Ben Pedigo

reducing `n_iter` may also help, though hard to say if it will look nice. How big is the graph (number of nodes and number of edges)?

@dfrancisco1998 are you mainly just working on https://github.com/microsoft/graspologic/issues/613 ? if so, can we move this conversation there instead (since it's the specific thing you're working on for the time being)?

just posting a link to some recent layouts I did using UMAP on ASE/LSE on a brain connectome with ~25,000 nodes, I thought the ASE/LSE stuff looked good! Have not...

Hi @dlee0156 - here's a high level overview of what we are doing for graph layouts conceptually. Input - network (could be networkx object, could be a sparse or dense...

@dlee0156 this is my janky prototype code for layouts: https://github.com/bdpedigo/giskard/blob/e88dbdb67c1c373f946f7295a598abaf375bd579/giskard/plot/graph_layout.py#L25 and this is the code in graspologic currently: https://github.com/microsoft/graspologic/blob/2125f27bc3f2739f4f2c784d5b700417df63c5d7/graspologic/layouts/auto.py#L112

@dwaynepryce @bryantower how do we feel about making a graph undirected (by averaging i -> j and j -> i edge weights) JUST for the purposes of running Leiden for...

I see we talked about this https://github.com/microsoft/graspologic/issues/655 my aggregate you mean sum?

since sum is literally avg * 2 I don't expect there to be any difference in modularity, leiden, etc. We could do either, I just said average because that's what...

Do you think we need to warn the user that this is happening or just do it automatically and explain in the docs?

crap, this is only on 3.10? it looks just like the one we had last week which makes me scared