Custom Parameterizations¶

Adding custom parameterizations to snowmicropyn¶

If you have derived or found a different way to deduce the density and/or specific surface area from an SMP signal (for example for a different version of the device or altogether different snow conditions) than what is included in this package (Proksch, Calonne/Richter, King, …) it is easy to add it requiring only minimal programming skills.

We hope to offer an interface which will let you focus on the Physics.

Implementation¶

Copy a fitting template.

Duplicate parameterizations/proksch2015.py and rename it as a starting point. Give your class a fitting name, here Custom2022.
Choose your attributes.

In the class’ __init__() function you will see a couple of variables being initialized, like a name and shortcut (here “C2022”) for your algorithm. Set these accordingly.
Define your density and SSA functions.

Your class must define the density() function and may define the ssa() function. These will be passed the lower level SMP properties (e. g. median of force, element size, and in case of SSA the density) by the calling chain to process to your liking. You can freely program your functions as long as they return a single float value.
Register your parameterization in the GUI.

The last line like
```
derivatives.parameterizations.register(Custom2022())
```
registers your parameterization in the GUI and for a unified API by creating an instance of your class. The appropriate menus, save- and export functionality will automatically be added to the GUI. Furthermore, you can now access your custom algorithm like all the others:
```
import snowmicropyn as smp
pro = smp.Profile.load('../examples/profiles/S37M0876.pnt')
c2022 = smp.params['C2022'].calc(pro.samples)
```
Document your work

Documentation is built from source code, so adhering to commenting standards (or sticking to the template) should ensure easy integration. To insert your module, add it in docs/api_reference.rst.

Colors¶

A color will be auto-chosen for your parameterization that is slightly different to the existing ones. If you wish to override it you can do so after registering the class:

derivatives.parameterizations.register('Custom2022())
rgb = lambda rr, gg, bb: '#%02x%02x%02x' % (rr, gg, bb)
derivatives.parameterizations['C2022']._density_color = rgb(255, 0, 0) # RGB color
derivatives.parameterizations['C2022']._ssa_color = 'C1' # matplotlib enumerated color

Publishing your Efforts¶

If your work is peer reviewed we are very happy to receive a pull request on GitHub.