Joint analysis of multiple statistics

[marcpaterno]

On a somewhat related note, is it possible to (manually) build a data vector that includes xi_plus, xi_minus, and Cl_EE that share the same systematics? So not necessarily with the automated tools like make_all_photoz_bin_combinations.
Such joint analyses of multiple statistics on the same catalogue are useful to check the consistency of different statistics (see e.g. https://arxiv.org/abs/2503.19442)

Originally posted by @tilmantroester in #508

[marcpaterno]

Please take a look at https://firecrown.readthedocs.io/en/latest/_static/two_point_factories.html#recommended-full-metadata-data-extraction. If two different TwoPoint objects were created, and they shared the values of all the keys in the different factories, then they would have distinct systematics objects, but the different objects would be identically configured. This identical configuration would extend even to the names of any parameters that are potentially sampled. Thus any parameters being sampled would obtain the same value for each MCMC sample.

Does this answer your question?

[tilmantroester]

Ok, that sounds like what I have in mind. I'll need to play around with this to understand better if there are edge cases that might break.