[Draft] Code prototypes

[Doomsk]

Writing all the code prototypes in Qadence 2 that @jpmoutinho is generating on Qadence 1.

Regardless the examples, some points to consider in the TODO list:
[ ] properly handle dtypes (João pointed out to include the set_default_dtype(torch.float64) in the __init__.py file of qadence2-core folder)
[ ] fix the issue with the inputs on the qadence2_platforms.backends.pyqtorch.embedding.Embedding.__call__ method containing variational parameters (variables) as well
[ ] replace dictionaries to OrderedDict on the inputs and/or variational parameters

[pimvenderbosch]

Hi Eduardo, thank you for the good work on this PR. I have a few comments to be addressed.
Also, I see that the test suite doesn't have 100% coverage yet, could you have a look into the parts of the code that are not tested yet?

[pimvenderbosch]

Can this import + default type setting be pushed down into the pyqtorch subfolder? If you're using a different backend importing torch is not necessary.

[pimvenderbosch]

Docstring should reflect this change.

[pimvenderbosch]

Docstring should reflect this change

[pimvenderbosch]

Docstring to be updated.

[pimvenderbosch]

used on the expectation method

[pimvenderbosch]

Why should the protocols in this module be runtime_checkable?

[pimvenderbosch]

For reducing the number of indents:

Suggested change

	if observable is not None:
	return platform.expect(
	obs_list=parse_native_observables(
	num_qubits=len(self.sequence.register.qubit_ids), observable=observable
	)
	)
	raise ValueError("observable cannot be None or empty on 'expectation' method.")
	if observable is None:
	raise ValueError("observable cannot be None or empty on 'expectation' method.")
	return platform.expect(
	obs_list=parse_native_observables(
	num_qubits=len(self.sequence.register.qubit_ids), observable=observable
	)
	)

[pimvenderbosch]

Would it be a good idea to group the tests per backend? This way they stay organized when there are many more backends implemented. Like so:

├── tests
│   ├── fresnel1
|   |   └── test_xxx.py
│   ├── pyqtorch
|   |   └── test_yyy.py

This way the folder structure also matches more closely with the source structure.

An additional suggestion is to define conftest.py per backend for fixtures that are backend specific.