Richard Höchenberger

> Upstream issue with legend and topo map positions? Yeah, and the big legend is clearly a problem. We'll need to address this upstream, or revise this plot entirely somehow...

> why `.mat` files? I believe these are actually just HDF5 files these days :) Any suggestions for a different data format that would make your life easier? Is it...

> or I've been known to dabble in XARRAY. I'm a big fan of xarray too, WDYT about that, @agramfort?

> sorry what is the question? Whether it could make sense to replace the .mat files we currently produce with something else. @ktavabi suggested xarray could make sense.

xarray allows for [various storage options](http://xarray.pydata.org/en/stable/io.html), including netCDF; but one caught my attention: they have a [dictionary export](http://xarray.pydata.org/en/stable/io.html#dictionary) that would then allow one to store the data in a JSON...

> > 04:31:34 [Step-05][sub-301] Saving epochs > > Overwriting existing file. > > 04:31:34 A critical error occurred. > > Traceback (most recent call last): > > ...events, mappings =...

> The pipeline doesn't re-run routines like epoching prior to grand averaging, decoding, or TF steps given necessary derivative files are present. Correct? Yes this is correct if you request...

Good to hear it's working! Please do feel free to share **anything** you don't like about the report. There's so many things to be improved there, but user feedback can...

What is the recommended way to programmatically retrieve versioned objects (files, directories) from a dataset? Also, will the `openneuro` CLI app continue to be supported?

Great, thank you for the explanation, @nellh!