Contraction order for "generalized tensor networks"

[GiggleLiu]

Wondering how difficult it is to extend optimaltree function to handle more general contraction, the "generalized tensor network" in

• https://arxiv.org/abs/1710.01437
• https://arxiv.org/pdf/1806.05964

Or, must we switch to treewidth like algorithm? like
https://arxiv.org/abs/quant-ph/0511069