0

I'm trying to add a new discrete distribution to PyMC3 (a Wallenius non-central hypergeometric) by wrapping Agner Fogs c++ version of it (https://www.agner.org/random/).

I have successfully put the relevant functions in a C++ extension and added broadcasting so that it behaves as scipy's distributions. (For now broadcasting is done in Python. I will later try the xtensor-python bindings for more performant vectorization in C++.)

I'm running into the following problem: when I instantiate an RV of the new distribution in a model context, I'm getting a "TypeError: an integer is required (got type FreeRV)" from where "value" is passed to the logp() function of the new distribution.

I understand that PyMC3 might need to connect RVs to the functions, but I find no way to cast them into something my new functions can work with.

Any hints on how to resolve this or general info on adding new distributions to PyMC3 or the internal workings of distributions would be extremely helpful.

I noticed that FreeRV inherits from theanos TensorVariable, so I tried calling .eval(). This leads to another error along the lines that no input is connected. (I don't have the exact error message right now). One thing which puzzles me is why logp is called at instantiation of the variable when setting up the model.

Mark Rotteveel
  • 100,966
  • 191
  • 140
  • 197
Jan Tünnermann
  • 91
  • 5
  • 1
    This is pretty specialized. You should post this to [PyMC Discourse](https://discourse.pymc.io/) where project devs will help you. Also, have you tried looking at [their distributions code base](https://github.com/pymc-devs/pymc3/blob/master/pymc3/distributions/discrete.py)? – merv Jun 10 '19 at 15:32
  • Thanks for the reply! Yes, I have looked at their dists but without much success yet. Thanks for the suggestion. I didn't know about PyMC discourse. – Jan Tünnermann Jun 10 '19 at 22:12

0 Answers0