def define(self):
self.con = ConnectionPatch(lw=1.5,
color=(.25, .25, 1),
picker=10,
xyA=(self.x0, self.y0),
coordsA='data',
axesA=self.ax0,
xyB=(self.xx, self.yy),
coordsB='data',
axesB=self.axx)
self.con.parent = self
super(BlitFigure, self.master).add_artist(self.con)
self.master.add_artist(self.con)
self.scat += [
self.axx.scatter(self.xx,
self.yy,
c=[],
s=200,
linewidth=1.5,
edgecolors=(.25, .25, 1))
]
self.master.add_artist(self.scat[-1])
self.master.del_artist(self.line)
del self.line
def __init__(self, outputPlot):
self.decayValue = 0
self.outputPlot = outputPlot
self.measurementPatch = Rectangle((0, 0),
1,
1,
linewidth=1,
color='gray',
facecolor='none')
self.statePatch = Circle((0, 0), 1, facecolor='blue')
self.speedPatch = Arrow(0, 0, 3, 3)
self.circlePatch = Circle((0, 0),
1,
linewidth=1,
color='red',
facecolor='None')
self.linePatch = ConnectionPatch((0, 0), (1, 1), 'data')
self.model = 0
self.startLine = (0, 0)
self.needToResetStartLine = True
self.outputPlot.add_patch(self.measurementPatch)
self.outputPlot.add_patch(self.statePatch)
self.outputPlot.add_patch(self.speedPatch)
self.outputPlot.add_patch(self.circlePatch)
self.outputPlot.add_patch(self.linePatch)
self.circlePatch.set_visible(False)
self.linePatch.set_visible(False)
def connect_plots(ax0, ax1, x, y, delta):
xy0 = (x + delta, y + delta)
xy1 = (x - delta, y + delta)
xy2 = (x + delta, y - delta)
xy3 = (x - delta, y - delta)
con0 = ConnectionPatch(xyA=xy0,
xyB=xy1,
coordsA="data",
coordsB="data",
axesA=ax0,
axesB=ax1,
color="m")
ax0.add_artist(con0)
con1 = ConnectionPatch(xyA=xy2,
xyB=xy3,
coordsA="data",
coordsB="data",
axesA=ax0,
axesB=ax1,
color="m")
ax0.add_artist(con1)
return con0, con1
def step_4(t,resonance,FID,fig,ax_one,ax_FID,ax_mult,ax_FT,freqs,nb_freq,trial,mult,integ,posinteg,neginteg,n):
con_3 = ConnectionPatch(xyB=(1.2,0.5), xyA=(-0.2,0.5), coordsA="axes fraction", coordsB="axes fraction",axesA=ax_mult, axesB=ax_FT,color="red")
con_3.set_arrowstyle("simple",head_length=0.5, head_width=1, tail_width=0.3)
ax_mult.add_patch(con_3)
ax_mult.fill_between(t,mult[n],where=mult[n]>0,interpolate=True,color="blue",label="Positive Areas")
ax_mult.fill_between(t,mult[n],where=mult[n]<0,interpolate=True,color="red",label="Negative Areas")
if ax_mult.get_legend() is None:
ax_mult.legend(loc=1)
ax_FT.plot(freqs[:n+1],integ[:n+1],linewidth=1,color="blue")
ax_FT.plot([freqs[n]],[integ[n]],"ro")
ann=ax_FT.annotate("Trial frequency {} Hz\nTotal area {}".format(round(freqs[n],1),round(round(posinteg[n],3)+round(neginteg[n],3),3)), xy=(freqs[n],integ[n]), xycoords='data',
xytext=(-25, 10), textcoords='offset points',
arrowprops=dict(facecolor='black',shrink=0.1,width=2,headwidth=10),
horizontalalignment='right', verticalalignment='bottom')
session["message"]="""The area under the multiplication curve (over the acquisition time) for a given trial frequency represents the relative intensity of the signal at this frequency.<br><br>
For instance, the calculation for the trial frequency {} Hz from your FID gives (see graph 3):<br><br>
<div style="background-color:white;font-weight:bold">Total area ( = Relative intensity) = <br><span style="color:blue">Positive areas</span> + <span style="color:red">Negative areas </span>= <span style="color:blue">({})</span> + <span style="color:red">({})</span> = {} area units</div><br>
This point is reported on the final spectrum presented in graph 4.<br><br>
The computer will repeat this operation with trial frequencies from {} to {} Hz in order to cover all our range of frequencies.<br><br>
Press OK to access the spectrum generation (loading of the animation can take up to 1 min).<br><input type=submit name="msgOK" value="OK" class="control">""".format(round(freqs[n],1),round(posinteg[n],3),round(neginteg[n],3),round(round(posinteg[n],3)+round(neginteg[n],3),3),session["freq_min"]-1,session["freq_max"]+1)
session["FT_step"]=4
return t,resonance,FID,fig,ax_one,ax_FID,ax_mult,ax_FT,freqs,nb_freq,trial,mult,integ,posinteg,neginteg,n,ann
def plot_warp(from_s, to_s, new_s, path, filename=None, fig=None, axs=None):
"""Plot the warped sequence and its relation to the original sequence
and the target sequence.
:param from_s: From sequence.
:param to_s: To sequence.
:param new_s: Warped version of from sequence.
:param path: Optimal warping path.
:param filename: Filename path (optional).
:param fig: Matplotlib Figure object
:param axs: Array of Matplotlib axes.Axes objects (length == 3)
:return: Figure, list[Axes]
"""
try:
import matplotlib.pyplot as plt
import matplotlib as mpl
from matplotlib.patches import ConnectionPatch
except ImportError:
logger.error(
"The plot_warp function requires the matplotlib package to be installed."
)
return
if fig is None and axs is None:
fig, axs = plt.subplots(nrows=3, ncols=1, sharex='all', sharey='all')
elif fig is None or axs is None:
raise TypeError(
f'The fig and axs arguments need to be both None or both instantiated.'
)
axs[0].plot(from_s, label="From")
axs[0].legend()
axs[1].plot(to_s, label="To")
axs[1].legend()
lines = []
line_options = {'linewidth': 0.5, 'color': 'orange', 'alpha': 0.8}
for r_c, c_c in path:
if r_c < 0 or c_c < 0:
continue
con = ConnectionPatch(xyA=[r_c, from_s[r_c]],
coordsA=axs[0].transData,
xyB=[c_c, to_s[c_c]],
coordsB=axs[1].transData,
**line_options)
lines.append(con)
axs[2].plot(new_s, label="From-warped")
axs[2].legend()
for i in range(len(to_s)):
con = ConnectionPatch(xyA=[i, to_s[i]],
coordsA=axs[1].transData,
xyB=[i, new_s[i]],
coordsB=axs[2].transData,
**line_options)
lines.append(con)
for line in lines:
fig.add_artist(line)
if filename:
plt.savefig(filename)
plt.close()
fig, axs = None, None
return fig, axs
def _connect_axes(self, ax1, ax2):
"""Draws an arrow from ax1 to ax2.
"""
axis_center = self.get_cell_center()
con = ConnectionPatch(xyA=axis_center,
xyB=axis_center,
coordsA="data",
coordsB="data",
axesA=ax1,
axesB=ax2,
color="red",
mutation_scale=40,
arrowstyle="->",
shrinkB=5,
shrinkA=5)
ax1.add_artist(con)
con = ConnectionPatch(xyA=axis_center,
xyB=axis_center,
coordsA="data",
coordsB="data",
axesA=ax2,
axesB=ax1,
color="red",
mutation_scale=40,
arrowstyle="<-",
shrinkB=5,
shrinkA=5)
ax2.add_artist(con)
ax1.plot(*axis_center, 'ro', markersize=10)
ax2.plot(*axis_center, 'ro', markersize=10)
def plot_image_matching(x1, y1, x2, y2, bx1, by1, bx2, by2, img1, img2):
f = plt.figure(figsize=(30, 30))
ax1 = f.add_subplot(121)
ax2 = f.add_subplot(122)
ax1.imshow(img1)
ax1.axis('off')
ax2.imshow(img2)
ax2.axis('off')
ax1.plot(x1, y1, 'ro', markersize=3)
ax2.plot(x2, y2, 'ro', markersize=3)
ax1.plot(bx1, by1, 'bo', markersize=3)
ax2.plot(bx2, by2, 'bo', markersize=3)
for i in range(len(x1)):
con = ConnectionPatch(xyA=(x2[i], y2[i]),
xyB=(x1[i], y1[i]),
coordsA="data",
coordsB="data",
axesA=ax2,
axesB=ax1,
color="red")
ax2.add_artist(con)
for i in range(len(bx1)):
con = ConnectionPatch(xyA=(bx2[i], by2[i]),
xyB=(bx1[i], by1[i]),
coordsA="data",
coordsB="data",
axesA=ax2,
axesB=ax1,
color="blue")
ax2.add_artist(con)
def plot_synth_data_with_gates(models, concatenated_data, hparams, plot_config={}):
plt.rcParams.update({'font.size': 20, 'font.family': 'serif'})
DEFAULT_PLOT_CONFIG.update(plot_config)
plot_config = DEFAULT_PLOT_CONFIG
features = ['M1', 'M2', 'M3', 'M4', 'M5', 'M6', 'M7', 'M8']
if plot_config['ncols'] < 7:
raise ValueError('Need at least seven columns to display tree')
input = SynthInput(hparams)
fig = plt.figure(figsize=plot_config['figsize'])
plt.tight_layout()
grid = gridspec.GridSpec(nrows=3, ncols=plot_config['ncols'], figure=fig)
grid.update(hspace=plot_config['hspace'], wspace=plot_config['wspace'])
axes = []
axes.append(fig.add_subplot(grid[0,plot_config['ncols']//2]))
axes.append(fig.add_subplot(grid[1,plot_config['ncols']//2 - 1]))
axes.append(fig.add_subplot(grid[1,plot_config['ncols']//2 + 1]))
axes.append(fig.add_subplot(grid[2,plot_config['ncols']//2 + 2]))
fig.suptitle(plot_config['title'] ,x=.53, y=0, fontsize=25)
for i in range(len(axes)):
axes[i].set_xlim(right=plot_config['right'])
axes[i].set_ylim(top=plot_config['top'])
axes[i].set_xticks([], [])
axes[i].set_yticks([], [])
axes[0].set_ylabel('M2')
axes[0].set_xlabel('M1')
con = ConnectionPatch(xyA=(0., 0.), xyB=(.5, 1.), coordsA="axes fraction", \
coordsB="axes fraction", axesA=axes[0], axesB=axes[1], color='k', arrowstyle='-|>')
axes[0].add_artist(con)
axes[1].set_ylabel('M4')
axes[1].set_xlabel('M3')
con = ConnectionPatch(xyA=(1., 0.), xyB=(.5, 1.), coordsA="axes fraction", \
coordsB="axes fraction", axesA=axes[0], axesB=axes[2], color='k', arrowstyle='-|>')
axes[0].add_artist(con)
axes[2].set_ylabel('M6')
axes[2].set_xlabel('M5')
con = ConnectionPatch(xyA=(1., 0.), xyB=(.5, 1.), coordsA="axes fraction", \
coordsB="axes fraction", axesA=axes[2], axesB=axes[3], color='k', arrowstyle='-|>')
axes[2].add_artist(con)
axes[3].set_ylabel('M8')
axes[3].set_xlabel('M7')
flat_axes = np.array([axes[0], axes[2], axes[3], axes[1]])
for axis in flat_axes:
axis.set_ylim(0, 1)
axis.set_xlim(0, 1)
init_gate_plotter = DataAndGatesPlotter(models['init'], concatenated_data)
init_gate_plotter.plot_on_axes(flat_axes, hparams)
final_gates_and_data_plotter = DataAndGatesPlotter(models['final'], concatenated_data, color='r')
final_gates_and_data_plotter.plot_on_axes(flat_axes, hparams)
def plot_2dmap(self, data=None):
""" Plots the distribution of the map in two dimensions,
including the connections for neighbouring neurons.
Only available if the dimensionality of the map is 2.
Args:
- data: (optional) an m by N array. If included, the
map is plotted together with the provided data.
"""
if self.m != 2:
raise ValueError("SOM dimensionality is not two")
# Plot points first
plt.figure()
if data is not None:
plt.scatter(data[0, :], data[1, :], alpha=0.5, edgecolors="none")
plt.scatter(self.neurons[0, :], self.neurons[1, :], s=10, c="k")
# Add the lines
axis = plt.gca()
i = 0
j = 0
for i in range(self.h - 1):
for j in range(self.w - 1):
# Connect to neurons in front of and below the neuron
neuron_idx = self.neurons[:,
np.ravel_multi_index(
(i, j), (self.h, self.w))]
neuron_front = self.neurons[:,
np.ravel_multi_index(
(i, j + 1), (self.h, self.w))]
neuron_below = self.neurons[:,
np.ravel_multi_index(
(i + 1, j), (self.h, self.w))]
axis.add_patch(
ConnectionPatch(neuron_idx, neuron_front, "data"))
axis.add_patch(
ConnectionPatch(neuron_idx, neuron_below, "data"))
# For the last column, connect to neuron below
neuron_idx = self.neurons[:,
np.ravel_multi_index((i, j + 1),
(self.h, self.w))]
neuron_below = self.neurons[:,
np.ravel_multi_index((i + 1, j + 1),
(self.h, self.w))]
axis.add_patch(ConnectionPatch(neuron_idx, neuron_below, "data"))
for j in range(self.w - 1):
# For the last row, connect to neurons in front of the neuron
neuron_idx = self.neurons[:,
np.ravel_multi_index((self.h - 1, j),
(self.h, self.w))]
neuron_front = self.neurons[:,
np.ravel_multi_index(
(self.h - 1, j + 1),
(self.h, self.w))]
axis.add_patch(ConnectionPatch(neuron_idx, neuron_front, "data"))
plt.show(block=False)
def add_connection(source_ax, source_data_coord, target_ax, target_data_coord):
con = ConnectionPatch(xyA=source_data_coord, xyB=target_data_coord,
coordsA='data', coordsB='data',
axesA=source_ax, axesB=target_ax,
arrowstyle='->', clip_on=False, linewidth=3)
con.set_zorder(10)
source_ax.add_artist(con)
def __init__(self, px_i, px_f, width):
self.px_i = px_i
self.px_f = px_f
self.px_width = width
# Find distance in pixels of line profile
self.px_dist = int(
np.round(
np.hypot(self.px_f[0] - self.px_i[0],
self.px_f[1] - self.px_i[1])))
self.s_dist = None
# Calculate the angle the line makes with the x-axis
self.line_vec = np.array(
[self.px_f[0] - self.px_i[0], self.px_f[1] - self.px_i[1]])
self.angle = np.angle(self.line_vec[0] + self.line_vec[1] * 1j,
deg=False)
# Calculate the offset in X and Y for the linewidth start
# and end points at the start point for the line profile
self.xyA_i = (
(self.px_i[0] - width / 2 * np.sin(self.angle)),
(self.px_i[1] + width / 2 * np.cos(self.angle)),
)
self.xyB_i = (
(self.px_i[0] + width / 2 * np.sin(self.angle)),
(self.px_i[1] - width / 2 * np.cos(self.angle)),
)
self.cpatch_i = ConnectionPatch(
xyA=self.xyA_i,
xyB=self.xyB_i,
coordsA="data",
coordsB="data",
)
self.cpatch_line = ConnectionPatch(
xyA=self.px_i,
xyB=self.px_f,
coordsA="data",
coordsB="data",
)
self.xyA_f = (
(self.px_f[0] - width / 2 * np.sin(self.angle)),
(self.px_f[1] + width / 2 * np.cos(self.angle)),
)
self.xyB_f = (
(self.px_f[0] + width / 2 * np.sin(self.angle)),
(self.px_f[1] - width / 2 * np.cos(self.angle)),
)
self.cpatch_f = ConnectionPatch(
xyA=self.xyA_f,
xyB=self.xyB_f,
coordsA="data",
coordsB="data",
)
def _plot_feature_to_kernel_lines(kernel_matrix_2d, feature_matrix_2d,
feature_row_at_center,
feature_column_at_center, kernel_axes_object,
feature_axes_object):
"""Plots lines between feature map and kernel.
:param kernel_matrix_2d: See doc for `_plot_kernel`.
:param feature_matrix_2d: Same.
:param feature_row_at_center: Same.
:param feature_column_at_center: Same.
:param kernel_axes_object: Axes on which kernel is plotted (instance of
`matplotlib.axes._subplots.AxesSubplot`).
:param feature_axes_object: Axes on which feature map is plotted (instance
of `matplotlib.axes._subplots.AxesSubplot`).
"""
num_feature_rows = feature_matrix_2d.shape[0]
num_feature_columns = feature_matrix_2d.shape[1]
num_kernel_rows = kernel_matrix_2d.shape[0]
first_feature_row = max([feature_row_at_center - 1.5, -0.5])
last_feature_row = min(
[feature_row_at_center + 1.5, num_feature_rows - 0.5])
last_feature_column = min(
[feature_column_at_center + 1.5, num_feature_columns - 0.5])
center_row_from_bottom = num_feature_rows - feature_row_at_center - 1
first_kernel_row = -0.5 + max([1 - feature_row_at_center, 0])
last_kernel_row = 0.5 + min(
[1 + center_row_from_bottom, num_kernel_rows - 1])
first_kernel_column = -0.5 + max([1 - feature_column_at_center, 0])
this_connection_object = ConnectionPatch(
xyA=(first_kernel_column, first_kernel_row),
xyB=(last_feature_column, first_feature_row),
coordsA='data',
coordsB='data',
axesA=kernel_axes_object,
axesB=feature_axes_object,
color=FEATURE_TO_KERNEL_LINE_COLOUR,
linewidth=FEATURE_TO_KERNEL_LINE_WIDTH,
linestyle='dashed')
kernel_axes_object.add_artist(this_connection_object)
this_connection_object = ConnectionPatch(
xyA=(first_kernel_column, last_kernel_row),
xyB=(last_feature_column, last_feature_row),
coordsA='data',
coordsB='data',
axesA=kernel_axes_object,
axesB=feature_axes_object,
color=FEATURE_TO_KERNEL_LINE_COLOUR,
linewidth=FEATURE_TO_KERNEL_LINE_WIDTH,
linestyle='dashed')
kernel_axes_object.add_artist(this_connection_object)
def step_1():
t,resonance,FID,fig,ax_one,ax_FID,ax_mult,ax_FT=draw_freq()
con_2 = ConnectionPatch(xyA=(0.5,-0.2), xyB=(0.5,1.15), coordsA="axes fraction", coordsB="axes fraction",axesA=ax_FID, axesB=ax_mult,color="red")
con_2.set_arrowstyle("simple",head_length=0.5, head_width=1, tail_width=0.3)
ax_FID.add_patch(con_2)
freqs=np.arange(session["freq_min"]-1,session["freq_max"]+1,session["freq_res"])
freqs=np.append(freqs,session["freq_max"]+1)
nb_freq=freqs.size
trial=np.zeros((nb_freq,session["nb_points"]))
mult=np.zeros((nb_freq,session["nb_points"]))
integ=np.zeros(nb_freq)
posinteg=np.zeros(nb_freq)
neginteg=np.zeros(nb_freq)
for i,freq in enumerate(freqs):
trial[i]=np.cos(2*np.pi*freq*t)
mult[i]=trial[i]*FID
posmult=np.zeros(session["nb_points"])
negmult=np.zeros(session["nb_points"])
for nn,ii in enumerate(mult[i]):
if ii >0:
posmult[nn]=ii
if ii <0:
negmult[nn]=ii
posinteg[i]=np.trapz(posmult,t)
neginteg[i]=np.trapz(negmult,t)
integ[i]=np.trapz(mult[i],t)
if freq<=session["freq_min"]:
n=i
ax_FT.set_ylim(integ.min()-(integ.max()-integ.min())*0.05,integ.max()+(integ.max()-integ.min())*0.4)
ax_FID.plot(t,trial[n],linewidth=1,color="magenta",alpha=0.5)
ax_mult.plot(t,mult[n],linewidth=1,color="black",alpha=0.5)
ax_FID.legend(("FID","Trial Frequency {} Hz".format(round(freqs[n],1))),loc=1)
session["message"]="""The acquired signal will be processed and the computer will try to find all the frequencies from which your FID is composed.<br><br>
To do this, it will multiply your FID by cosine functions with increasing frequencies, covering all our frequencies range from {} to {} Hz.<br>
We will call these cosine functions "Trial Frequencies".<br><br>
The multiplication is done "point by point", as shown in the example below: <br><br>
<img src="/static/img1.png" style="width:95%"/><br><br>
The result of the multiplication of your FID by the trial frequency {} Hz is shown in graph 3.<br>
<input type=submit name="msgOK" value="Next" class="control">""".format(session["freq_min"],session["freq_max"],round(freqs[n],1))
session["FT_step"]=1
return t,resonance,FID,fig,ax_one,ax_FID,ax_mult,ax_FT,freqs,nb_freq,trial,mult,integ,posinteg,neginteg,n
def add_connection(source_ax, source_data_coord, target_ax, target_data_coord):
con = ConnectionPatch(xyA=source_data_coord,
xyB=target_data_coord,
coordsA='data',
coordsB='data',
axesA=source_ax,
axesB=target_ax,
arrowstyle='->',
clip_on=False,
linewidth=3)
con.set_zorder(10)
source_ax.add_artist(con)
def linking_lines(
fig_: Figure,
axes_A: Axes,
axes_B: Axes,
axes_C: Axes,
axes_D: Axes,
color_: str = brown_,
) -> None:
r"""
Draw lines linking features in the main plot to zoom insets.
Args:
fig_:
Figure to act on
axes_A:
Axes of zoom box A
axes_B:
Axes of zoom box B
axes_C:
Axes of zoom box C
axes_D:
Axes of zoom box D
"""
joins: List[ConnectionPatch] = [None] * 3
kwargs = dict(color=color_, linestyle=":")
joins[0] = ConnectionPatch(
xyA=(0.2, 0),
coordsA=axes_D.transData,
xyB=(0.4, 1),
coordsB=axes_A.transData,
**kwargs,
)
joins[1] = ConnectionPatch(
xyA=(1, 0.00),
coordsA=axes_D.transData,
xyB=(0, 0.9),
coordsB=axes_B.transData,
**kwargs,
)
joins[2] = ConnectionPatch(
xyA=(1, 0.60),
coordsA=axes_D.transData,
xyB=(0, 0.8),
coordsB=axes_C.transData,
**kwargs,
)
for join_ in joins:
fig_.add_artist(join_)
def plotting_pairs(self, BLPoints, FUPoints):
'''
Plotting the point pairs next to each other in one plot.
:param BLPoints: Baseline points -- shape (N,2)
:param FUPoints: Follow-up points -- shape (N,2)
Function for plotting pairs of points in two transformed images.
'''
fig = plt.figure(1)
a1 = fig.add_subplot(1, 2, 1)
self.plotting_points(self.bl, BLPoints, number=True)
a2 = fig.add_subplot(1, 2, 2)
self.plotting_points(self.fu, FUPoints, number=True)
for idx in range(BLPoints.shape[0]):
xbl = BLPoints[idx][0]
ybl = BLPoints[idx][1]
xfu = FUPoints[idx][0]
yfu = FUPoints[idx][1]
con = ConnectionPatch(xyA=(xfu, yfu),
xyB=(xbl, ybl),
coordsA="data",
coordsB="data",
axesA=a2,
axesB=a1,
color="red")
a2.add_artist(con)
def vmarker(x0, x1, ax0, ax1, **kwargs):
xy0 = (x0, ax0.get_ylim()[0])
xy1 = (x1, ax1.get_ylim()[1])
ax0.axvline(x0, **kwargs)
ax1.axvline(x1, **kwargs)
con = ConnectionPatch(xy0, xy1, 'data', 'data', ax0, ax1, **kwargs)
ax0.add_artist(con)
def visualize_pair(self, image1, image2):
img1 = mpimg.imread(image1)
img2 = mpimg.imread(image2)
x1 = self.point1.x
y1 = self.point1.y
x2 = self.point2.x
y2 = self.point2.y
fig = plt.figure()
ax1 = fig.add_subplot(1, 2, 1)
ax1.imshow(img1)
ax1.plot(x1, y1, 'r*')
ax2 = fig.add_subplot(1, 2, 2)
ax2.imshow(img2)
ax2.plot(x2, y2, 'r*')
con = ConnectionPatch(xyA=(x2, y2),
xyB=(x1, y1),
coordsA="data",
coordsB="data",
axesA=ax2,
axesB=ax1,
color="red")
ax2.add_artist(con)
fig.savefig('result.png')
def show(self, filename=None, **kwargs):
tsx, tsy = kwargs.get('tsx', 0), kwargs.get('tsy', 0)
fig, ax = plt.subplots(figsize=self.figsize)
ax.axis('equal')
ax.set_xticks([]), ax.set_yticks([])
sns.despine(left=True, bottom=True, right=True, top=True)
ax.set_xlim((min(self.ymin, self.xmin), max(self.ymax, self.xmax)))
ax.set_ylim((min(self.ymin, self.xmin), max(self.ymax, self.xmax)))
for name, v in self._vertices.items():
vx, vy = v.position
c = Circle((vx, vy), v.radius, zorder=20)
c.set_edgecolor(v.rcolor)
c.set_facecolor(v.color)
ax.annotate(name,
xy=(vx, vy),
fontsize=v.tsize,
ha='center',
xytext=(vx + tsx, vy + tsy),
color=v.tcolor,
zorder=21)
ax.add_artist(c)
for e in self._alist[name]:
l = ConnectionPatch(e.a.position, e.b.position, 'data')
ax.add_artist(l)
plt.close()
return fig
def animate(frame):
ax.patches = []
c_B = cum_orbits[0, int(frame)]
for j, i in enumerate(draw_idx):
c_A, c_B = cum_orbits[i, int(frame)], cum_orbits[i + 1, int(frame)]
r = radius[i]
ax.add_patch(
plt.Circle((c_A.real, c_A.imag),
r,
fill=False,
linestyle="dotted",
linewidth=args.circle_width,
color=args.circle_color))
ax.add_patch(
ConnectionPatch((c_A.real, c_A.imag), (c_B.real, c_B.imag),
ax.transData,
arrowstyle="-",
linewidth=args.vector_width,
color=args.vector_color))
x, y = cum_orbits[-1, :int(frame) +
1].real, cum_orbits[-1, :int(frame) + 1].imag
pl.set_data(x, y)
return ax.patches + [
pl,
]
def example():
from matplotlib import pyplot as plt
import numpy as np
# todo: read the example image from the papers source directory
image1 = imread(wrapper.source_directory + '/adam1.png')
image2 = imread(wrapper.source_directory + '/adam2.png')
# todo: call the wrapped function(s) with a valid set of arguments
matches, keys1, keys2 = wrapper.asift(image1, image2)
# todo: display the results
fig = plt.figure(figsize=(10, 5))
ax1 = fig.add_subplot(121)
plt.imshow(image1, cmap='Greys_r')
ax2 = fig.add_subplot(122)
plt.imshow(image2, cmap='Greys_r')
for x1, y1, x2, y2 in matches[::1, :]:
con = ConnectionPatch(xyA=(x2, y2),
xyB=(x1, y1),
coordsA="data",
coordsB="data",
axesA=ax2,
axesB=ax1,
shrinkB=5,
color='r')
ax2.add_artist(con)
plt.draw()
plt.show()
print('done')
def draw_pitch(ax):
# focus on only half of the pitch
#Pitch Outline & Centre Line
Pitch = Rectangle([0,0], width = 120, height = 80, fill = False)
#Left, Right Penalty Area and midline
LeftPenalty = Rectangle([0,22.3], width = 14.6, height = 35.3, fill = False)
RightPenalty = Rectangle([105.4,22.3], width = 14.6, height = 35.3, fill = False)
midline = ConnectionPatch([60,0], [60,80], "data", "data")
#Left, Right 6-yard Box
LeftSixYard = Rectangle([0,32], width = 4.9, height = 16, fill = False)
RightSixYard = Rectangle([115.1,32], width = 4.9, height = 16, fill = False)
#Prepare Circles
centreCircle = plt.Circle((60,40),8.1,color="black", fill = False)
centreSpot = plt.Circle((60,40),0.71,color="black")
#Penalty spots and Arcs around penalty boxes
leftPenSpot = plt.Circle((9.7,40),0.71,color="black")
rightPenSpot = plt.Circle((110.3,40),0.71,color="black")
leftArc = Arc((9.7,40),height=16.2,width=16.2,angle=0,theta1=310,theta2=50,color="black")
rightArc = Arc((110.3,40),height=16.2,width=16.2,angle=0,theta1=130,theta2=230,color="black")
element = [Pitch, LeftPenalty, RightPenalty, midline, LeftSixYard, RightSixYard, centreCircle,
centreSpot, rightPenSpot, leftPenSpot, leftArc, rightArc]
for i in element:
ax.add_patch(i)
def draw(self, img_q, img_t, pt_qt):
import matplotlib
matplotlib.use('Agg')
from matplotlib import pyplot as plt
from matplotlib.patches import ConnectionPatch
fig, (ax_q, ax_t) = plt.subplots(1, 2, figsize=(8, 4))
for pt_q, pt_t in pt_qt:
con = ConnectionPatch(pt_t,
pt_q,
coordsA='data',
coordsB='data',
axesA=ax_t,
axesB=ax_q,
color='g',
linewidth=0.5)
ax_t.add_artist(con)
ax_q.plot(pt_q[0], pt_q[1], 'rx')
ax_t.plot(pt_t[0], pt_t[1], 'rx')
ax_q.imshow(img_q)
ax_t.imshow(img_t)
ax_q.axis('off')
ax_t.axis('off')
plt.subplots_adjust(wspace=0, hspace=0)
plt.show()
def drawSPC(self, currentClass, fig, ogX, ogY): # Draws the CPC by resizing the figure and creating the graph
for currentIndex in range(len(ogX)):
n = len(fig.axes)
for b in range(n):
fig.axes[b].change_geometry(currentClass, len(ogX), b + 1) # Resizing (row, column, position)
ax = fig.add_subplot(currentClass, len(ogX), n + 1)
ax.set_xlabel("X " + str(currentIndex * 2 + 1)) # Styling the graph
ax.set_ylabel("X " + str(currentIndex * 2 + 2))
if (currentIndex == 0):
ax.set_title("SPC for Class: " + str(currentClass))
xMin = min(ogX) - 1
xMax = max(ogX) + 1
yMin = min(ogY) - 1
yMax = max(ogY) + 1
plt.xlim(float(xMin), float(xMax))
plt.ylim(float(yMin), float(yMax))
ax.plot(ogX[currentIndex], ogY[currentIndex], 'r') # Plotting the graph
pax = fig.axes[n]
if (currentIndex > 0):
xy = (
ogX[currentIndex - 1], ogY[currentIndex - 1]) # These are for original values to be graphed in SPC
ab = (ogX[currentIndex], ogY[currentIndex])
con = ConnectionPatch(xyA=ab, xyB=xy, coordsA="data", coordsB="data",
axesA=fig.axes[n], axesB=fig.axes[n - 1], color="purple")
pax.add_artist(con)
pax.plot(ogX[currentIndex], ogY[currentIndex], 'ro', markersize=10) # Plot the SPC points on each graph
else:
pax.plot(ogX[currentIndex], ogY[currentIndex], 'ro', markersize=10) # Plot the SPC points on each graph
def plot_corr(img1, img2, img1_pts, img2_pts, vsplit=False, figsize=(20, 15)):
plot1 = 211
plot2 = 212
if vsplit:
plot1 = 121
plot2 = 122
fig = plt.figure(figsize=figsize)
ax1 = plt.subplot(plot1)
ax1.imshow(img1, cmap='gray', aspect = "auto")
ax1.scatter(img1_pts[:, 0], img1_pts[:, 1], marker='+')
ax1.set_xticklabels([])
ax1.set_yticklabels([])
ax2 = plt.subplot(plot2)
ax2.imshow(img2, cmap='gray', aspect = "auto")
ax2.scatter(img2_pts[:, 0], img2_pts[:, 1], marker='+')
ax2.set_xticklabels([])
ax2.set_yticklabels([])
for i in range(len(img1_pts)):
xy1 = (img1_pts[i,0],img1_pts[i,1])
xy2 = (img2_pts[i,0],img2_pts[i,1])
con = ConnectionPatch(xyA=xy2, xyB=xy1, coordsA="data", coordsB="data",
axesA=ax2, axesB=ax1, color='#53F242')
ax2.add_artist(con)
plt.subplots_adjust(wspace=0, hspace=0)
# plt.tight_layout()
plt.show()
def addArrow(ax,
cx,
xdiff,
cy,
ydiff,
text,
color="k",
ha="left",
va="center",
arrowstyle="-|>",
shrinkA=2,
shrinkB=4,
xlog=False,
ylog=True):
x = cx * xdiff if xlog else cx + xdiff
y = cy * ydiff if ylog else cy + ydiff
con = ConnectionPatch(xyA=(x, y),
xyB=(cx, cy),
coordsA="data",
coordsB="data",
arrowstyle=arrowstyle,
shrinkA=shrinkA,
shrinkB=shrinkB,
mutation_scale=20,
fc="w",
color=color)
# connectionstyle="arc3,rad=0.3"
ax.add_artist(con)
ax.annotate(text, xy=(x, y), ha=ha, va=va,
color=color) #,box=dict(boxstyle="round", fc="w")
def plot_expected_lengths(lengths, batch_sizes, choose_length, markers=[], n_batches=10000):
fig, axarr = plt.subplots(len(batch_sizes), 1, figsize=(14, 20), sharex=True)
expected_lengths = {}
for i, batch_size in enumerate(batch_sizes):
maxs = []
for _ in tqdm(range(n_batches), disable=False):
val = choose_length(np.random.choice(lengths, batch_size))
maxs.append(math.ceil(val))
pd.Series(maxs).plot.hist(bins=50, ax=axarr[i], density=True, color='black', edgecolor='white', alpha=0.1)
expected = np.mean(maxs)
expected_lengths[batch_size] = expected
max_y = axarr[i].get_ylim()[1]
axarr[i].vlines([expected], 0, 1e3, 'limegreen', lw=4)
axarr[i].set_ylim([0, max_y])
axarr[i].set_xlim([0, max(lengths)])
axarr[i].set_ylabel(f'batch_size={batch_size}', rotation=0)
axarr[i].yaxis.set_label_coords(-0.1, 0.45)
axarr[i].set_yticks([])
for marker in markers:
con = ConnectionPatch(xyA=(marker, axarr[0].get_ylim()[1]), xyB=(marker, 0), coordsA='data',
coordsB='data', axesA=axarr[0], axesB=axarr[-1], color='red', lw=4)
axarr[0].add_artist(con)
axarr[0].set_zorder(1)
axarr[0].set_title(f'Expected sequence lengths with various batch sizes (n per batch = {n_batches})')
plt.subplots_adjust(hspace=0)
return expected_lengths
def plot_graph(graph, ax, n, weighted, shrinkA, shrinkB, layout,
polygon_radius, attr, arrowstyle, weight_color, arrow_color):
"""
creates a plot of nodes and edges connected by a ConnectionPatch object.
"""
for node in plot_graph.nodes:
x_i, y_i = plot_graph.points[node]
for adj_node in node.adjacent_nodes:
_from = graph.get_node_id(node)
_to = graph.get_node_id(adj_node)
edge = graph.connections[_from][_to]
A_color = "k" if not arrow_color else edge.color
A_width = edge.linewidth if not edge.linewidth else edge.linewidth * 4
x_prime, y_prime = plot_graph.points[adj_node]
con = ConnectionPatch((x_i, y_i), (x_prime, y_prime),
COORDSA,
COORDSB,
arrowstyle=arrowstyle,
shrinkA=shrinkA,
shrinkB=shrinkB,
mutation_scale=plot_graph.scale,
fc=A_color,
linewidth=A_width,
edgecolor=A_color)
if weighted:
T_color = "k" if not weight_color else A_color
x_mid, y_mid = (x_prime + x_i) * 0.5, (y_prime + y_i) * 0.5
weight = edge.weight
fontsize = 12 / n + 12
ax.text(x_mid,
y_mid,
str(weight),
fontsize=edge.fontsize * fontsize + fontsize,
color=T_color)
ax.add_artist(con)
def draw_matches(img1, keypoints1, img2, keypoints2, plot_title=""):
"""
Draws matches of keypoints between img1 and img2.
:param img1: the first image
:param keypoints1: the first image's keypoints
:param img2: the second image
:param keypoints2: the second image's keypoints
:param plot_title: the title of the plot
"""
figure = plt.figure(figsize=(10, 5))
ax1 = plt.subplot(1, 2, 1)
ax2 = plt.subplot(1, 2, 2)
img1 = cv2.drawKeypoints(img1, keypoints1, None)
img2 = cv2.drawKeypoints(img2, keypoints2, None)
ax1.imshow(img1)
ax2.imshow(img2)
for kp1, kp2 in zip(keypoints1, keypoints2):
con = ConnectionPatch(xyA=kp2.pt, xyB=kp1.pt,
coordsA="data", coordsB="data",
axesA=ax2, axesB=ax1, color=np.random.rand(3, ))
ax2.add_patch(con)
plt.title(plot_title)
plt.show()
figure.savefig("data/results/" + plot_title.replace(" ", "-").replace(".", "-") + '.png', dpi=100,
bbox_inches='tight')
def example():
# todo: read the example image from the papers source directory
image1=imread(wrapper.source_directory+'/examples/gobelet.png')
image2=imread(wrapper.source_directory+'/examples/gobelet2.png')
# todo: call the wrapped function(s) with a valid set of arguments
#keys1=wrapper.
keys1=wrapper.extract_surf(image1)
keys2=wrapper.extract_surf(image2)
matches=wrapper.match_surf(keys1,keys2)
# todo: display the results
fig = plt.figure(figsize=(10,5))
ax1 = fig.add_subplot(121)
plt.imshow(image1)
ax2 = fig.add_subplot(122)
plt.imshow(image2)
for x1,y1,x2,y2 in matches:
con = ConnectionPatch(xyA=(x2, y2), xyB=(x1,y1), coordsA="data", coordsB="data" ,
axesA=ax2, axesB=ax1,
shrinkB=5,color='r')
ax2.add_artist(con)
plt.draw()
plt.show()
print 'done'
def plot_transfer(self, transfer, source_plates, target_plates, axes=None):
"""Plot the plates and add and arrow for the transfer"""
fig, axes, axes_dict = self.plot_plates(source_plates, target_plates,
axes=axes)
source_well = transfer.source_well
source_plate = source_well.plate
source_ax = axes_dict[source_plate.name]
x, y = source_well.coordinates[::-1]
source_coord = x, source_plate.num_rows - y + 1
target_well = transfer.destination_well
target_plate = target_well.plate
target_ax = axes_dict[target_plate.name]
x, y = target_well.coordinates[::-1]
target_coord = x, target_plate.num_rows - y + 1
target_ax.set_zorder(source_ax.get_zorder() - 1)
arrow = ConnectionPatch(
xyA=source_coord, xyB=target_coord,
coordsA="data", coordsB="data",
axesA=source_ax, axesB=target_ax,
shrinkB=5.0, lw=2,
facecolor='black',
arrowstyle="wedge", zorder=1000)
source_ax.add_artist(arrow)
return fig, axes
ax2.set_title('Age of approvers')
ax2.legend(('50-65', 'Over 65', '35-49', 'Under 35'))
ax2.axis('off')
ax2.set_xlim(- 2.5 * width, 2.5 * width)
# use ConnectionPatch to draw lines between the two plots
# get the wedge data
theta1, theta2 = ax1.patches[0].theta1, ax1.patches[0].theta2
center, r = ax1.patches[0].center, ax1.patches[0].r
bar_height = sum([item.get_height() for item in ax2.patches])
# draw top connecting line
x = r * np.cos(np.pi / 180 * theta2) + center[0]
y = np.sin(np.pi / 180 * theta2) + center[1]
con = ConnectionPatch(xyA=(- width / 2, bar_height), xyB=(x, y),
coordsA="data", coordsB="data", axesA=ax2, axesB=ax1)
con.set_color([0, 0, 0])
con.set_linewidth(4)
ax2.add_artist(con)
# draw bottom connecting line
x = r * np.cos(np.pi / 180 * theta1) + center[0]
y = np.sin(np.pi / 180 * theta1) + center[1]
con = ConnectionPatch(xyA=(- width / 2, 0), xyB=(x, y), coordsA="data",
coordsB="data", axesA=ax2, axesB=ax1)
con.set_color([0, 0, 0])
ax2.add_artist(con)
con.set_linewidth(4)
plt.show()