There are 2 boxes of marbles: box I contains 7 red marbles and 3 black makbles, box II contains 2 red marbles and 8 black marbles. Drawing marble arcording to follow steps:

- Step 1: select a box randomly with probability box I is p (0 <= p <= 1)
- Step 2: drawing a marbles from selected box and without replacement

After doing above process 3 times, we get the results of red, black and red marbles, respectively.

Use Bayesian interfere with pymc, determine the p values?

I want to find the posterior p, which is the probability of selecting box I when knowing the result of 3 times drawing without replacement. I want to do with pymc python library