Service Desk (from dreardon@swin.edu.au): Passing **kwargs inputs to models via likelihood functions

Hi Bilby devs,

I have some model functions that require auxiliary data. This data needs to be pre-computed so the model evaluation is quick. It's therefore necessary to be able to pass additional information (via **kwargs for example) to model functions.

However the built-in likelihood functions do not allow this, with the inputs to GaussianLikelihood for example being just (x, y, func, sigma=None), with the expectation of func to be of the form (x, param1, param2, param3, ...), where paramX are the model parameters to sample. Can the likelihood functions be altered to allow for additional arguments to be passed to the models, without these arguments being confused for model parameters? e.g. GaussianLikelihood(x, y, func, sigma=None, **kwargs) and func(x, param1, param2, param3, ..., **kwargs) ?

Thanks for your help with this, Daniel