Included PTMCMC and Bilby bayesian sampling techniques in anis_pta.py. Power_basis for now.

[kushagrann]

Added the LogLikelihood, LogPrior and set_ptmcmc functions inside anis_pta class suitable to perform bayesian inference with PTMCMC. Also created a handy function inject_anisotropy suitable for injection studies which creates injected instance attributes of anis_pta for helping with post-processing. The set_ptmcmc function also has a flag argument save_anisotropy which saves the anis_pta instance as a pickle in the PTMCMC outdir for a handy post-processing.
In addition, created a sperate class named set_bilby inside anis_pta.py suitable for performing bayesian inference with Bilby.
Both works with pair covariance. Hence, for it a handy (hidden as it starts with _) instance attribute ._lik_denom is created when we use set_data function which helps reduce computation cost by overcoming calculation of the determinant of the pair covariance matrix every time in the likelihood functions. This instance attribute is only activated if we supply a pair covariance matrix to set_data function.