Changepoints detection and magnitude in a time series

[drbenvincent]

Hi. There are a number of other threads here on change point models which could be useful. Eg.

Hierarchical changepoint detection

I’m trying to build a changepoint detection model where both the number of changepoints is sampled as well as their respective locations. From what I understand, this would mean the number of switch statements within the defined model is dynamic, depending on the number of changepoints sampled. So far, this kind of hierarchical model can’t be expressed with PyMC. Am I correct, or is there a way to express this model in PyMC?