New functions for posterior evaluation

[Wenda-Lou]

This PR introduces two helper modules for evaluating the quality of MCMC posterior distributions from ECM (Equivalent Circuit Model) fits and integrating those evaluations into the AutoEIS pipeline.

New Modules
1. evaluate_posterior_distribution.py

• Implements a scoring function for MCMC posterior samples.
• Evaluates each component using:
• R² from a Q-Q plot against a normal distribution.
• Skewness of the sample.
• Number of peaks in the KDE-estimated distribution.
• Optionally saves KDE plots and histograms for each parameter.
• Returns a functionality score and list of problematic parameters.

2. update_circuits_info.py

• Applies the scoring function to each InferenceResult in a circuit list.
• Updates the circuits_df DataFrame with:
• Posterior scores.
• Bad parameter flags.
• Sorted ranking based on score and parameter quality.
• Computes and appends fitness metrics using AutoEIS core tools.

These tools enhance AutoEIS by:

• Quantitatively assessing posterior reliability.
• Flagging multi-peak, skewed, or poorly fitted parameters.
• Enabling better model selection via posterior-informed ranking.

[ma-sadeghi]

Hi @Wenda-Lou, thanks for the PR! You can find my comments under "Files changed". I'm still trying understand the PR better. I'll give it a try tonight or tomorrow and will get back to you.

[Wenda-Lou]

Thank you for your feedback. I developed this module alongside Zhecheng Yin, another undergraduate student. He should have started a separate PR outlining the basic workflow of this module. It might be easier to review the code in conjunction with the workflow outlined in his PR.