Open stiber opened 2 years ago
Actually, a (maybe sorted, maybe that doesn't matter) vector of ints is better; we can just iterate through that and indirectly access the listed neurons.
After some thought, I think that it's better to make our code simpler and just always record from all vertices. We have run simulations with 10,000 vertices and 600,000,000 time steps and the output files, while large, were still workable for analysis. We can always select a subset of the data to analyze after the fact. This change is also consistent with us ripping out all of the code that "pre-analyzes" data within the simulation and writes out that information; instead, it's better to record the raw data and do all analysis afterwards, using Matlab or whatever.
So, I will change the title of this issue to reflect that.
Note that this will also involve a change to the config files. (As an aside, since it is a removal of something, the old config files will still work with the updated simulator.)
Right now, we retain a copy of the probed neurons list. We can simplify our code by making this a vector of bools, either true or false, with size equal to number of vertices.
https://github.com/UWB-Biocomputing/Graphitti/blob/eccbec4bf65666580bd4b7f36589666ebd0fde2d/Simulator/Recorders/Hdf5Recorder.cpp#L262