Interface starting from total causal effect estimates

[seongwoohan]

Hi LLCB developers,

I am interested in running LLCB on a simulation dataset. It seems like the current version assumes that the input data should be structured as indicated in the repo rather than from total causal effect estimates. The example simulation data looks like a 30x30 adjacency matrix representing the causal graph, where each entry indicates the strength and direction of the causal effect between variables like the table below (showing only 5x5). This is a ground true graph and it would be nice to have an interface starting from total causal effect estimates from this graph via something like (I-G)^-1 to see how LLCB can recover the underlying structure

I really appreciate it.

[weinstockj]

Great, thank you for creating an issue. An interface for this is definitely possible; I'll start thinking about what that would look like. Also happy to review a PR.

Josh