sample.theta.sam: Sample theta
In drvalle1/EcoCluster: Bayesian Clustering using Truncated Stick-Breaking priors

This function generates a sample of the vector of probabilities associated with each species group

1	sample.theta.sam(cs, ngroups, gamma, burnin, gibbs.step, betas, theta)

`cs`	vector of size S with group assignment for each species
`ngroups`	maximum number of species groups (KS)
`gamma`	the parameter for the stick-breaking prior
`burnin`	number of iterations to discard as part of burn-in phase
`gibbs.step`	current iteration of the gibbs sampler
`betas`	P x KS matrix containing slope parameters for each covariate (row) and each group (column)
`theta`	vector of size KS with the probability of each species group