def load_cluster(self, cluster):
'''
Load necessary information from cluster object.
'''
self.cluster_data = PlotSingleCluster(cluster)
self.name = cluster.identity()
class PlotSingle(PlotMPL):
'''
Ploting with single cluster data.
'''
def load_cluster(self, cluster):
'''
Load necessary information from cluster object.
'''
self.cluster_data = PlotSingleCluster(cluster)
self.name = cluster.identity()
def plot_original(self, fig = None, savefig = True, display = False):
'''
Plot the original trace.
'''
if fig is None:
fig = plt.figure()
# Build X axis
time, amp = self.cluster_data.compute_original()
ax1 = fig.add_subplot(111)
ax1.plot(time, amp, color = 'black', lw=1,
label = 'Original trace')
ax1.set_xlim([time[0], time[-1]])
ax1.set_xlabel('Time (ms)')
ax1.set_ylabel('Amplitude (pA)')
ax1.set_title('Original trace')
self._plot_dict['original'] = fig
if savefig:
fig.savefig(os.path.join(self.filepath,self.name+'_Original.png'),dpi=300)
if display:
plt.show()
def plot_popen_on_original(self, savefig = True):
'''
Plot the Popen of different modes on the original trace.
'''
if 'original' in self._plot_dict:
self.plot_original()
fig = self._plot_dict['original']
if len(fig.get_axes()) == 1:
ax2 = fig.get_axes()[0].twinx()
else:
raise IndexError('More than one axis in the figure')
time, popen = self.cluster_data.compute_popen()
ax2.plot(time, popen, color = 'blue', label = 'Popen')
ax2.set_ylabel('Popen')
ax2.yaxis.label.set_color('blue')
ax2.set_ylim([0, 1])
ax2.set_xlim([time[0], time[-1]])
for tl in ax2.get_yticklabels():
tl.set_color('blue')
if savefig:
fig.savefig(os.path.join(self.filepath,self.name+'Popen.png'),dpi=300)
def plot_multitrace(self, func_list = ('compute_original', 'compute_popen'),
tracelength = 5e3, trace_number = 10):
'''
Get the data from the func list and plot the data.
'''
x_list = []
y_list = []
for func in func_list:
x,y = getattr(self.cluster_data, func)()
x_list.append(x)
y_list.append(y)
sep_x_list = []
sep_y_list = []
for x,y in zip(x_list, y_list):
sep_x, sep_y = separate_multiply_line(x,y,tracelength = tracelength)
sep_x_list.append(sep_x)
sep_y_list.append(sep_y)
page_number = int(np.ceil(len(sep_x)/trace_number))
fig_list = []
for i in range(page_number):
fig = plt.figure()
fig.set_size_inches(22,17)
for j in range(min([trace_number, len(sep_x_list[0]) - i*trace_number])):
ax1 = fig.add_subplot(trace_number, 1, j+1)
# plot original trace
x = sep_x_list[0][i*trace_number+j]
y = sep_y_list[0][i*trace_number+j]
ax1.plot(x, y, color = 'black', lw=1)
ax1.set_xlim([x[0], x[0] + tracelength])
ax1.set_ylim([-1, 9])
# plot Popen
ax2 = ax1.twinx()
x = sep_x_list[1][i*trace_number+j]
y = sep_y_list[1][i*trace_number+j]
ax2.plot(x, y, color = 'blue')
ax2.set_xlim([x[0], x[0] + tracelength])
ax2.set_ylim([0, 1])
fig.suptitle(self.name+'/'+str(i))
fig.savefig(os.path.join(self.filepath,self.name+'Joint'+str(i)+'.png'))
fig_list.append(fig)
def plot_open_close(self, fig = None, savefig = True):
'''
Plot the open period versus shut periods.
'''
stretch_list = self.cluster_data.compute_open_close()
mode_num = len(stretch_list)
if fig is None:
fig = plt.figure()
ax = fig.add_subplot(111)
cmap = np.linspace(0,1,mode_num)
for index, stretch in enumerate(stretch_list):
ax.scatter(stretch['open_period'], stretch['shut_period'],
facecolors='none',
edgecolors=plt.cm.spectral(cmap[index]),
s=1, label = str(index + 1))
#ax.scatter(stretch['mean_open'], stretch['mean_shut'],
# color=plt.cm.spectral(cmap[index]),
# s=50, label = str(index + 1))
ax.legend()
ax.set_xscale('log')
ax.set_yscale('log')
ax.set_ylim([0.3, np.exp(7)])
ax.set_xlim([0.3, np.exp(5)])
ax.set_xlabel('Open period (ms in log scale)')
ax.set_ylabel('Shut period (ms in log scale)')
ax.set_title('Open/Shut')
if savefig:
fig.savefig(os.path.join(self.filepath,self.name+'Open_Shut.png'),dpi=150)
def plot_cost_difference(self, fig = None, savefig = True):
'''
Plot the cost function, and the difference between the cost and
normalised cost function, which are all normalised by dividing by the
cluster length.
'''
if fig is None:
fig = plt.figure()
# get the data
dataset = self.cluster_data.compute_cost_diff()
mode_num = dataset['mode_num']
cost = dataset['cost']
mean_cost = dataset['mean_cost']
difference = dataset['difference']
mean_difference = dataset['mean_difference']
x = range(1, len(cost)+1)
# Plot cost and mean cost function
ax1 = fig.add_subplot(211)
ax1.plot(x, cost, color = 'blue', label = 'minimum cost')
ax1.plot(x, mean_cost, color = 'red', label = 'average cost')
ax1.plot(mode_num, cost[mode_num - 1], 'o')
ax1.plot(mode_num, mean_cost[mode_num - 1], 'o')
ax1.set_xlabel('Mode number')
ax1.set_ylabel('Cost')
ax1.legend()
# ax1.set_ylim(bottom=0)
# Plot the difference between two cost
x = range(2, len(cost)+1)
ax2 = fig.add_subplot(212)
ax2.plot(x, difference, 'o')
ax2.plot(x, mean_difference)
ax2.set_xlabel('Mode number')
ax2.set_ylabel('effectiveness')
ax1.set_title('Optimum number of stretches')
if savefig:
fig.savefig(os.path.join(self.filepath,self.name+'Cost_Difference.png'),dpi=300)