Very simple implementation of gibbs sampling on the classical unsupervised bag-of-words bayesian clustering problem for natural language document topics.