Binary Markov-K Random Generator

Hello Stackoverflow Community,

currently I'm working on an entropy encoder (MQ-coder) implementation (cython wrapper and internal c source code). To create a test setting, I want to use a binary markov-k random generator, that outputs numpy arrays as input for the encoder. What would be the easiest way to implement such a generator in python, numpy, scipy, or tensorflow? The transition probabilities p(x|x1,...,xk) and the order k should be possible to specify.

Thanks in advance, meridius