refactor: rename RunMetrics to SingleRunMetric and MultipleRunMetric

[iraedeus]

This pull request refactors the metric base classes to clarify the distinction between metrics that operate on a single run and those that aggregate results over multiple runs. The main change is the renaming of RunMetric to SingleRunMetric for single-run metrics and to MultipleRunMetric for benchmark-level metrics, along with corresponding updates throughout the codebase.

Metric base class renaming and refactoring:

• Renamed RunMetric to SingleRunMetric in pysatl_cpd/analysis/metrics/single_run_metric.py and updated all single-run metric classes (e.g., ClassificationMetric, ConfusionMatrix, DelayMetric, RunLengthMetric) to inherit from SingleRunMetric instead of RunMetric. [1] [2] [3] [4] [5] [6]

• Renamed pysatl_cpd/benchmark/metrics/run_metric.py to multiple_run_metric.py, and renamed the base class from RunMetric to MultipleRunMetric for benchmark-level metrics. Updated all relevant imports and class usages. [1] [2] [3]

Codebase consistency and clarity:

• Updated all imports and type aliases throughout the codebase and in test mocks to use the new SingleRunMetric and MultipleRunMetric names, improving clarity and reducing ambiguity between metric types. [1] [2]