I need to explore the properties of the matrix and gradient in jointcal's ConstrainedPhotometyModel, but the sparse matrix code in Eigen doesn't include many options for doing so. To better facilitate this, Pierre Astier suggested outputting the matrix elements and creating a dense matrix from it. This ticket is just for the output and checking that it's sensible: detailed exploration will be in a different ticket.
A quick search suggests that one can pass the sparse matrix to one of Eigen's constructors to directly get the dense matrix. It might be easier to explore the matrix properties with numpy.