Add SparseLabelOp base class

[javabster]

Summary

closes #669

Details and comments

[mrossinek]

I finally had time for some more feedback on the open discussion points.

Besides that, please also relocate (now that #746 has been merged) this class to qiskit_nature/second_q/operators/sparse_label_op.py (and the matching test location).

[mrossinek]

Some feedback for now. I think in general this looks good.

It would be good, if you could also add a release note 👍

I think, we can leave the following aspects for an extension later during the refactoring process:

• more detailed documentation
• addition of Parameter support

[coveralls]

Pull Request Test Coverage Report for Build 3016572442

• 53 of 57 (92.98%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.02%) to 85.315%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
qiskit_nature/second_q/operators/sparse_label_op.py	52	56	92.86%

Totals
Change from base Build 3008133675:	0.02%
Covered Lines:	16744
Relevant Lines:	19626

---

💛 - Coveralls

[mrossinek]

Some minor comments but other than this, this LGTM 👍

[mrossinek]

Following my suggestion over in #756, I will make a similar comment here:

I had another discussion with @woodsp-ibm and we said that it might be beneficial to directly implement SparseLabelOp as a subclass of Mapping. That would give full flexibility when accessing the internal _data object.
The way that you could achieve this, is by something similar to the following:

from collections.abc import Mapping
from typing import Iterator

class SparseLabelOp(..., Mapping):
    ...

    def __getitem__(self, __k: str) -> Number:
        return self._data.__getitem__(__k)

    def __len__(self) -> int:
        return self._data.__len__()

    def __iter__(self) -> Iterator[str]:
        return self._data.__iter__()

This also shows an even simpler implementation of __iter__ which is probably better than what we have right now. It will change it such that the default iteration will only return the keys (and not items) but this is inline with what Mapping does 👍

This will directly enable users to use len(...), .keys(), .values(), .items() and even use op["key"] to access specific entries 👍

[mrossinek]

Other than these suggestions, this LGTM 👍

Note: if you go to the Files tab of the PR, you can batch all of these suggestions into a single commit, which will trigger fewer CI runs 👍

[mrossinek]

LGTM 👍
Thanks a lot, Abby!

I'll await another official review from @woodsp-ibm before merging 👍

[woodsp-ibm]

Could this be fleshed out a little to explain what this is a bit more. Maybe an example usage to or something to give someone a better idea of whats this all about.

[mrossinek]

Thanks, Abby! 🙂