Closed lockwo closed 1 month ago
I don't think so. That's not a concept that necessarily makes sense for all paths -- only really Brownian ones. Other paths might wish to represent themselves as e.g. a signature (=Magnus expansion / noncommutative exponential / ...), for example.
Should the abstract path class also accept a space time time levy area: https://github.com/patrick-kidger/diffrax/blob/main/diffrax/_brownian/base.py#L22?