Inspiration from NeuroKit?

[jdkent]

I've used the NeuroKit package for analyzing my physio related data in the past: https://github.com/neuropsychology/NeuroKit

Don't know what the mission of this package is, but in case no one knew about NeuroKit, it may be a useful resource for inspiration (or using as a dependency?)

[tsalo]

It does seem like NeuroKit could be a very useful dependency. It doesn't seem to do anything specific to MRI/imaging, but there are a number of more basic functions (e.g., ecg_phase) that could be used instead of trying to write them from scratch. Thank you!

@smoia, @CesarCaballeroGaudes WDYT? Does NeuroKit seem useful? I know you want to minimize dependencies, but it could be worth it.

[CesarCaballeroGaudes]

Indeed, we'll consider it. Thanks!!

[smoia]

I've been keeping an eye on NeuroKit2 since it (re)started, as we actually started at the same time with physiopy. I remember we had a quick discussion about it during one of our early community meeting too!
I don't have anything against it - in fact, it's a solid, well tested and maintained library, and reinventing the wheel is not a good idea.
I do wonder how much of it we need though (since they don't spare dependencies) - and I am quite confident about the fact that if we want to adopt it, it should go into physutils rather than in phys2denoise. Once again, this is a discussion we started elsewhere (physiopy/phys2bids#186) and never concluded. Feel free to comment there!

[DominiqueMakowski]

Hey guys, neurokit's maintainer here ☺️

I've been likewise following your progress, and I'm glad to see the vitality of python's physio ecosystem! I am very keen on looking for synergies, and we should indeed look for instances where we can integrate and work together, rather than having too much redundant features/code, that in the end increases the burden for maintainers and users.

Our goal is indeed to be viable and useful as a dependency and a basis for other packages (hence we try to provide well-implemented and tested (and efficient) low-level signal processing tools as well as a variety of published pipelines). We are in very active development, so If you need anything or have some ideas, please do let us know so we can adapt and add features or anything if need be.

since they don't spare dependencies

Just to clarify here, we aim and try to keep the core dependencies to a minimum (i;e., ['numpy', 'pandas', 'scipy', 'sklearn', 'matplotlib']). Aside from that, we then have a larger number of deps that are used for testing (for instance, some of our pipelines are tested against other existing implementations in other packages, such as biosppy etc.) and we also have in a few places some optional dependencies (e.g., this function will run only if the user has mne installed).

In any case, let us know if we can help 🤗