Post-processing

This notebook demonstrates how to produce plots to represent the information extracted from the morphodynamics behavior. All plots are returned as Matplotlib figures that can be customized and exported easily.

#%load_ext autoreload
#%autoreload 2

import matplotlib
%matplotlib widget

import numpy as np
import matplotlib.pyplot as plt
import ipywidgets as ipw
from IPython.display import display

import morphodynamics.plots
from morphodynamics.plots.ui_wimage import Wimage
from morphodynamics.plots import animate_plots, show_plots
from morphodynamics.utils import load_alldata

/Users/gw18g940/miniconda3/envs/morphodynamics/lib/python3.9/site-packages/dask_jobqueue/core.py:20: FutureWarning: tmpfile is deprecated and will be removed in a future release. Please use dask.utils.tmpfile instead.
  from distributed.utils import tmpfile

Define a path to a folder containin the Parameters, Results etc. and use the utils.load_alldata to import the data:

default_path = '../synthetic/data/Results_step/'
param, res, data = load_alldata(default_path, load_results=True)

Display the windows

wimage = Wimage(param, data, res)
wimage.interface

Curvature, length etc.

Create a plot containing geometrical properties of the contour over time:

fig_curv, ax_curv = show_plots.show_geometry_props(data, res, size=(7,3), titles=['Length [um]', 'Area', '$p^2/4*\pi*a^2$'])

As the previous function returned the corresponding figure and axis, we can adjust now our plt. For example we can override the title of the first figure:

ax_curv[0].set_title('My new title')
fig_curv

We can also plot single properties:

fig, ax = show_plots.show_geometry(data, res, prop='circularity', size=(4,2))

Edge overview

fig_edge_overview, ax_edge_overview = show_plots.show_edge_overview(
    param, data, res, lw=0.6, size=(7, 5), cmap_image='Blues',
    cmap_contour='jet', show_colorbar=False)

As an example, to save a figure, you can just use the standard Matplotlib commands:

fig_edge_overview.savefig('edge_overview.pdf')

Edge vectorial

For the edge plots, we can either create a static plot or an animation. To create the animation, we first create a figure and then re-use it. This means that any customization made on the static figure (as here the xlabel) are conserved in the animation:

out = ipw.Output()
with plt.ioff():
    fig, ax = plt.subplots(figsize=(5,5))
    fig_edge_vect, ax_edge_vect = show_plots.show_edge_vectorial_aux(
        param, data, res, k=1, curvature=False, fig_ax=(fig, ax))
    ax.set_xlabel('my test');

int_box = animate_plots.interact_edge_vect(data, param, res, fig, ax)
ipw.VBox([fig.canvas, int_box])

Edge rasterized

The same is true for the visualization of the rasteried contour plot:

from morphodynamics.splineutils import edge_colored_by_features                                            

with plt.ioff():
    fig, ax = show_plots.show_edge_raster_coloured_by_feature(
        data, res, 10, feature='curvature', N=None, width=3, normalize=True, cmap_name='jet')

t_slider = animate_plots.interact_edge_raster_coloured_by_feature(data, param, res, 100, feature='displacement_cumul',
                                                        fig=fig, ax=ax, width=3,normalize=True, cmap_name='jet', alpha=0.8)
ipw.VBox([fig.canvas, t_slider])

The colouring of the edge can represent several varaibles: displacement, cumulative displacement or curvature. These options can be set via the feature option:

im_disp,_ = edge_colored_by_features(data, res, t=0, feature='displacement', N=100)
im_disp_cumul,_ = edge_colored_by_features(data, res, t=0, feature='displacement_cumul', N=100)
im_curv,_ = edge_colored_by_features(data, res, t=20, feature='curvature', N=100)

plt.figure()
plt.imshow(im_curv);

fig, ax = show_plots.show_edge_raster_coloured_by_feature(data, res,20, 'displacement_cumul', width=4, cmap_name='jet'
                                               )

Curvature

fig, ax = show_plots.show_curvature(data, res, cmap_name='jet', show_colorbar=False)

Displacement

fig, ax = show_plots.show_displacement(res, size=(4, 3), cmap_name='jet');

fig, ax = show_plots.show_cumdisplacement(res, size=(4, 3), show_colorbar=False);

Signals

Signal list:

for ind,s in enumerate(data.signal_name):
    print(f'index {ind}: {s}')

index 0: synth_ch2.h5
index 1: synth_ch3.h5

fig, ax = show_plots.show_signals_aux(
    data, res, signal_index=0, layer_index=0, mode='Mean', fig_ax=None, size=(5, 3),
    title='Signal', xlabel='Frame', cmap_name='jet', show_colorbar=True
)

Here again one can easily customize the plot.

ax.set_title("Signal: Channel 1")

Text(0.5, 1.0, 'Signal: Channel 1')