Structuring our Plasma factory

1 minute read


We’re mostly done with structuring our Plasma factory. The PR (#459) just got merged yesterday! I’d have to say we weren’t expecting this to work out this quickly.

The reference to SunPy’s Map source made it a tad easier to understand how things are supposed to work for our PlasmaPy’s Plasma. :D

Also, SunPy’s BasicRegistrationFactory was so basic that we could directly import it in PlasmaPy without having to work out on one on our own (wow, that’s a lot of “o”s!). But really, I wonder if it makes more sense that they should just publish a separate PyPi package for just this registration factory. =)

If we didn’t had any such reference, I am certain that our factory implementation would have consumed a lot more time than we actually took.

Anyway, there are still plenty of things that probably would need work in near future, like overriding methods under BasicRegistraionFactory being inherited in PlasmaFactory to make it more specific as we get to know our subclass needs, and defining generic plasma methods which are common in most plasmas (electronTemperature, ionTemperature, etc.) under GenericPlasma class. We’ll get to learn more about these as we work on subclasses for our Plasma factory as well.

Currently, We’ve been discussing about what subclasses would be nice to have (some bits can be found in #458) and what we really want them to do for us. And so, this is our next plan, to create variety of subclasses that deal with different plasma datasets.

Some of the openly available datasets we’ve located are Dense plasma database, Johns Hopkins Turbulence Database and some example datasets using the OpenPMD stanadard. By the way, if you’re working on a dataset, I’d recommend you to use the conventions under OpenPMD standard. It has the potential to make things easier for other people to read and automate!

Lately, I’ve been working with h5py python package which can parse the HDF5 binary data format. I’m very excited since this is going to be my first time working so closely with such large scientific datasets!

That will be all for now. See ya later.