Added support for error based metrics

[Thematiq]

No description provided.

[Copilot]

Pull request overview

This PR adds support for error-based metrics (like RMSE) where lower scores indicate better performance, complementing the existing support for metrics where higher scores are better (like accuracy or F1). The implementation introduces a maximize parameter that, when set to False, internally negates the scores before computing win/tie/loss comparisons.

Changes:

• Added maximize boolean parameter to PyBBT class with default value of True
• Modified _construct_win_table to negate scores when maximize=False
• Added test cases verifying correct behavior for both maximize modes
• Updated README documentation with usage example for error-based metrics

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 6 comments.

File	Description
bbttest/bbt/py_bbt.py	Added maximize parameter to PyBBT class constructor and docstring; parameter is passed to _construct_win_table during fitting
bbttest/bbt/alg.py	Modified _construct_win_table to accept maximize parameter and negate algorithm scores when maximize=False; added warning for unnamed columns; changed from set to list for algorithm names
tests/bbt/test_alg.py	Added parametrized test cases for _construct_win_table with maximize=True and maximize=False; added test for unnamed column warning
README.md	Added new section documenting how to use the maximize parameter for error-based metrics

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.