Redundancy SortingAnalyzer computation with kilosort output

[RobertoDF]

Hi,

I have a question about the SortingAnalyzer object. Right now I use it to compute various extensions, though, some of these are already computed by kilosort: amplitudes, spike positions, templates, pc_features. Would it be a good idea to make it possible to load this already computed .npy files in the analyzer? I don´t see a way to do that. Btw, I am willing to put some time into this if makes sense.

[zm711]

In general the goal of the analyzer is to "analyze" rather than to load. This is because SpikeInterface gives you fine control over exactly how you want to do that analysis. This allows for reproducibility in your analysis. Also thinks like amplitude from kilosort are not true amplitudes, but scaling factors for templates, so the values might not be exactly what you think they are. All the extensions for the Analyzer play well together and can be scaled automatically with the return_scaled parameter (as long as your raw data format allows it).

Does all that make sense?