Complete sample overload signatures

[danielgaskins]

Reference issue

Closes #28

What does this implement/fix?

Lists every keyword-only sample() parameter in both overloads instead of hiding them behind **kwargs. The Literal[True] and Literal[False] annotations still narrow the return type to torch.Tensor or dimod.SampleSet.

This also adds a release note.

Additional information

• python3 -m pytest -q tests/test_boltzmann_machine.py (27 passed, 6 subtests passed)
• python3 -m pytest -q --deselect='tests/test_samplers/test_base.py::TestTorchSampler::test_parameters' (156 passed, 70 subtests passed)
• reno lint
• The deselected test has an existing local PyTorch CPU/meta-device failure that is reproducible on unchanged main.

AI Generation Disclosure

I used Codex to inspect the overloads, update the signatures, draft the release note, and run validation. I reviewed the changes and test results before submitting.