Can JacobianPropagatorConverter be relaxed to support more than NumericalPropagator?

Hi all,

I noticed that JacobianPropagatorConverter only works for NumericalPropagatorBuilder/NumericalPropagator.

As far as I can tell, the implementation just relies on MatricesHarvester, which is generic and already implemented for (semi-)analytical propagators as well. So the limitation to NumericalPropagator/NumericalPropagatorBuilder seems overly restrictive.

Is there a specific reason why this converter was kept numerical-only, or could the input be relaxed to accept a general PropagatorBuilder?

Thanks!

Hi,

must have been for historical reasons, thanks for reporting it. I’ve created an MR for generalization:

Cheers,
Romain.

Edit: the converter still needs Cartesian as dependant variables, so it won’t work with DSST for example

Thanks for the swift response, Romain!

And thanks for the review! It has been merged on the develop branch and will be in the long-awaited 14.0 release.

Cheers,
Romain.