Add population analysis functionality

[kazewong]

The main focus of jim has been on individual event parameter estimation. One obvious extension that can leverage the infrastructure in jim and flowMC is the addition of population analysis, at least in the way that is commonly done with hierarchical Bayesian analysis.

This would require the following steps:

• 1. Data ingestion pipeline for posterior samples
• 2. Selection function data
• 3. Hierarchical Bayesian model likelihood.

[Qazalbash]

@kazewong will you use vanilla JAX to write a Hierarchical Bayesian model likelihood or any other ppl like NumPyro?

[kazewong]

I think I am more inclined to use vanilla JAX and existing interfaces defined in Jim and flowMC instead of ppl like NumPyro, as users do not need to learn the specific ppl interface in order to understand the code. Also, vanilla Jax integrate with flowMC better.

[CharmaineWONG2]

To-do:

1. Include the priors of individual events in the population likelihood: do we save the priors after the PE?
2. selection effect
3. Is the m_1 inferred the redshifted mass? If so, we may need to add cosmology
4. Composite of population model: eg. for both mass and mass ratio