bsotomayorg / spikeship Goto Github PK

Hi Boris,

SpikeShip itself is running smooth and fast but getting my spiking data in the right input format for SpikeShip is currently taking very long and I was wondering if you have conversion functions already. The required input format is very unintuitive to me and I'm losing a lot of time in for loops but I can't easily see how to make it faster.

My data is in the format used by the IBL (and I believe Allen as well). It consist of an array 1xN spike_times which contains all the times of N spikes. Then there is a 1xN spike_clusters array which contains the neuron identity of each spike. Then I have a Nx2 events array which contains the start (first column) and stop (second column) times of all my N events.

Do you happen to have a conversion script to go from this data format to the one required by SpikeShip or do you have an idea on how to do this efficiently?

I'm working on implementing SpikeShip on my data but I often have the issue that after running diss_spikeship = spikeship.distances(ripple_spike_times, ii_spike_times) there are NaNs in diss_spikeship. I've tried setting various minimum criteria (minimum number of neurons/events/spikes etc) which sometimes works, but not always. Under which conditions does the algorithm output a NaN exactly? Then I can make sure to exclude those.
Cheers!

How can we install on Mac? setup.py is deprecated...