PowerGridModel/power-grid-model

AI Summary: The project needs to replace several `TypeAlias` definitions for `np.ndarray` with custom classes inheriting from `np.ndarray`. This change is necessary because MyPy currently treats distinct aliases as the same type, preventing effective use of function `@overloading` and clear type distinctions. The task involves creating 7 new classes, each with specific validation in `__init__` and a `from` static method to aid migration.

AI Summary: The Power Grid Model (PGM) currently suffers from suboptimal performance because its caching mechanism for expensive calculations (like topology and Y_bus construction) invalidates the cache after every scenario if it couldn't be reused, even if subsequent scenarios could benefit. This issue proposes a more intelligent caching strategy that combines "reverse update data" and "new update data" to prevent premature cache invalidation. The goal is to maximize the reuse of pre-calculated data across scenarios, thereby improving overall performance.

AI Summary: The issue proposes adding a utility function to `power_grid_model.utils` to reorder scenarios within a `BatchDataset`. This reordering aims to improve calculation performance by grouping scenarios with identical topologies and, within those groups, scenarios with identical electrical parameters. The function will return a new, reordered `BatchDataset` copy and the original indices mapping.

AI Summary: Implement a validation test case runner within the power-grid-model package, allowing users to easily run their own validation cases and potentially submit them to the project. This also involves creating a separate `dataset_equal` function for robust comparison of datasets, handling angular values and tolerances.