ImportError when running yasa.art_detect and fix about yasa

[shengzizeng]

In addition, I think it would be nice if the output of yasa.art_detect could also contain annotation information (including onset, duration [=5, if using window of 5], and description [start with "bad_", so that later epoching can use it for exclusion]). This annotation info can be easily applied to mne annotation through mne.Annotations and will allow for visual inspection of the auto-marked artifacts on the time-domain data.

Although this could be done within a few costumed code lines, I guess it would be nice if this can be done along with the detection function.

[raphaelvallat]

Hi @shengzizeng,

Thanks for opening the issue. It is very unfortunate that pyRiemann isn't maintained anymore. I've actually been considering removing the pyRiemann-based covariance method for artefact rejection because of this exact issue. I hope they release a new version soon, but it doesn't seem like it's gonna happen.

I agree for MNE Annotations. More generally, one could think of a "events_to_annot" function that would allow converting any kind of event detection (artifact, spindles, etc) to MNE-compatible annotations (see also #20). I won't have the time to work on this in the near future, but please feel free to open a PR if you'd like!

Thanks

[ErikBjare]

@raphaelvallat Good news, Alexandre finally merged my PR yesterday! pyRiemann/pyRiemann#93

Thanks for the work on YASA btw, it's been a very helpful reference for my MSc thesis: https://github.com/ErikBjare/thesis

[raphaelvallat]

Thanks @ErikBjare for letting me know, and glad to hear that YASA was useful to your thesis (thanks for citing it btw)! I'm just waiting for Alexandre to release a new version of pyRiemann (hopefully soon) and then I'll update YASA!

[raphaelvallat]

Fixed in release 0.5.0.