fix: Fix targetting class to be neuron compatible #898
Conversation
| if (model := by_name.get(model_name)) is not None: | ||
| ps_ids = list(simdata.placement[model].load_ids()) | ||
| ids_here = [ps_ids.index(ids_i) for ids_i in ids if ids_i in ps_ids] | ||
| dict_target[model] = simdata.populations[model][ids_here] |
There was a problem hiding this comment.
What is this piece of code supposed to do? It's written very inefficiently, if I understand the input/output relation perhaps we can optimize it.
There was a problem hiding this comment.
yes, my understanding of ByIdTargetting is that it should select those cell models with (global) id as in ids.
Originally ids was used as index in the subpopulation of each simdata: simdata.populations[model][ids] (kind of local id).
Here ps_ids has the cellmodel ids of that specific simdata, then ids_here finds whether and where any of the ids that we are looking for is in ps_ids (like np.where).
There was a problem hiding this comment.
Could you give a numeric example with let's say 3 or 4 elements per array so I can visualize the transformation?
shortcut Co-authored-by: Robin De Schepper <[email protected]>
Describe the work done
CellTypeFilterwhenget_targets()method is called.