Add the ability to use arbitrary Ansatzes to calculateP about bet HOT 4 OPEN

Just to clarify, is this still true or does the sample object's probabilities attribute take care of this?
From my glance at calculateP (very early stages of learning to read code written in parallel), it seems like this is no longer an issue.

from bet.

This is still true. The standard ansatz is used in all of the prob methods in bet.calculateP.calculateP as the line:

P[Itemp] = discretization._output_probability_set._probabilities[i]/Itemp_sum

and :

P_local[Itemp] = discretization._output_probability_set._probabilities[i]*discretization._input_sample_set._volumes_local[Itemp]/Itemp_sum``

from bet.

I think this could be done by using the emulated input sample set appropriately.

from bet.

I actually accomplished this by setting probabilities to 1/num_samples if empty, and then using that vector in place of volumes in the main loop of calcP.prob.

@lcgraham This feature is coming down the line very soon, I think I can make it it's own PR. It will be aligned with setting an "initial" distribution which we "update". If unspecified, uniform is assumed.

Check out the animation in #161 to see it in action.

from bet.