InferenceEngine on infinite loop after creating latent variable

[achuinar]

Description

I am working on the creation of a model to predict fire occurrences.

I generated this graph:

and extracted the CPDs from it. Then initiating the InferenceEngine works fine.

Then, I tried to introduce a latent variable, run the latent CPDs calculation and re-create the InferenceEngine but it keeps running forever which makes me suspect an infinite loop like what was previously reported in a similar ticket.

Context

I wanted to perform a

Steps to Reproduce

1. Add node to network

bn.add_node(
    node="high_number_of_people_prone_to_inattention",
    edges_to_add=[
        ("high_number_of_people_with_not_enough_sleep", "high_number_of_people_prone_to_inattention"),
        ("high_number_of_single_parents", "high_number_of_people_prone_to_inattention"),
        ("high_number_of_people_with_disabilities", "high_number_of_people_prone_to_inattention"),
        ("high_number_of_elderly_people_75plus", "high_number_of_people_prone_to_inattention"),
        ("high_number_of_people_prone_to_inattention", "fire_density")
    ],
    edges_to_remove=[],
) 

2. Set boundaries

from causalnex.estimator.em import EMSingleLatentVariable

boundaries_inat = EMSingleLatentVariable.get_default_box(
    sm=bn.structure,
    node_states={
        **bn.node_states,
        'high_number_of_people_prone_to_inattention': {0, 1}
    },
    lv_name='high_number_of_people_prone_to_inattention'
)

boundaries_inat['fire_density'][1].loc[1 ,0] = .8 

3. Fit latent CPDs

bn = bn.fit_latent_cpds(
    lv_name="high_number_of_people_prone_to_inattention",
    lv_states=[0, 1],
    data=full_training_sample,
    box_constraints=boundaries_inat,
    n_runs=30,)

4. Try to generate InferenceEngine

from causalnex.inference import InferenceEngine
ie = InferenceEngine(bn)

Expected Result

Should generate InferenceEngine object

Actual Result

Runs forever

Your Environment

Include as many relevant details about the environment in which you experienced the bug:

• CausalNex version used (pip show causalnex): Version: 0.11.0
• Python version used (python -V): 3.8.7
• Operating system and version: Mac OS Big Sur 11.6.7