wf_dual), max(wf_dual)
if ymin < gmin: gmin = ymin
if ymax > gmax: gmax = ymax
plt.plot(wf, drawstyle="steps", label=str(ID[0]), c=color)
plt.plot(wf_dual,
drawstyle="steps",
label=str(ID[0]),
c=next(_colors))
ylim = (0.99 * ymin, 1.01 * ymax)
customize_plot(zoomx, zoomy if zoomy else ylim, wf_type, evt,
ID[0])
show_and_wait()
else:
for wf, ID, color in zip(waveforms[0][range], sensors[range], _colors):
ymin, ymax = min(wf), max(wf)
if ymin < gmin: gmin = ymin
if ymax > gmax: gmax = ymax
if sum:
bls_wf = wf - mode(wf)
sum_wf = sum_wf + bls_wf * (1 if "SiPM" in wf_type else -1)
else:
plt.plot(wf, drawstyle="steps", label=str(ID[0]), c=color)
if not overlay and not sum:
ylim = (0.99 * ymin, 1.01 * ymax)
customize_plot(zoomx, zoomy if zoomy else ylim, wf_type, evt,
ID[0])
show_and_wait()
if overlay:
ylim = 0.99 * gmin, 1.01 * gmax
customize_plot(zoomx, zoomy if zoomy else ylim, wf_type, evt)
show_and_wait()
if sum:
ylim = np.min(sum_wf) - 50, np.max(sum_wf) + 50
plt.plot(sum_wf, drawstyle="steps", c="k")
customize_plot(zoomx, zoomy if zoomy else ylim, wf_type, evt)
show_and_wait()
def plot_results(detector,
img,
x,
small,
mixcomp=None,
bounding_boxes=[],
img_resized=None):
# Get max peak
#print ix, iy
#print '---'
#print x.shape
#print small.shape
plt.clf()
if small is None and x is None:
plt.subplot(111)
else:
plt.subplot(121)
plt.title('Input image')
plt.imshow(img, cmap=plt.cm.gray)
for dbb in bounding_boxes[::-1]:
bb = dbb.box
color = 'cyan' if dbb.correct else 'red'
plt.gca().add_patch(
plt.Rectangle((bb[1], bb[0]),
bb[3] - bb[1],
bb[2] - bb[0],
facecolor='none',
edgecolor=color,
linewidth=2.0))
#plt.text(bb[1], bb[0], "{0:.2f}".format(dbb.confidence), color='white', backgroundcolor=color, size=8, ha='left', va='bottom')
plt.text(bb[1],
bb[0],
"{0:.2f}".format(dbb.confidence),
color='yellow',
size=6,
ha='left',
va='bottom')
if x is not None:
plt.subplot(122)
#plt.title('Response map ({:.2f}, {:.2f})'.format(float(x.min()), float(x.max())))
plt.title('Response map')
plt.imshow(x, interpolation='nearest') #, vmin=-40000, vmax=-36000)
#plt.colorbar()
if 0:
if small is not None:
plt.subplot(223)
plt.title('Feature activity')
plt.imshow(small.sum(axis=-1), interpolation='nearest')
plt.colorbar()
if img_resized is not None:
plt.subplot(224)
plt.title('Resized image')
plt.imshow(img_resized, interpolation='nearest', cmap=plt.cm.gray)
if 0:
pass
plt.title('Normalized stuff')
plt.imshow(x / np.clip(small.sum(axis=-1), 5, np.inf),
interpolation='nearest')
plt.colorbar()
else:
#if mixcomp is not None:
#plt.title('Kernel Bernoulli probability averages')
#plt.imshow(detector.kernels[mixcomp].mean(axis=-1), interpolation='nearest', cmap=plt.cm.RdBu, vmin=0, vmax=1)
#plt.colorbar()
pass
import sys
from os import listdir
from os.path import isfile, join
import numpy as np
import matplotlib.pylab as plt
if __name__ == "__main__":
data_path = sys.argv[1]
data_files = []
for f in listdir(data_path):
if isfile(join(data_path, f)):
data_files.append(join(data_path, f))
fig, ax = plt.subplots()
for data_file in data_files:
data = np.loadtxt(open(data_file, "r"), delimiter=",")
track = np.array([data[:, 0], data[:, 1]]).T
plt.plot(track[:, 0], track[:, 1])
# ax.suptitle("Feature Track")
ax.set_xlim([-1.0, 1.0])
ax.set_ylim([-1.0, 1.0])
ax.xaxis.tick_top()
plt.gca().invert_yaxis()
plt.show()
def wiggle(Data,
SH={},
maxval=-1,
skipt=1,
lwidth=.5,
x=[],
t=[],
gain=1,
type='VA',
color='black',
ntmax=1e+9):
"""
wiggle(Data,SH)
"""
import matplotlib.pylab as plt
import numpy as np
import copy
yl = 'Sample number'
ns = Data.shape[0]
ntraces = Data.shape[1]
if ntmax < ntraces:
skipt = int(np.floor(ntraces / ntmax))
if skipt < 1:
skipt = 1
if len(x) == 0:
x = range(0, ntraces)
if len(t) == 0:
t = range(0, ns)
else:
yl = 'Time [s]'
# overrule time form SegyHeader
if 'time' in SH:
t = SH['time']
yl = 'Time [s]'
dx = x[1] - x[0]
if (maxval <= 0):
Dmax = np.nanmax(Data)
maxval = -1 * maxval * Dmax
print('segypy.wiggle: maxval = %g' % maxval)
#fig, (ax1) = plt.subplots(1, 1)'
fig = plt.gcf()
ax1 = plt.gca()
for i in range(0, ntraces, skipt):
# use copy to avoid truncating the data
trace = copy.copy(Data[:, i])
trace[0] = 0
trace[-1] = 0
ax1.plot(x[i] + gain * skipt * dx * trace / maxval,
t,
color=color,
linewidth=lwidth)
if type == 'VA':
for a in range(len(trace)):
if (trace[a] < 0):
trace[a] = 0
# pylab.fill(i+Data[:,i]/maxval,t,color='k',facecolor='g')
#ax1.fill(x[i] + dx * Data[:, i] / maxval, t, 'k', linewidth=0, color=color)
ax1.fill(x[i] + gain * skipt * dx * trace / (maxval),
t,
'k',
linewidth=0,
color=color)
ax1.grid(True)
ax1.invert_yaxis()
plt.ylim([np.max(t), np.min(t)])
plt.xlabel('Trace number')
plt.ylabel(yl)
if 'filename' in SH:
plt.title(SH['filename'])
def streamplot(
UV,
ax=None,
map=None,
geodata=None,
drawlonlatlines=False,
basemap_resolution="l",
cartopy_scale="50m",
lw=0.5,
cartopy_subplot=(1, 1, 1),
axis="on",
**kwargs,
):
"""Function to plot a motion field as streamlines.
.. _`mpl_toolkits.basemap`: https://matplotlib.org/basemap/api/basemap_api.html#module-mpl_toolkits.basemap
.. _SubplotSpec: https://matplotlib.org/api/_as_gen/matplotlib.gridspec.SubplotSpec.html?highlight=subplotspec#matplotlib.gridspec.SubplotSpec
.. _cartopy: https://scitools.org.uk/cartopy/docs/latest/
Parameters
----------
UV : array-like
Array of shape (2, m,n) containing the input motion field.
ax : axis object
Optional axis object to use for plotting.
map : {'basemap', 'cartopy'}, optional
Optional method for plotting a map: 'basemap' or 'cartopy'.
The former uses `mpl_toolkits.basemap`_, while the latter uses cartopy_.
geodata : dictionary
Optional dictionary containing geographical information about
the field.
If geodata is not None, it must contain the following key-value pairs:
+----------------+----------------------------------------------------+
| Key | Value |
+================+====================================================+
| projection | PROJ.4-compatible projection definition |
+----------------+----------------------------------------------------+
| x1 | x-coordinate of the lower-left corner of the data |
| | raster |
+----------------+----------------------------------------------------+
| y1 | y-coordinate of the lower-left corner of the data |
| | raster |
+----------------+----------------------------------------------------+
| x2 | x-coordinate of the upper-right corner of the data |
| | raster |
+----------------+----------------------------------------------------+
| y2 | y-coordinate of the upper-right corner of the data |
| | raster |
+----------------+----------------------------------------------------+
| yorigin | location of the first element in the data raster |
| | element in the data raster w.r.t. y-axis: |
| | |
| | 'upper' = upper border, 'lower' = lower border |
+----------------+----------------------------------------------------+
drawlonlatlines : bool, optional
If set to True, draw longitude and latitude lines. Applicable if map is
'basemap' or 'cartopy'.
basemap_resolution : str, optional
The resolution of the basemap, see the documentation of
`mpl_toolkits.basemap`_.
Applicable if map is 'basemap'.
cartopy_scale : {'10m', '50m', '110m'}, optional
The scale (resolution) of the map. The available options are '10m',
'50m', and '110m'. Applicable if map is 'cartopy'.
lw: float, optional
Linewidth of the map (administrative boundaries and coastlines).
cartopy_subplot : tuple or SubplotSpec_ instance, optional
Cartopy subplot. Applicable if map is 'cartopy'.
axis : {'off','on'}, optional
Whether to turn off or on the x and y axis.
Other Parameters
----------------
density : float
Controls the closeness of streamlines.
Default : 1.5
color : string
Optional streamline color. This is a synonym for the PolyCollection
facecolor kwarg in matplotlib.collections.
Default : black
Returns
-------
out : axis object
Figure axes. Needed if one wants to add e.g. text inside the plot.
"""
if map is not None and geodata is None:
raise ValueError("map!=None but geodata=None")
# defaults
density = kwargs.get("density", 1.5)
color = kwargs.get("color", "black")
# prepare x y coordinates
reproject = False
if geodata is not None:
x = (
np.linspace(geodata["x1"], geodata["x2"], UV.shape[2])
+ geodata["xpixelsize"] / 2.0
)
y = (
np.linspace(geodata["y1"], geodata["y2"], UV.shape[1])
+ geodata["ypixelsize"] / 2.0
)
extent = (geodata["x1"], geodata["x2"], geodata["y1"], geodata["y2"])
# check geodata and project if different from axes
if ax is not None and map is None:
if type(ax).__name__ == "GeoAxesSubplot":
try:
ccrs = utils.proj4_to_cartopy(geodata["projection"])
except UnsupportedSomercProjection:
# Define fall-back projection for Swiss data(EPSG:3035)
# This will work reasonably well for Europe only.
t_proj4str = "+proj=laea +lat_0=52 +lon_0=10 +x_0=4321000 +y_0=3210000 +ellps=GRS80 +units=m +no_defs"
reproject = True
elif type(ax).__name__ == "Basemap":
utils.proj4_to_basemap(geodata["projection"])
if reproject:
geodata = utils.reproject_geodata(
geodata, t_proj4str, return_grid="coords"
)
extent = (geodata["x1"], geodata["x2"], geodata["y1"], geodata["y2"])
X, Y = geodata["X_grid"], geodata["Y_grid"]
else:
x = np.arange(UV.shape[2])
y = np.arange(UV.shape[1])
if not reproject:
X, Y = np.meshgrid(x, y)
# draw basemaps
if map is not None:
try:
ax = basemaps.plot_geography(
map,
geodata["projection"],
extent,
UV.shape[1:],
drawlonlatlines,
basemap_resolution,
cartopy_scale,
lw,
cartopy_subplot,
)
except UnsupportedSomercProjection:
# Define default fall-back projection for Swiss data(EPSG:3035)
# This will work reasonably well for Europe only.
t_proj4str = "+proj=laea +lat_0=52 +lon_0=10 +x_0=4321000 +y_0=3210000 +ellps=GRS80 +units=m +no_defs"
geodata = utils.reproject_geodata(geodata, t_proj4str, return_grid="coords")
extent = (geodata["x1"], geodata["x2"], geodata["y1"], geodata["y2"])
X, Y = geodata["X_grid"], geodata["Y_grid"]
ax = basemaps.plot_geography(
map,
geodata["projection"],
extent,
UV.shape[1:],
drawlonlatlines,
basemap_resolution,
cartopy_scale,
lw,
cartopy_subplot,
)
else:
ax = plt.gca()
# plot streamplot
ax.streamplot(
x,
np.flipud(y),
UV[0, :, :],
-UV[1, :, :],
density=density,
color=color,
zorder=1e6,
)
if geodata is None or axis == "off":
axes = plt.gca()
axes.xaxis.set_ticks([])
axes.xaxis.set_ticklabels([])
axes.yaxis.set_ticks([])
axes.yaxis.set_ticklabels([])
return plt.gca()
def plot_network(self,
state_sizes=None,
state_scale=1.0,
state_colors='#ff5500',
state_labels='auto',
arrow_scale=1.0,
arrow_curvature=1.0,
arrow_labels='weights',
arrow_label_format='%10.2f',
max_width=12,
max_height=12,
max_flux=None,
figpadding=0.2,
xticks=False,
yticks=False,
show_frame=False,
**textkwargs):
"""
Draws a network using discs and curved arrows.
The thicknesses and labels of the arrows are taken from the off-diagonal matrix elements in A.
"""
plt = self.plt
if self.pos is None:
self.layout_automatic()
# number of nodes
n = len(self.pos)
# get bounds and pad figure
xmin = np.min(self.pos[:, 0])
xmax = np.max(self.pos[:, 0])
Dx = xmax - xmin
xmin -= Dx * figpadding
xmax += Dx * figpadding
Dx *= 1 + figpadding
ymin = np.min(self.pos[:, 1])
ymax = np.max(self.pos[:, 1])
Dy = ymax - ymin
ymin -= Dy * figpadding
ymax += Dy * figpadding
Dy *= 1 + figpadding
# sizes of nodes
if state_sizes is None:
state_sizes = 0.5 * state_scale * \
min(Dx, Dy)**2 * np.ones(n) / float(n)
else:
state_sizes = 0.5 * state_scale * \
state_sizes / (np.max(state_sizes) * float(n))
#min(Dx, Dy)**2 * state_sizes / (np.max(state_sizes) * float(n)) # JFR
# automatic arrow rescaling **JFR - Soooo confusing, don't do this!
#arrow_scale *= 1.0 / \
# (np.max(self.A - np.diag(np.diag(self.A))) * math.sqrt(n))
# size figure
if (Dx / max_width > Dy / max_height):
figsize = (max_width, Dy * (max_width / Dx))
else:
figsize = (Dx / Dy * max_height, max_height)
fig = plt.gcf()
fig.set_size_inches(figsize, forward=True)
# font sizes
from matplotlib import rcParams
old_fontsize = rcParams['font.size']
rcParams['font.size'] = 20
# remove axis labels
frame = plt.gca()
if not xticks:
frame.axes.get_xaxis().set_ticks([])
if not yticks:
frame.axes.get_yaxis().set_ticks([])
# show or suppress frame
frame.set_frame_on(show_frame)
# set node labels
if state_labels is 'auto':
state_labels = [str(i) for i in np.arange(n)]
else:
assert len(
state_labels
) == n, "Mistmatch between nstates and nr. state_labels (%u vs %u)" % (
n, len(state_labels))
# set node colors
if state_colors is None:
state_colors = '#ff5500' # None is not acceptable
if isinstance(state_colors, str):
state_colors = [state_colors] * n
if isinstance(state_colors, list):
assert len(
state_colors
) == n, "Mistmatch between nstates and nr. state_colors (%u vs %u)" % (
n, len(state_colors))
try:
colorscales = _types.ensure_ndarray(state_colors,
ndim=1,
kind='numeric')
colorscales /= colorscales.max()
state_colors = [
plt.cm.binary(int(256.0 * colorscales[i])) for i in range(n)
]
except:
pass # assume we have a list of strings now.
# set arrow labels
if isinstance(arrow_labels, np.ndarray):
L = arrow_labels
else:
L = np.empty(np.shape(self.A), dtype=object)
if arrow_labels is None:
L[:, :] = ''
elif arrow_labels.lower() == 'weights':
for i in range(n):
for j in range(n):
L[i, j] = arrow_label_format % self.A[i, j]
else:
rcParams['font.size'] = old_fontsize
raise ValueError('invalid arrow label format')
# Set the default values for the text dictionary
textkwargs.setdefault('size', 14)
textkwargs.setdefault('horizontalalignment', 'center')
textkwargs.setdefault('verticalalignment', 'center')
textkwargs.setdefault('color', 'black')
# draw circles
circles = []
for i in range(n):
fig = plt.gcf()
# choose color
c = plt.Circle(self.pos[i],
radius=math.sqrt(0.5 * state_sizes[i]) / 2.0,
color=state_colors[i],
zorder=2)
circles.append(c)
fig.gca().add_artist(c)
# add annotation
plt.text(self.pos[i][0],
self.pos[i][1],
state_labels[i],
zorder=3,
**textkwargs)
assert len(circles) == n, "%i != %i" % (len(circles), n)
# draw arrows
# my own colormap
from matplotlib import pyplot as plt
from matplotlib.pyplot import *
# define the colormap
#print self.A
#print np.all(self.A >= 0)
if (np.all(self.A >= 0)):
mycmap = plt.cm.Greys
#mycmap = plt.cm.winter
mycmap_max = np.max(np.abs(self.A))
mycmap_min = -1. * mycmap_max #0. # np.min(self.A[self.A != 0])
else:
mycmap = plt.cm.bwr
#mycmap = plt.cm.jet
if (max_flux is None):
mycmap_max = np.max(np.abs(self.A))
mycmap_min = -mycmap_max
else:
mycmap_max = max_flux
mycmap_min = -mycmap_max
# extract all colors from the .jet map
mycmaplist = [mycmap(i) for i in range(mycmap.N)]
# create the new map
mycmap = mycmap.from_list('Custom cmap', mycmaplist, mycmap.N)
# define the bins and normalize
bounds = np.linspace(mycmap_min, mycmap_max, mycmap.N)
norm = matplotlib.colors.BoundaryNorm(bounds, mycmap.N)
mycmaplist = [mycmap(i) for i in range(mycmap.N)]
dx = bounds[1] - bounds[0]
for i in range(n):
for j in range(i + 1, n):
if (abs(self.A[i, j]) > 0):
# JFR - let's allow for neg delta-F
grid = int((self.A[i, j] - mycmap_min) / dx + 0.5)
color = mycmaplist[grid]
self._draw_arrow(self.pos[i, 0],
self.pos[i, 1],
self.pos[j, 0],
self.pos[j, 1],
Dx,
Dy,
label=str(L[i, j]),
width=arrow_scale * abs(self.A[i, j]),
color=color,
arrow_curvature=arrow_curvature,
patchA=circles[i],
patchB=circles[j],
shrinkA=3,
shrinkB=0)
if (abs(self.A[j, i]) > 0):
grid = int((self.A[j, i] - mycmap_min) / dx + 0.5)
color = mycmaplist[grid]
self._draw_arrow(self.pos[j, 0],
self.pos[j, 1],
self.pos[i, 0],
self.pos[i, 1],
Dx,
Dy,
label=str(L[j, i]),
width=arrow_scale * abs(self.A[j, i]),
color=color,
arrow_curvature=arrow_curvature,
patchA=circles[j],
patchB=circles[i],
shrinkA=3,
shrinkB=0)
# plot
plt.xlim(xmin, xmax)
plt.ylim(ymin, ymax)
rcParams['font.size'] = old_fontsize
return fig
def _plot_seasonality(self, alpha: float, plot_kwargs: bool):
# two_tailed_alpha = int(alpha / 2 * 100)
periods = list(set([float(i.split("_")[1]) for i in self.seasonality]))
additive_ts, multiplicative_ts = self._fit_seasonality()
all_seasonalities = [("additive", additive_ts)]
if len(self.multiplicative_data):
all_seasonalities.append(("multiplicative", multiplicative_ts))
for sn, ts in all_seasonalities:
if (sn == "multiplicative"
and np.sum(ts) == 1) or (sn == "additive"
and np.sum(ts) == 0):
continue
ddf = pd.DataFrame(
np.vstack([
np.percentile(ts[:, :, self.skip_first:], 50, axis=-1),
np.percentile(ts[:, :, self.skip_first:],
alpha / 2 * 100,
axis=-1),
np.percentile(ts[:, :, self.skip_first:],
(1 - alpha / 2) * 100,
axis=-1),
]).T,
columns=[
"%s_%s" % (p, l) for l in ["mid", "low", "high"]
for p in periods[::-1]
],
)
ddf.loc[:, "ds"] = self.data["ds"]
for period in periods:
if int(period) == 0:
step = int(self.data["ds"].diff().mean().total_seconds() //
float(period))
else:
step = int(period)
graph = ddf.head(step)
if period == 7:
ddf.loc[:, "dow"] = [i for i in ddf["ds"].dt.weekday]
graph = (ddf[[
"dow",
"%s_low" % period,
"%s_mid" % period,
"%s_high" % period,
]].groupby("dow").mean().sort_values("dow"))
graph.loc[:, "ds"] = [[
"Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"
][i] for i in graph.index]
graph = graph.sort_index()
plt.figure(**plot_kwargs)
graph.plot(y="%s_mid" % period,
x="ds",
color="C0",
legend=False,
ax=plt.gca())
plt.grid()
if period == 7:
plt.xticks(range(7), graph["ds"].values)
plt.fill_between(
np.arange(0, 7),
graph["%s_low" % period].values.astype(float),
graph["%s_high" % period].values.astype(float),
alpha=0.3,
)
else:
plt.fill_between(
graph["ds"].values,
graph["%s_low" % period].values.astype(float),
graph["%s_high" % period].values.astype(float),
alpha=0.3,
)
plt.title("Model Seasonality (%s) for period: %s days" %
(sn, period))
plt.axes().xaxis.label.set_visible(False)
plt.show()
def peerSensitivity():
space = 1
af = 4
allasCount = {}
resultsFile = "../results/peerSensitivity/KLdiv_%s.pickle"
collectorsDataFile = "../results/peerSensitivity/collectorData.pickle"
allPeersFile = "../results/peerSensitivity/allPeersDist.pickle"
collectorsDist = []
results = defaultdict(list)
if not os.path.exists(resultsFile % "Betweenness"):
ribFiles = glob.glob("/data/routeviews/archive.routeviews.org/*/*/RIBS/rib.20160601.0000.bz2")
ribFiles.extend(glob.glob("/data/routeviews/archive.routeviews.org/*/*/*/RIBS/rib.20160601.0000.bz2"))
ribFiles.extend(glob.glob("/data/ris/*/*/bview.20160601.0000.gz"))
ribFiles.append("/data/bgpmon/ribs/201606/ribs")
for i, ribFile in enumerate(ribFiles):
words = ribFile.split("/")
if "routeviews" in ribFile:
if words[-4] == "route-views3":
label = "rv3"
elif words[-5] == "archive.routeviews.org" and words[-4] == "bgpdata":
label = "rv2"
elif not "." in words[-5] and words[-5].startswith("route-views"):
label = "rv"+words[-5][-1]
else:
label = words[-5].split(".")[-1]
elif "ris" in ribFile:
label = words[-3]
else:
label = "bgpmon"
asCountFile = "../results/peerSensitivity/20160601.0000_asCount%s.pickle" % (i)
if not os.path.exists(asCountFile):
rtree, _ = ashash.readrib(ribFile, space, af)
asCount = rtree.search_exact("0.0.0.0/0").data
asHegemony, _, nbPeers = ashash.computeCentrality(asCount, space)
# asBetweenness = ashash.computeBetweenness(asCount, space)
pickle.dump((asCount, asHegemony, nbPeers), open(asCountFile, "wb"))
else:
asCount, asHegemony, nbPeers = pickle.load(open(asCountFile,"rb"))
collectorsDist.append( (label, nbPeers, asHegemony) )
print "%s: %s peers" % (label, len(asCount))
for peer, count in asCount.iteritems():
if count["totalCount"]>2000000000:
if not peer in allasCount:
allasCount[peer] = count
else:
print "Warning: peer %s is observed multiple times (%s)" % (peer, ribFile)
asHegemonyRef, _, nbPeers = ashash.computeCentrality(allasCount, space)
asBetweennessRef, _, _ = ashash.computeBetweenness(allasCount, space)
pickle.dump((asHegemonyRef, asBetweennessRef, nbPeers), open(allPeersFile,"wb"))
for metricLabel, ref, computeMetric in [("Hegemony", asHegemonyRef, ashash.computeCentrality),
("Betweenness", asBetweennessRef, ashash.computeBetweenness)]:
if not os.path.exists(resultsFile % metricLabel):
# Remove AS with a score of 0.0
toremove = [asn for asn, score in ref.iteritems() if score==0.0]
for asn in toremove:
del ref[asn]
minVal = min(ref.values())
nbPeersList = range(0, len(allasCount), 10)
nbPeersList[0] = 1
for nbPeers in nbPeersList:
tmp = []
for _ in range(10):
# Randomly select peers
peersIndex = random.sample(range(len(allasCount)), nbPeers)
asCount = {}
for p in peersIndex:
asCount[allasCount.keys()[p]] = allasCount.values()[p]
asMetric, _, _ = computeMetric(asCount, space)
# Remove AS with a score == 0.0
toremove = [asn for asn, score in asMetric.iteritems() if score==0.0]
if not toremove is None:
for asn in toremove:
del asMetric[asn]
# Set the same number of ASN for both distributions
missingAS = set(ref.keys()).difference(asMetric.keys())
if not missingAS is None:
for asn in missingAS:
asMetric[asn] = minVal
# Compute the KL-divergence
dist = [asMetric[asn] for asn in ref.keys()]
kldiv = sps.entropy(dist, ref.values())
tmp.append(kldiv)
results[metricLabel].append(tmp)
print tmp
# save final results
pickle.dump((nbPeersList, results[metricLabel]),open(resultsFile % metricLabel,"wb"))
pickle.dump(collectorsDist, open(collectorsDataFile,"wb"))
else:
(nbPeersList, results[metricLabel]) = pickle.load(open(resultsFile % metricLabel,"rb"))
collectorsDist = pickle.load(open(collectorsDataFile,"rb"))
else:
for metricLabel in ["Hegemony", "Betweenness"]:
nbPeersList, results[metricLabel] = pickle.load(open(resultsFile % metricLabel,"rb"))
collectorsDist = pickle.load(open(collectorsDataFile,"rb"))
asHegemonyRef, asBetweennessRef, allFullFeedPeers = pickle.load(open(allPeersFile,"rb"))
#return asHegemonyRef
def plotRef(references, legendFmt="%s", alpha=1.0):
for asDistRef, metricLabel, color in references:
m = np.mean(results[metricLabel][1:], axis=1)
s = np.std(results[metricLabel][1:], axis=1)
# mi = m-np.min(results[metricLabel][1:], axis=1)
# ma = np.max(results[metricLabel][1:], axis=1)-m
# mi, ma = sms.DescrStatsW(results[metricLabel][1:]).tconfint_mean(alpha=0.1)
mi = np.min(results[metricLabel][1:], axis=1)
ma = np.max(results[metricLabel][1:], axis=1)
x = nbPeersList[1:]
plt.fill_between(x, mi, ma, facecolor=color, alpha=alpha*0.5)
# plt.plot(x, m,"-+", ms=4, color="0.6", label="Randomly selected")
# plt.plot(x, m,"-+", ms=4, color="0.6", label="Random peers (%s)" % metricLabel)
try:
plt.plot(x, m,"-o", ms=3, label=legendFmt % metricLabel, color=color, alpha=alpha)
except TypeError:
plt.plot(x, m,"-o", ms=3, label=legendFmt, color=color, alpha=alpha)
# plt.errorbar(x,m, [mi, ma], fmt="C3.", ms=4)
plt.xlabel("Number of peers")
plt.ylabel("KL divergence")
# plt.yscale("log")
plt.legend()
plt.tight_layout()
# Compare betweenness and hegemony
plt.figure()
plotRef([(asHegemonyRef, "Hegemony", "C0"), (asBetweennessRef, "Betweenness", "C5")])
plt.ylim([0.00001, 1])
plt.xscale("log")
plt.savefig("../results/peerSensitivity/meanKL_%s.pdf" % metricLabel)
# Compare collectors and random sample (hegemony):
# Remove AS with a score of 0.0
toremove = [asn for asn, score in asHegemonyRef.iteritems() if score==0.0]
for asn in toremove:
del asHegemonyRef[asn]
minVal = min(asHegemonyRef.values())
plt.figure()
plotRef([(asHegemonyRef, "Hegemony", "C0")], legendFmt="Random peers", alpha=0.5)
plt.xlabel("Number of peers")
plt.ylabel("KL divergence")
for collectorLabel, nbPeers, asHegemony in collectorsDist:
if asHegemony is None :
print "warning: ignore collector %s" % collectorLabel
continue
# Remove AS with a score == 0.0
toremove = [asn for asn, score in asHegemony.iteritems() if score==0.0]
if not toremove is None:
for asn in toremove:
del asHegemony[asn]
# Set the same number of ASN for both distributions
missingAS = set(asHegemonyRef.keys()).difference(asHegemony.keys())
if not missingAS is None:
for asn in missingAS:
asHegemony[asn] = minVal
# Compute the KL-divergence
dist = [asHegemony[asn] for asn in asHegemonyRef.keys()]
kldiv = sps.entropy(dist, asHegemonyRef.values())
print "%s:\t %s peers \t %s " % (collectorLabel, nbPeers, kldiv)
if kldiv>0.4 :
continue
if collectorLabel.startswith("rrc"):
plt.plot(nbPeers, kldiv,"C1x", label="RIS")
collectorLabel = collectorLabel.replace("rrc","")
elif collectorLabel == "bgpmon":
plt.plot(nbPeers, kldiv,"C3^")
else:
plt.plot(nbPeers, kldiv,"C3+", label="Route Views")
if kldiv<1 or nbPeers>10:
plt.text(nbPeers+0.005, kldiv+0.005, collectorLabel, fontsize=8)
# plt.yscale("log")
# plt.xscale("log")
handles, labels = plt.gca().get_legend_handles_labels()
by_label = OrderedDict(zip(labels, handles))
plt.legend(by_label.values(), by_label.keys())
plt.ylim([0.00001, 0.4])
plt.xlim([9, 50])
plt.tight_layout()
plt.savefig("../results/peerSensitivity/collectorDiversity.pdf")
return (nbPeersList, results)
def predict(
self,
forecasting_periods: int = 10,
freq: str = "D",
extra_data: pd.DataFrame = None,
include_history: bool = True,
alpha: float = 0.05,
plot: bool = False,
):
"""Predict using the PMProphet model.
Parameters
----------
forecasting_periods : int, > 0
Number of future points to forecast
freq : string, default: 'D'
extra_data : pd.DataFrame
include_history : bool
If True, predictions are concatenated to the data.
alpha : float
Width of the the credible intervals.
plot : bool
Plot the predictions.
Returns
-------
A pd.DataFrame with the forecast components.
"""
if self.auto_changepoints:
self._finalize_auto_changepoints(plot=False)
last_date = self.data["ds"].max()
dates = pd.date_range(
start=last_date,
periods=forecasting_periods +
1, # An extra in case we include start
freq=freq,
)
dates = dates[dates > last_date] # Drop start if equals last_date
dates = dates[:forecasting_periods] # Return correct number of periods
new_df = pd.DataFrame()
if include_history:
new_df["y"] = np.concatenate(
[self.data["y"],
np.zeros(forecasting_periods) * np.nan])
new_df["ds"] = np.concatenate([self.data["ds"], dates])
else:
new_df["y"] = np.zeros(forecasting_periods)
new_df["ds"] = dates
for regressor in self.regressors:
new_df[regressor] = self.data[regressor]
if extra_data is not None:
for column in extra_data.columns:
if column not in ["y", "ds"]:
new_df[column] = extra_data[column]
m = PMProphet(
data=new_df,
growth=self.growth,
intercept=self.intercept,
model=self.model,
name=self.name,
)
m.changepoints = self.changepoints
periods = {}
for column in self.data.columns:
if column.startswith("f_"):
period, order = column[2:].split("_")
periods.setdefault(period, [])
periods[period].append(int(order))
for period, orders in periods.items():
m.add_seasonality(seasonality=float(period),
fourier_order=max(orders) + 1)
m.priors = self.priors
m.priors_names = self.priors_names
m.trace = self.trace
m.multiplicative_data = self.multiplicative_data
draws = max(self.trace[var].shape[-1] for var in self.trace.varnames
if "hat_{}".format(self.name) not in var)
if self.growth:
# Start with the trend
y_hat = m._fit_growth(prior=False)
else:
y_hat = np.zeros((len(m.data.ds.values), draws))
multiplicative_seasonality = np.zeros((len(m.data.ds.values), draws))
additive_seasonality = np.zeros((len(m.data.ds.values), draws))
if self.seasonality:
# Add seasonality
additive_seasonality, multiplicative_seasonality = m._fit_seasonality(
flatten_components=True)
if self.intercept:
# Add intercept
y_hat += self.trace[self.priors_names["intercept"]]
# Add regressors
multiplicative_regressors = np.zeros((len(m.data.ds.values), draws))
additive_regressors = np.zeros((len(m.data.ds.values), draws))
for idx, regressor in enumerate(self.regressors):
trace = m.trace[m.priors_names["regressors"]][:, idx]
if regressor in self.multiplicative_data:
multiplicative_regressors += trace * np.repeat(
[m.data[regressor]], len(trace)).reshape(
len(m.data), len(trace))
else:
additive_regressors += trace * np.repeat(
[m.data[regressor]], len(trace)).reshape(
len(m.data), len(trace))
# Add holidays
additive_holidays = np.zeros((len(m.data.ds.values), draws))
multiplicative_holidays = np.zeros((len(m.data.ds.values), draws))
for idx, holiday in enumerate(self.holidays):
trace = m.trace[m.priors_names["holidays"]][:, idx]
if holiday in self.multiplicative_data:
multiplicative_holidays += trace * np.repeat(
[m.data[holiday]], len(trace)).reshape(
len(m.data), len(trace))
else:
additive_holidays += trace * np.repeat(
[m.data[holiday]], len(trace)).reshape(
len(m.data), len(trace))
if (np.sum(multiplicative_holidays + multiplicative_seasonality +
multiplicative_regressors) == 0):
multiplicative_term = 1
else:
multiplicative_term = (multiplicative_holidays +
multiplicative_seasonality +
multiplicative_regressors)
y_hat *= multiplicative_term
y_hat += additive_seasonality + additive_holidays + additive_regressors
y_hat_noised = np.random.normal(
y_hat[:, self.skip_first:], self.data['y'].std() *
self.trace[self.priors_names["sigma"]][self.skip_first:])
ddf = pd.DataFrame([
np.percentile(y_hat_noised, 50, axis=-1),
np.percentile(y_hat_noised, alpha / 2 * 100, axis=-1),
np.percentile(y_hat_noised, (1 - alpha / 2) * 100, axis=-1),
]).T
ddf["ds"] = m.data["ds"]
ddf.columns = ["y_hat", "y_low", "y_high", "ds"]
if plot:
plt.figure(figsize=(20, 10))
ddf.plot("ds", "y_hat", ax=plt.gca())
ddf["orig_y"] = self.data["y"]
plt.fill_between(
ddf["ds"].values,
ddf["y_low"].values.astype(float),
ddf["y_high"].values.astype(float),
alpha=0.3,
)
ddf.plot("ds", "orig_y", style="k.", ax=plt.gca(), alpha=0.2)
for idx, change_point in enumerate(self.changepoints):
if self.auto_changepoints:
plt.axvline(
change_point,
color="C2",
lw=1,
ls="dashed",
alpha=self.changepoint_weights[idx],
)
else:
plt.axvline(change_point, color="C2", lw=1, ls="dashed")
plt.axvline(pd.to_datetime(self.data.ds).max(),
color="C3",
lw=1,
ls="dotted")
plt.grid(axis="y")
plt.show()
return ddf
##########################################################
# plot energies
plt.subplot(3, 1, 2)
plt.plot(data2[:, 0], data2[:, 1], ':', label='potential energy', alpha=0.5)
plt.plot(data2[:, 0], data2[:, 2], ':', label='kinetic energy', alpha=0.5)
plt.plot(data2[:, 0], data2[:, 3], '-', label='total energy')
#plt.plot(data2[:,0], np.ones((data2.shape[0],1)) * (data1.shape[1] - 1),'-',label='total energy')
# labels
plt.xlabel(r'time $t$', fontsize=12)
plt.ylabel(r'total energies', fontsize=12)
# legend
plt.legend()
leg = plt.gca().get_legend()
ltext = leg.get_texts()
plt.setp(ltext, fontsize=12)
# axis limits
#plt.xlim([0,10000])
# tick fontsize
plt.xticks(fontsize=12)
plt.yticks(fontsize=12)
##########################################################
# plot energy modes
plt.subplot(3, 1, 3)
for i in range(1, data3.shape[1]):
plt.plot(data3[:, 0], data3[:, i], '-', alpha=0.5, label='mode ' + str(i))
