def on_button_clicked(b):
# add new point to the Montecarlo experiment
monte.addNewPoint(dropdown_newpoints.value)
# Update the plot
with plot_output:
clear_output(wait=True)
fig = plotTheMonteCarloExperiment(monte.points)
display( fig )
close_fig( fig ) #close figs to liberate memory
# Update the widget printing the value of pi
html_output.value = html_message_template.format(pi_value = str(monte) )
def plot_collections(shifted_data: tuple, thresh: ThreshType, rate: int,
location: str):
normal, delayed = shifted_data
low, high = thresh
fig_format = 'pdf' # figure_settings['format']
linewidth = 0.01 if fig_format in ('svg', 'pdf') else 0.5
name = f'time delay {low}-{high}Hz'
f_name = name.replace(' ', '_')
file_name = os_path.join(location, results_directory['general'],
f'{f_name}.{fig_format}')
nrows = int(round(delayed.columns.size / 8))
delay_size = int32(round(rate / (((low + high) / 2) * 4)))
sns.set(style='white')
default_cols = 8
ncols = default_cols if delayed.columns.size > default_cols else delayed.columns.size
fig, _axes = subplots(nrows=nrows if ncols *
nrows == delayed.columns.size else nrows + 1,
ncols=ncols,
figsize=[16, 20],
subplot_kw={
'aspect': 'equal',
'adjustable': 'box'
})
fig.suptitle(f'{name.title()}\nDelay = {delay_size}', fontsize=12)
axes = _axes.ravel()
for ind, col in enumerate(delayed.columns):
ax = axes[ind]
ax.plot(normal[col], delayed[col], lw=linewidth)
ax.set_xticklabels([])
ax.set_yticklabels([])
ax.set_title('Channel {}'.format(col))
subplots_adjust(top=0.95, bottom=.01, left=.01, right=.99)
fig.savefig(file_name,
format=fig_format,
bbox_inches='tight',
dpi=figure_settings['resolution'])
close_fig(fig)
return True
def plot_heatmap(matrix: ArrayOrDataFrame, location: str, thresh: ThreshType,
kind: str):
low, high = thresh
fig_format = figure_settings['format']
name = f'{kind} heatmap {low}-{high}Hz'
f_name = name.replace(' ', '_')
file_name = os_path.join(location, kind.replace(' ', '_'),
f'{f_name}.{fig_format}')
sns.set(style='ticks')
fig, ax = subplots()
# col, ind = matrix.columns, matrix.index
# matrix_arr = fill_diagonal(matrix.as_matrix(), NaN)
# matrix = DataFrame(matrix_arr, columns=col, index=ind)
sns.heatmap(matrix,
square=matrix.shape[0] == matrix.shape[1],
ax=ax,
cmap=colormap,
xticklabels=10,
yticklabels=10,
robust=True,
cbar_kws={'shrink': .5})
ax.set_title(name.title().replace('_', ' '), fontsize=12)
subplots_adjust(top=0.98, bottom=.02, left=.06, right=.99)
fig.savefig(file_name,
format=fig_format,
bbox_inches='tight',
dpi=figure_settings['resolution'])
close_fig(fig)
return True
def returnPebblessWidget():
# Set an instance of the MontecarloApprox class
monte = MontecarloApprox()
# Set an output widget to display the plot, an HTML widget
# to display the value of pi and buttons widget to update everything
plot_output = ipw.Output()
html_output = ipw.HTML()
update_button = ipw.Button(description="Add points!")
# A dropdown widget to choose how many new points are added when
# update_button is clicked
dropdown_newpoints = ipw.Dropdown(
options=[1, 10, 100],
value=1,
description='New Points:',
disabled=False,
)
# Add Initial plot to the plot_output
with plot_output:
fig = plotTheMonteCarloExperiment( [] )
display( fig )
close_fig( fig ) #close figs to liberate memory
# define initial value for the html_output
html_message_template ="""<div class="jumbotron">
<h3>
π ≈ {pi_value}
</h3>
</div>
"""
html_output.value = html_message_template.format(pi_value = str(monte) )
# Add a call back to the button that will update the Montecarlo Approx with a new point
# and update the scatter plot.
# Using a decorator
@update_button.on_click
def on_button_clicked(b):
# add new point to the Montecarlo experiment
monte.addNewPoint(dropdown_newpoints.value)
# Update the plot
with plot_output:
clear_output(wait=True)
fig = plotTheMonteCarloExperiment(monte.points)
display( fig )
close_fig( fig ) #close figs to liberate memory
# Update the widget printing the value of pi
html_output.value = html_message_template.format(pi_value = str(monte) )
add_points_widget = ipw.VBox(
children = [
dropdown_newpoints,
update_button
],
layout = ipw.Layout(align_items = 'flex-end')
)
return ipw.HBox(
children= [
plot_output,
ipw.VBox(
children = [
add_points_widget,
html_output
]
)
],
layout = ipw.Layout(justify_content = 'center')
)
def plot_windowed_comparison(dt: DataFrame, thresh: ThreshType, rate: int,
location: str, columns, corr_lag):
low, high = thresh
primary_title = f'Moving Windows - Filter threshold {low}-{high}Hz'
# matrix, sum_of_medians = results
moving_mean = DataFrame(dt, columns=columns)
# Setting up the figure -------------------------------------------------------------
sns.set(style="ticks")
fig = figure(figsize=[16, 16])
fig.suptitle(primary_title, fontsize=14)
ax1 = subplot2grid((2, 5), (0, 0), colspan=4, rowspan=1)
# Heatmap ---------------------------------------------------------------------------
sns.heatmap(moving_mean.T,
cmap=colormap,
ax=ax1,
xticklabels=rate * 2,
yticklabels=10,
robust=True,
cbar_kws={
'orientation': 'horizontal',
'shrink': 0.20,
'pad': 0.07
})
sns.despine(ax=ax1, right=True, top=True)
ax1.set_title(
f'Auto-correlation (lag={corr_lag}) of running window (length={SAMPLING_FREQ})',
fontsize=12)
ax1.set_yticklabels(ax1.get_yticklabels(), rotation=60)
# Horizontal sum of the heatmap -----------------------------------------------------
ax2 = subplot2grid((2, 5), (0, 4), colspan=1, rowspan=1)
sns.heatmap(asarray([moving_mean.sum(axis=0)] * 2).T,
cmap=colormap,
ax=ax2,
yticklabels=10,
xticklabels=False,
robust=True,
cbar_kws={
'orientation': 'horizontal',
'pad': 0.07
})
sns.despine(ax=ax2, right=True, top=True)
ax2.set_title('Sum of medians\nper channel', fontsize=12)
ax2.set_yticklabels([])
# Vertical sum of the heatmap -------------------------------------------------------
ax3 = subplot2grid((4, 1), (2, 0), colspan=1, rowspan=1)
time_delta = linspace(0,
max(moving_mean.shape) / rate,
max(moving_mean.shape))
ax3.plot(time_delta,
moving_mean.sum(axis=1).tolist(),
'--',
lw=.75,
label='Normal',
alpha=.5)
ax3_se = ax3.twinx()
ax3_se.plot(time_delta,
abs(moving_mean).sum(axis=1).tolist(),
lw=1,
label='Modulus')
ax3.legend(loc='upper left')
ax3_se.legend(loc='upper right')
ax3.set_title('Sum of medians per window (second)'.format(*thresh),
fontsize=12)
ax3.set_xlim(time_delta.min(), time_delta.max())
# Saving the figure -----------------------------------------------------------------
path = os_path.join(location, results_directory['windowed'],
primary_title + '.' + figure_settings['format'])
fig.savefig(path,
format=figure_settings['format'],
bbox_inches='tight',
dpi=figure_settings['resolution'])
close_fig(fig)
return True
def plotter(data: DataFrame,
location: str,
freq: int,
thresh: ThreshType,
save=True):
# fig_format = 'pdf' # figure_settings['format']
extension = 'pdf' # figure_settings['format']
amplify, linewidth = .5, 0.01 if extension in ('svg', 'pdf') else 0.5
if not save:
linewidth = .6
low, high = thresh
title = f'Filtered Data {low}-{high}Hz'
index_size = data.index.size
index = linspace(0, index_size / freq, index_size, dtype=float32)
index_max_squared = index.max() * 2
sns.set(style='ticks')
fig_width = 60 if index_max_squared > 60 else index_max_squared
fig_height = data.columns.size / 4
fig_kws = dict(figsize=(fig_width, fig_height))
fig, ax = subplots(fig_kw=fig_kws)
data_min, data_max = data.quantile(0.02).mean(), data.quantile(0.98).mean()
dr = (data_max - data_min) * (1 / amplify) # Amplify
func = partial(_stack_channel,
index=index,
index_size=index_size,
data=data)
# [::-1] is to display the matrix in the actual order.
segs = map(func, data.columns[::-1])
offsets = zeros((data.columns.size, 2), dtype=float32)
offsets[:, 1] = arange(-1, data.columns.size - 1) * dr
lines = LineCollection(segs,
offsets=offsets,
transOffset=None,
linewidth=linewidth,
cmap=_edged_colormap())
lines.set_array(arange(0, data.columns.size))
ax.add_collection(lines)
# Display settings:
# ----------------------------------------------------------
ax.set_title(title, fontsize=14)
ax.set_xlim([index.min(), index.max()])
ax.set_ylim([data_min - dr * 2, (data.columns.size - 1) * dr + data_max])
ax.set_yticks(offsets[:, 1])
# [::-1] is to display the matrix in the actual order.
ax.set_yticklabels(data.columns[::-1], fontsize=12)
ax.set_xlabel('Time [sec]')
ax.set_ylabel('Channels [$\mu V$]')
subplots_adjust(top=0.98, bottom=.02, left=.08, right=.99)
ax.set_xticks(arange(0, index.max() + 1))
ax.xaxis.grid(b=True,
which='major',
color='lightgray',
linestyle=':',
linewidth=1,
alpha=.3,
antialiased=True)
subplots_adjust(top=0.98, bottom=.02, left=.08, right=.99)
sns.despine(right=True, top=True, bottom=True, left=True)
location = os_path.join(location, results_directory['general'],
title.replace(' ', '_').lower() + '.' + extension)
if save:
fig.savefig(location,
format=extension,
bbox_inches='tight',
dpi=figure_settings['resolution'])
close_fig(fig)
return True
return fig, ax