Faster relative score calculation

[Yuki-Imajuku]

No description provided.

[Copilot]

Pull request overview

Refactors RelativeResults.recalculate_relative_score for better performance by avoiding deepcopy, eliminating the per-user sort, and precomputing rank-based scores once per distinct absolute score via Counter. Also adds a test for the existing MIN-with-zero-score error path.

Changes:

• Replace deepcopy(existing_scores) + append with list-spread to build case_scores.
• Compute best_score for MAX/MIN with a single pass instead of sorting; compute RANK scores once per unique value via Counter and reuse for ties.
• Add test_calculate_min_relative_score_zero_new_score covering the MIN zero-score RuntimeError.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/ale_bench/data.py	Performance refactor of recalculate_relative_score; switches sort-based scoring to single-pass min/max and Counter-based rank computation.
tests/test_data.py	Adds a test for MIN relative-score-type rejecting a new score of 0.

---

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