Add CI for ingesting neuroscope spiking data

[bendichter]

https://gin.g-node.org/NeuralEnsemble/ephy_testing_data/src/master/neuroscope/dataset_1

[samuelgarcia]

@bendichter @CodyCBakerPhD I tried to add this to SI tests. The reader we have in Si in only for recording.

In the (old) neo implmentation, the dataset is supposed to contain a ".dat" file. Here there is a ".eeg" file instead. The xml do not related any eeg file except in the help section. Do you have a reference information about this format and what are the extension for "coninuous" (dat/eeg/lfp/xxx) ? We need to make a patch on neo side.

[samuelgarcia]

I find it. http://neurosuite.sourceforge.net/formats.html The official format page is mentioning .dat, .lfp, .fil as possible extensions for continuous traces. The archive of neuroscope/dataset_1 have a eeg I am not sure this eeg is part of the official format but looks a home made preprocessing (filetering). Do you have info why they change lfp to eeg ? Any way the neo patch should explore some extension for the continuous dat/lfp/eeg/fil

[CodyCBakerPhD]

All the working code for Neuroscope types can be found in the LEGACY spikeextractors: https://github.com/SpikeInterface/spikeextractors/blob/master/spikeextractors/extractors/neuroscopeextractors/neuroscopeextractors.py

.lfp files are read in an equivalent manner to the .dat (same binary structure) but they do have separate metadata (namely the different sampling frequency) in the .xml

.lfp files are equivalent to .eeg in every way. Don't know exactly when or why, or if it's required or just convention.

I've never seen a .fil file over the course of the past ~10 conversions.

[samuelgarcia]

OK thanks. I will add this in neo soon.

[bendichter]

@samuelgarcia since this is for spike sorting data specifically, I think .eeg can be safely ignored here