BNLikelihood.compute error "AttributeError: type object 'BayesianNetwork' has no attribute 'from_samples'"

[franz3000]

Environment Details

Please indicate the following details about the environment in which you found the bug:

• SDMetrics version: 0.16.0
• Python version: 3.10.15
• Operating System: ubuntu-24.04.1-desktop-amd64

Actual version of SDV: 1.17.0
Actual version of SDMetrics: 0.16.0
Actual version of Pomegranate: 1.0.0

PyCharm 2024.2.3 (Professional Edition)
Build #PY-242.23339.19, built on September 25, 2024
Virtual Environment

Error Description

AttributeError: type object 'BayesianNetwork' has no attribute 'from_samples'

Steps to reproduce

bnLikelihood = BNLikelihood.compute(real_data=real_data,synthetic_data=synthetic_data)
print (bnLikelihood,"BNLikelihood on data set (categorical colmuns only considered")

Paste the command(s) you ran and the output.
If there was a crash, please include the traceback here.

Traceback (most recent call last):
  File "/home/franz/PycharmProjects/SDV_Metrics/Evaluate/evaluate_sdv_metrics.py", line 207, in <module>
    bnLikelihood = BNLikelihood.compute(real_data=real_data,synthetic_data=synthetic_data)
  File "/home/franz/PycharmProjects/SDV_Metrics/.venv/lib/python3.10/site-packages/sdmetrics/single_table/bayesian_network.py", line 125, in compute
    return np.mean(cls._likelihoods(real_data, synthetic_data, metadata, structure))
  File "/home/franz/PycharmProjects/SDV_Metrics/.venv/lib/python3.10/site-packages/sdmetrics/single_table/bayesian_network.py", line 43, in _likelihoods
    bn = BayesianNetwork.from_samples(real_data[fields].to_numpy(), algorithm='chow-liu')

[srinify]

Hi @franz3000 I was able to reproduce this issue. It's clear that SDMetrics is calling code from pomegranate that's no longer valid in more recent versions of that library but I also couldn't identify a specific, recent-enough version of pomegranate that would still work with our SDMetrics code.

With all of this in mind, I'm going to open a new issue for the team to determine the best path forward!