Users’ perspective

In PySDM and PyMPDATA packages, we enforce a consistent structure across Jupyter-notebooks. This includes three standard badges enabling execution of the notebook on different platforms: Google Colab, Binder, and GitHub.

Take-home message

Notebooks are self-contained and ready to run! User can open them directly in Colab, Binder, etc.

Visualisations

In all example-notebooks, visuals are embedded within the file without storing separate output files in the repository (self-contained). This approach supports:

maintainability for developers, and
usability for researchers who want to run and modify the code themselves.

Take-home message

Inline visuals make research notebooks easier to maintain and adapt!

We provide utilities for visualisations via open-atmos-jupyter-utils package:

`show_plot()`

show_plot()

renders inline SVG graphics
adds buttons to save figures as SVG or PDF
provides Google Drive integration on Colab
displays correctly on GitHub

Take-home message

Vector graphics = journal-ready plots.

Minimal example

See this minimal usage demo in the open-atmos-jupyter-utils repository:

This notebook reproduces the Mandelbrot set, first published by Brooks and Matelski (see Wikipedia: Mandelbrot set)

import numpy as np
from matplotlib import pyplot 
from open_atmos_jupyter_utils import show_plot

np.seterr(all='ignore')

def is_stable(c, n_iters):
    z = 0
    for _ in range(n_iters):
        z = z ** 2 + c 
    return abs(z) <= 2

data = np.array([
    re + im * 1j  
    for re in np.arange(-2,   1,   3/64)
    for im in np.arange(-1.5, 1.5, 3/64)
])
data = data[is_stable(data, n_iters=20)]

pyplot.scatter(data.real, data.imag, marker='.')
show_plot()

With CI automation, testing routines and version-controlled environments, we ensure notebooks stay up-to-date, satisfying reproducibility requirements.

Take-home message

Notebooks are powerful for tutorials, but only if kept up-to-date with ongoing development (with CI automation!)

PySDM examples with `show_plot()` usage

based on Fig. 2 from Lowe et al. [2019]
based on Fig. 1 from Shipway and Hill [2012] — Including BREAKUP process to demonstrate physical changes to cloud.
Miyake et al. [1968]
and more…

`show_anim()`

show_anim()

uses matplotlib and imageio
embeds animations as base64 GIFs in .ipynb
adds “Save as GIF” button and Colab support
renders correctly on GitHub

Minimal example

Example usage available in the open-atmos-jupyter-utils repository:

from matplotlib import pyplot
from open_atmos_jupyter_utils import show_anim 

def anim_func(frame):
    pyplot.plot([(1j**((i+frame)/100)).real for i in range(500)])    
    return pyplot.gcf()

show_anim(anim_func, frame_range=range(50))

Users’ perspective

Visualisations

show_plot()

Minimal example

PySDM examples with show_plot() usage

show_anim()

Minimal example

DEMO - Users’ perspective

`show_plot()`

PySDM examples with `show_plot()` usage

`show_anim()`