Also fix type errors from now validated function bodies

This pull request improves the test coverage and maintainability of the V4 data components by parameterizing tests to run against both "itar" and "directory" store types, and by modernizing random number generation for better reproducibility. The changes also remove redundant assertions in array equality checks, making the tests cleaner and easier to read.

Test parameterization and coverage improvements:

• Converted several test classes (TestData4Reload, TestDataNewComponent, and TestPointCloudsSourceIntegration) to use @parameterized_class, ensuring all tests run for both "itar" and "directory" store types. This required changing test signatures and using self.store_type instead of a parameter. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]

• Added parameterized and parameterized_class dependencies to the Bazel test target and test files. [1] [2] [3]

Random data generation modernization:

• Replaced usage of np.random.rand and np.random.randint with np.random.default_rng().random and np.random.default_rng().integers for all random data generation in tests, improving reproducibility and aligning with best practices. [1] [2] [3] [4] [5]

Test code simplification:

• Removed redundant self.assertIsNone(...) wrappers around np.testing.assert_array_equal calls in multiple places, as the array equality assertion is sufficient on its own. [1] [2] [3]