Revert "Feature/dz0scf gradient clean"

[wtpeter]

This pull request removes two example scripts and makes several code cleanups and small refactorings related to the Dz0SCF and NTTDA gradient functionality. The main changes include deleting the analytic gradient example scripts, removing the reference_energy method override in the Dz0SCF implementation, and cleaning up test assertions and mixin methods that are no longer necessary.

Example and Documentation Cleanup

• Deleted the example scripts examples/grad/02_dz0scf_grad.py and examples/nttda/02_nttda_dz0scf_grad.py, which demonstrated analytic gradient calculations for Dz0SCF and NTTDA. [1] [2]

Dz0SCF Implementation Refactoring

• Removed the _charge_rks, make_rdm1s, and gen_response methods from the _DZ0VeffMixin class in src/nest/dz0scf/dz0scf.py, as well as related attributes, simplifying the code and removing unused or redundant logic.
• Removed the override of the reference_energy method in the Dz0SCF implementation, relying on the default behavior.

Test and Assertion Cleanup

• Removed a redundant assertion in the test_svwn_dz0scf test, streamlining the test logic.

NTTDA Gradient Module

• Deleted the src/nest/grad/nttda/__init__.py file, which contained the implementation for analytic and finite-difference NTTDA gradients.Reverts Feature/dz0scf gradient clean #10