'ContinuousFactor' object has no attribute 'variables'

[ilincaburdulea]

Subject of the issue

I am trying to set up a simple graph structure of 2 nodes (x and y) and 3 factors: f_x, f_y, f_x_y. The probability distribution functions I used for now are not relevant, I wanted to have a working toy example using pgmpy first and then I'll adapt my whole problem statement. However I have problems with using the function add_factors, because I get an Attribute error. Does anybody have any advice? It might be that I'm doing something wrong.

Your environment

• pgmpy version 0.1.25
• Python version 3.10
• Operating System

Steps to reproduce

Tell us how to reproduce this issue. Please provide a minimal reproducible code of
the issue you are facing if possible.

import numpy as np
from pgmpy.models import FactorGraph
from pgmpy.factors.continuous import ContinuousFactor

np.random.seed(0)
num_points = 30
x = np.linspace(0, 10, num_points)
y = x + np.random.normal(0, 0.5, num_points)

fg = FactorGraph()

def pdf_2d(x,y):
return np.exp(-((y - x) ** 2) / (2 * 0.5))
def pdf_1d(x):
mean = 0
variance = 2
return np.exp(-((x - mean) ** 2) / (2 * variance)) / np.sqrt(2 * np.pi * variance)

factor_x_y = ContinuousFactor(variables = ['x', 'y'], pdf = pdf_2d)
fg.add_nodes_from(['x', 'y'])
fg.add_edges_from([('x', factor_x_y), ('y', factor_x_y)])
fg.add_factors(factor_x_y)

Expected behaviour

Tell us what should happen
add_factors should work so I can continue with belief propagation

Actual behaviour

---

AttributeError: 'ContinuousFactor' object has no attribute 'variables'

[ankurankan]

@ilincaburdulea Sorry for the late reply. Unfortunately, continuous factors aren't supported fully yet, and I would suggest not using them at the moment.