Add Person 3 ML training and evaluation pipeline

[ChenouZhang]

Summary

Add the Person 3 machine learning training and evaluation workflow for annotation-based models.

What changed

• add dataset loading utilities under confostate.data
• add model registry, baseline training, training pipeline, and evaluation modules under confostate.models
• expose new package entry points in confostate.__init__
• add command-line scripts for dataset loading, model registry, baseline models, training, and reporting
• include a baseline logistic regression model artifact and metrics in data/models
• update usage and packaging docs for the new ML workflow
• add Person 3 planning and progress documents

Why

This branch establishes the initial end-to-end ML workflow for Person 3 so training, evaluation, and reporting can be run from a consistent code path instead of ad hoc scripts.

Notes for reviewers

Please focus on the model/data APIs, script entry points, and whether storing the baseline model artifact in-repo matches project expectations.

[ChenouZhang]

@copilot resolve the merge conflicts in this pull request

[Copilot]

@copilot resolve the merge conflicts in this pull request

Resolved the merge conflicts in commit 439a57a.

[orbeckst]

✅

[ChenouZhang]

LGTM. @orbeckst