#################################################################################
## ##
## Copyright C 2018 Juan P. Dominguez-Morales ##
## ##
## This file is part of pyNAVIS. ##
## ##
## pyNAVIS is free software: you can redistribute it and/or modify ##
## it under the terms of the GNU General Public License as published by ##
## the Free Software Foundation, either version 3 of the License, or ##
## (at your option) any later version. ##
## ##
## pyNAVIS is distributed in the hope that it will be useful, ##
## but WITHOUT ANY WARRANTY; without even the implied warranty of ##
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.See the ##
## GNU General Public License for more details. ##
## ##
## You should have received a copy of the GNU General Public License ##
## along with pyNAVIS. If not, see <http://www.gnu.org/licenses/>. ##
## ##
#################################################################################
import math
import random
from bisect import bisect_left, bisect_right
from matplotlib.colors import LinearSegmentedColormap
import matplotlib.pyplot as plt
import numpy as np
import time
from .utils import Utils
[docs]class Plots:
[docs] @staticmethod
def spikegram(spikes_file, settings, dot_size = 0.2, dot_freq = 1, graph_tile = 'Spikegram', start_at_zero = True, verbose = False):
"""
Plots the spikegram (also known as cochleogram or raster plot) of a SpikesFile.
This is, a graph where the X axis means time and the Y axis represents addresses (or cochlea channels), and where every spike is plotted as a dot.
Parameters:
spikes_file (SpikesFile): File to plot.
settings (MainSettings): Configuration parameters for the file to plot.
dot_size (float): Size of the dots used in the spikegram plot.
dot_freq (int): Set the frequency of spikes that will be represented in the spikegram.
graph_tile (string, optional): Text that will appear as title for the graph.
start_at_zero (boolean, optional): If set to True, the X axis will start at 0, instead of starting at the minimum timestamp.
verbose (boolean, optional): Set to True if you want the execution time of the function to be printed.
Returns:
None.
Note:
A value of 10 in dot_size means that for every 10 spikes, only 1 will be plotted. This helps reducing lag when plotting heavy files.
"""
if verbose == True: start_time = time.time()
#REPRESENTATION
plt.style.use('seaborn-whitegrid')
spk_fig = plt.figure()
spk_fig.canvas.set_window_title(graph_tile)
random.seed(0)
if settings.mono_stereo == 0:
plt.scatter(spikes_file.timestamps[0::dot_freq], spikes_file.addresses[0::dot_freq], s=dot_size)
else:
aedat_addr_ts = list(zip(spikes_file.addresses, spikes_file.timestamps))
addr, ts = zip(*[(evt[0], evt[1]) for evt in aedat_addr_ts if evt[0] < settings.num_channels*(settings.on_off_both + 1)])
plt.scatter(ts[0::dot_freq], addr[0::dot_freq], s=dot_size, label='Left cochlea')
addr, ts = zip(*[(evt[0], evt[1]) for evt in aedat_addr_ts if evt[0] >= settings.num_channels*(settings.on_off_both + 1) and evt[0] < settings.num_channels*(settings.on_off_both + 1)*2])
plt.scatter(ts[0::dot_freq], addr[0::dot_freq], s=dot_size, label='Right cochlea')
plt.legend(fancybox=False, ncol=2, loc='upper center', markerscale=2/dot_size, frameon=True)
if verbose == True: print('SPIKEGRAM CALCULATION', time.time() - start_time)
plt.title(graph_tile, fontsize='x-large')
plt.xlabel('Timestamp ($\mu$s)', fontsize='large')
plt.ylabel('Address', fontsize='large')
plt.ylim([0, settings.num_channels*(settings.on_off_both + 1)*(settings.mono_stereo + 1)])
if start_at_zero:
plt.xlim([0, np.max(spikes_file.timestamps)])
plt.tight_layout()
spk_fig.show()
[docs] @staticmethod
def sonogram(spikes_file, settings, return_data = False, graph_tile = 'Sonogram', start_at_zero = True, verbose = False):
"""
Plots the sonogram of a SpikesFile.
This is, a graph where the X axis means time and the Y axis represents addresses (or cochlea channels), and where the spiking activity is shown with color.
Parameters:
spikes_file (SpikesFile): File to plot.
settings (MainSettings): Configuration parameters for the file to plot.
return_data (boolean, optional): When set to True, the sonogram matrix will be returned instead of plotted.
graph_tile (string, optional): Text that will appear as title for the graph.
start_at_zero (boolean, optional): If set to True, the X axis will start at 0, instead of starting at the minimum timestamp.
verbose (boolean, optional): Set to True if you want the execution time of the function to be printed.
Returns:
int[ , ]: Sonogram matrix. Only returned if return_data is set to True.
"""
if verbose == True: start_time = time.time()
if start_at_zero:
total_time = max(spikes_file.timestamps)
last_time = 0
else:
total_time = max(spikes_file.timestamps) - min(spikes_file.timestamps)
last_time = min(spikes_file.timestamps)
sonogram = np.zeros((settings.num_channels*(settings.on_off_both + 1)*(settings.mono_stereo+1), int(math.ceil(total_time/settings.bin_size))))
its = int(math.ceil(total_time/settings.bin_size))
#THIS IS NOT NEEDED IF TS ARE SORTED
aedat_addr_ts = zip(spikes_file.addresses, spikes_file.timestamps)
aedat_addr_ts = sorted(aedat_addr_ts, key=Utils.getKey)
spikes_file = Utils.extract_addr_and_ts(aedat_addr_ts)
for i in range(its):
a = bisect_left(spikes_file.timestamps, last_time)
b = bisect_right(spikes_file.timestamps, last_time + settings.bin_size)
blockAddr = spikes_file.addresses[a:b]
spikes = np.bincount(blockAddr, minlength=settings.num_channels*(settings.on_off_both + 1)*(settings.mono_stereo+1))
last_time += settings.bin_size
sonogram[:, i] = spikes
if verbose == True: print('SONOGRAM CALCULATION', time.time() - start_time)
if return_data == False:
# REPRESENTATION
plt.style.use('default')
sng_fig = plt.figure()
sng_fig.canvas.set_window_title(graph_tile)
plt.imshow(sonogram, aspect="auto", cmap='hot') #, aspect="auto")
plt.gca().invert_yaxis()
plt.xlabel('Bin ('+str(settings.bin_size) + '$\mu$s width)', fontsize='large')
plt.ylabel('Address', fontsize='large')
plt.title(graph_tile, fontsize='x-large')
"""
plt.annotate('Right cochlea | Left cochlea',
xy=(1.00, 0.5), xytext=(5, 0),
xycoords=('axes fraction', 'figure fraction'),
textcoords='offset points',
size=11, ha='center', va='center', rotation=270)
"""
if settings.mono_stereo == 1:
plt.annotate('Right cochlea',
xy=(1.00, 0.75), xytext=(5, 0),
xycoords=('axes fraction', 'axes fraction'),
textcoords='offset points',
size=11, ha='center', va='center', rotation=270)
plt.annotate('Left cochlea',
xy=(1.00, 0.25), xytext=(5, 0),
xycoords=('axes fraction', 'axes fraction'),
textcoords='offset points',
size=11, ha='center', va='center', rotation=270)
colorbar = plt.colorbar()
colorbar.set_label('No. of spikes', rotation=270, fontsize='large', labelpad= 10)
sng_fig.show()
else:
return sonogram
[docs] @staticmethod
def histogram(spikes_file, settings, bar_line = 1, graph_tile = 'Histogram', verbose = False):
"""
Plots the histogram of a SpikesFile.
This is, a graph where addresses (or cochlea channels) are represented in the X axis, and number of spikes in the Y axis.
Parameters:
spikes_file (SpikesFile): File to plot.
settings (MainSettings): Configuration parameters for the file to plot.
bar_line (int, optional): Select wether to plot the histogram as bar plot (0) or as a line graph (1).
graph_tile (string, optional): Text that will appear as title for the graph.
verbose (boolean, optional): Set to True if you want the execution time of the function to be printed.
Returns:
int[]: Histogram array.
"""
start_time = time.time()
spikes_count = np.bincount(spikes_file.addresses, minlength=settings.num_channels * (settings.on_off_both + 1) * (settings.mono_stereo + 1))
if verbose == True: print('HISTOGRAM CALCULATION:', time.time() - start_time)
plt.style.use('seaborn-whitegrid')
hst_fig = plt.figure()
hst_fig.canvas.set_window_title(graph_tile)
plt.title(graph_tile, fontsize='x-large')
plt.xlabel('Address', fontsize='large')
plt.ylabel('No. of spikes', fontsize='large')
if bar_line == 0:
if settings.mono_stereo == 1:
plt.bar(np.arange(settings.num_channels * (settings.on_off_both + 1)), spikes_count[0:settings.num_channels*(settings.on_off_both + 1)])
plt.bar(np.arange(settings.num_channels * (settings.on_off_both + 1)), spikes_count[settings.num_channels*(settings.on_off_both + 1):settings.num_channels*2*(settings.on_off_both + 1)])
else:
plt.bar(np.arange(settings.num_channels * (settings.on_off_both + 1) * (settings.mono_stereo + 1)), spikes_count)
else:
if settings.mono_stereo == 1:
plt.plot(np.arange(settings.num_channels * (settings.on_off_both + 1)), spikes_count[0:settings.num_channels*(settings.on_off_both + 1)], label='Left cochlea')
plt.plot(np.arange(settings.num_channels * (settings.on_off_both + 1)), spikes_count[settings.num_channels*(settings.on_off_both + 1):settings.num_channels*2*(settings.on_off_both + 1)], label='Right cochlea')
plt.legend(loc='best', frameon=True)
else:
plt.plot(np.arange(settings.num_channels * (settings.on_off_both + 1) * (settings.mono_stereo + 1)), spikes_count)
plt.tight_layout()
hst_fig.show()
return spikes_count
[docs] @staticmethod
def average_activity(spikes_file, settings, graph_tile = 'Average activity', verbose=False):
"""
Plots the average activity plot of a SpikesFile.
This is, a graph where time is represented in the X axis, and average number of spikes in the Y axis.
Parameters:
spikes_file (SpikesFile): File to plot.
settings (MainSettings): Configuration parameters for the file to plot.
graph_tile (string, optional): Text that will appear as title for the graph.
verbose (boolean, optional): Set to True if you want the execution time of the function to be printed.
Returns:
None.
"""
aedat_addr_ts = zip(spikes_file.addresses, spikes_file.timestamps)
total_time = int(max(spikes_file.timestamps))
last_ts = 0
average_activity_L = np.zeros(int(math.ceil(total_time/settings.bin_size))+1)
if(settings.mono_stereo == 1):
average_activity_R = np.zeros(int(math.ceil(total_time/settings.bin_size))+1)
#THIS TWO ONLY IF CHECK DETECTS AEDAT NOT IN ORDER
aedat_addr_ts = sorted(aedat_addr_ts, key=Utils.getKey)
spikes_file = Utils.extract_addr_and_ts(aedat_addr_ts)
if verbose == True: start_time = time.time()
if settings.mono_stereo == 1:
for i in range(0, total_time, settings.bin_size):
evtL = 0
evtR = 0
a = bisect_left(spikes_file.timestamps, last_ts)
b = bisect_right(spikes_file.timestamps, last_ts + settings.bin_size)
events_list = spikes_file.addresses[a:b]
for j in events_list:
if j < settings.num_channels*(1 + settings.on_off_both):
evtL = evtL + 1
elif j >= settings.num_channels*(1 + settings.on_off_both) and j < settings.num_channels*(1 + settings.on_off_both)*2:
evtR = evtR + 1
average_activity_L[int(i/settings.bin_size)] = evtL
average_activity_R[int(i/settings.bin_size)] = evtR
last_ts = last_ts + settings.bin_size
elif settings.mono_stereo == 0:
for i in range(0, total_time, settings.bin_size):
evtL = 0
a = bisect_left(spikes_file.timestamps, last_ts)
b = bisect_right(spikes_file.timestamps, last_ts + settings.bin_size)
events_list = spikes_file.addresses[a:b]
for j in events_list:
evtL = evtL + 1
average_activity_L[i//settings.bin_size] = evtL
last_ts = last_ts + settings.bin_size
if verbose == True: print('AVERAGE ACTIVITY CALCULATION', time.time() - start_time)
plt.style.use('seaborn-whitegrid')
avg_fig = plt.figure()
avg_fig.canvas.set_window_title(graph_tile)
plt.title(graph_tile, fontsize='x-large')
plt.xlabel('Bin ('+str(settings.bin_size) + '$\mu$s width)', fontsize='large')
plt.ylabel('No. of spikes', fontsize='large')
plt.plot(np.arange(math.ceil(total_time/settings.bin_size)+1), average_activity_L, label='Left cochlea')
if(settings.mono_stereo == 1):
plt.plot(np.arange(math.ceil(total_time/settings.bin_size)+1), average_activity_R, label='Right cochlea')
plt.legend(loc='best', ncol=2, frameon=True)
plt.tight_layout()
avg_fig.show()
[docs] @staticmethod
def difference_between_LR(spikes_file, settings, return_data = False, graph_tile = 'Diff. between L and R cochlea', verbose = False):
"""
Plots a plot showing the differente between the left and the right activity of a SpikesFile.
Parameters:
spikes_file (SpikesFile): File to plot.
settings (MainSettings): Configuration parameters for the file to plot.
return_data (boolean, optional): When set to True, the sonogram matrix will be returned instead of plotted.
graph_tile (string, optional): Text that will appear as title for the graph.
verbose (boolean, optional): Set to True if you want the execution time of the function to be printed.
Returns:
int[ , ]: Disparity matrix. Only returned if return_data is set to True.
Raises:
SettingsError: if settings.mono_stereo == 0
Note:
This function can only be called if the mono_stereo parameter in settings is set to 1.
"""
if settings.mono_stereo == 1:
total_time = max(spikes_file.timestamps) - min(spikes_file.timestamps)
diff = np.zeros((settings.num_channels*(settings.on_off_both + 1), int(math.ceil(total_time/settings.bin_size))))
last_time = min(spikes_file.timestamps)
its = int(math.ceil(total_time/settings.bin_size))
#THIS IS NOT NEEDED IF TS ARE SORTED
aedat_addr_ts = list(zip(spikes_file.addresses, spikes_file.timestamps))
aedat_addr_ts = sorted(aedat_addr_ts, key=Utils.getKey)
spikes_file = Utils.extract_addr_and_ts(aedat_addr_ts)
if verbose == True: start_time = time.time()
for i in range(its):
a = bisect_left(spikes_file.timestamps, last_time)
b = bisect_right(spikes_file.timestamps, last_time + settings.bin_size)
blockAddr = spikes_file.addresses[a:b]
spikes = np.bincount(blockAddr, minlength=settings.num_channels*(settings.on_off_both + 1)*(settings.mono_stereo+1))
last_time += settings.bin_size
diff[:, i] = [x1 - x2 for (x1, x2) in list(zip(spikes[0:settings.num_channels*(settings.on_off_both + 1)], spikes[settings.num_channels*(settings.on_off_both + 1):settings.num_channels*(settings.on_off_both + 1)*2]))]
if verbose == True: print('DIFF CALCULATION', time.time() - start_time)
if max(abs(np.min(diff)), np.max(diff)) != 0: diff = diff*100/max(abs(np.min(diff)), np.max(diff))
if return_data == False:
# REPRESENTATION
plt.style.use('default')
sng_fig = plt.figure()
sng_fig.canvas.set_window_title(graph_tile)
#cmap = 'RdBu'
colors = [(1, 0.49803921568627450980392156862745, 0.05490196078431372549019607843137), (1, 1, 1), (0.12156862745098039215686274509804, 0.46666666666666666666666666666667, 0.70588235294117647058823529411765)] # R -> G -> B
n_bins = [3, 6, 10, 100] # Discretizes the interpolation into bins
cmap_name = 'my_list'
cm = LinearSegmentedColormap.from_list(cmap_name, colors, N=100)
plt.imshow(diff, vmin=-100, vmax=100, aspect="auto", cmap=cm) #, aspect="auto")
plt.gca().invert_yaxis()
plt.xlabel('Bin ('+str(settings.bin_size) + '$\mu$s width)', fontsize='large')
plt.ylabel('Address', fontsize='large')
plt.title(graph_tile, fontsize='x-large')
colorbar = plt.colorbar(ticks=[100, 50, 0, -50, -100], orientation='horizontal')
colorbar.set_label('Cochlea predominance', rotation=0, fontsize='large', labelpad= 10)
colorbar.ax.set_xticklabels(['100% L\nCochlea', '50%', '0%\nL==R', '50%', '100% R\nCochlea'])
colorbar.ax.invert_xaxis()
sng_fig.show()
else:
return diff
else:
#print("This functionality is only available for stereo AEDAT files.")
print("[Plots.difference_between_LR] > SettingsError: This functionality is only available for stereo files.")