fix: handle scoring=None in GridSearchCV._fit to avoid AttributeError (#1009)

[ekkoitac]

Fix Issue #1009: GridSearchCV scoring=None gives AttributeError

Problem: When scoring=None is passed to GridSearchCV.fit(), line 111 crashes with AttributeError: 'NoneType' object has no attribute 'name' because scoring.name is called without checking if scoring is None.

Fix: Check if scoring is not None before accessing scoring.name. When None, use "test_CPRS" as the scoring name, matching the documented behavior that scoring=None defaults to CRPS.

# Before
scoring_name = f"test_{scoring.name}"  # crashes when scoring is None

# After
scoring_name = f"test_{scoring.name}" if scoring is not None else "test_CPRS"  # safe fallback