How to generate synthetic data with Approx-Private-PGM (LocalInference)?

[RexYuan]

As titled, I'd like to know how to call .synthetic_data on a LocalInference like with FactoredInference

[ryan112358]

This library does not support generating synthetic data with Approx-PGM, that's one of the limitations of that approach. In principle, you could use the GUM algorithm described here to do it: https://www.usenix.org/conference/usenixsecurity21/presentation/zhang-zhikun

I'd be happy to accept a pull request if you want to implement that!