What would you like to see in an open-source plasma physics Python package? about plasmapy HOT 13 CLOSED

from plasmapy.

Should we create a TODO.md file to list that ?
Some points I listed :

• characteristic frequency/distance
• Easy transformation of units eV <-> K, Pa <-> Torr, T <-> Gauss, an so on
• Dispersion relation of common waves
• the legend - wait for it - dary Maxwellian function and its integrations
• A simple particle pusher in given fields
• A Child Languir sub-package

from plasmapy.

from plasmapy.

Just a quick update on a few of the points raised, keep up the good discussion: would everyone be okay with using github issues instead of a todo file? Everyone can open them, they're easier to track than a single file... it just seems better than a static file in most respects, as long as we keep using them dilligently.

We already kind of have a particle pusher in #54, I'll be finishing that (docs etc) today or tomorrow. The summer heat is making it difficult to work on this :(

from plasmapy.

@StanczakDominik I'm not used to github projects, so I'll refer to your opinion. However, if I like Github issues to deal with each issue, I think it is quite scattered. I like Todos because it sums up the goal of the project.

from plasmapy.

from plasmapy.

I agree we should stick to GitHub issues for tracking new features, it keeps bundled in with the codebase itself. We can group issues together into self-contained projects (e.g. v0.1) and if we don't already we should have a 'Feature request' tag for easily searching issues that relate to new things we want to implement.

We also might want different sub-packages for data analysis, simulation, etc.
We do :) (or at least we intend to).

Easy transformation of units eV <-> K, Pa <-> Torr, T <-> Gauss, an so on

astropy.units should make that fairly straightforward.

Dispersion relation of common waves

See #11.

A simple particle pusher in given fields

See #40, #54.

@antoinelpp, the other suggestions are good :) Do you want to go ahead and make some issues for them?

@bla1089, to clarify, you're saying we need data reduction and curve fitting? Could you expand a little on what pandas' groupby does?

from plasmapy.

@SolarDrew I'm waiting for the unit discussion to finish and the PR #54 to be merged. But I'll make some issues !

Easy transformation of units eV <-> K, Pa <-> Torr, T <-> Gauss, an so on

astropy.units should make that fairly straightforward.

Gauss, Torr and eV are not common units. I don't think there already are functions do convert them directly.

from plasmapy.

I'm waiting for the unit discussion to finish and the PR #54 to be merged. But I'll make some issues !

Cool, no problem :)

Gauss, Torr and eV are not common units. I don't think there already are functions do convert them directly.

True. To clarify: astropy.units allows you to define units and equivalencies between units that aren't technically converible (e.g. nm <--> Hz for wavelength/frequency). What I meant was that it should (in theory) be reasonably simple to use these systems to implement the conversions you want, rather than that they would just work out of the box (although Gauss actually is supported in astropy already and does just work).

from plasmapy.

from plasmapy.

For the record I'm not arguing that pandas and xarray aren't good tools. I guess what I should have asked was "what does groupby do that we need in PlasmaPy?"

This issue asks about what tasks people want to accomplish - I think we should decide first what physics people want to do, and then talk about what tools we want to do it with.

from plasmapy.

from plasmapy.

My focus is mostly on experiments in high energy density plasmas (though I do some theory), and most of my tasks involve processing raw data into some usable format (e.g. applying calibrations) and then doing some plasma physics analysis with that processed data (e.g. fit simulations, take line ratios to get temperature, run processed data through an inverse model for the diagnostic to get plasma parameters). Here we can add functions like the Saha-Boltzmann equation; Bohm-Gross dispersion equation; a calibration function that given a measured emission spectrum, a whitelight calibration, and a dispersion calibration, then does the correct interpolations and applies those calibrations to the raw measurement and correctly propagates the uncertainties. (On this note, I am trying to make an error propagation class in python. This would handle asymmetric error propagation, unlike other python projects I've seen.)

Another side to experimental plasma physics is exploratory data analysis, mainly plotting and looking for patterns in data when there is no readily applicable theory. On this front it would be handy to have some ready-made plotting utilities and some more advanced data analysis techniques that aren't already found in other packages (e.g. covariance mapping)

And yet another side of this is diagnostic design, where you want to model the geometry of your diagnostic to make sure you are measuring the right thing, or model a certain signal propagating through your optical lines. Here we can add things like the thin lens formula, Bragg diffraction, grating diffraction, etc.

I would suggest putting most of this stuff into a separate "plasmapy/diagnostics" directory and then further separate these by category (e.g. optics, lasers, spectrometers). Other things that overlap could go in other sections. For example, I often use a vector class that I created for simulations with 3D distribution functions or for raytracing. Something like this could go under "plasmapy/math"

from plasmapy.