Open JoeZiminski opened 10 months ago
This would be great. So many probes are moving to multi-shank these days.
Hi @rossant would you mind taking a look at this one too? cheers!
Actually apologies @rossant some further testing shows the assumption that channel positions will contain 1 unique x-position per shank is wrong. Will convert to draft and have a think.
Currently in Phy the
sh
entry on the ClusterView and SimilarityView table's is always zero. It seems this is because the files these were originally read from are not longer available. This PR reads them instead from thechannel_positions.npy
.This may require re-running tests somewhere, please let me know if there any other considerations, I have not tested this past loading Phy and checking the
sh
column now contains indices which match the shank with the largest loading for a unit.