A function for enumerating backdoor adjustment sets in causal graphs

[adam2392]

Is your feature request related to a problem? Please describe. In causal estimation tasks, one is interested in typically using covariate adjustment to estimate the total causal effects given a fully specified causal graph (i.e. nx.DiGraph, or pywhy_graphs.ADMG).

In addition, these sets exist for Markov equivalence classes as well: pywhy_graphs.CPDAG and pywhy_graphs.PAG

Describe the solution you'd like Implement https://www.jmlr.org/papers/volume18/16-319/16-319.pdf construction algorithm.

The implementation should be careful to note computational complexity of the algorithms. For example, many graph set listing algorithms are exponential in the number of variables, so ideally this function returns a generator, rather than a fully specified list.