Refactor Vasp output classes

[pmrv]

The vasp.interactive sub module defined subclasses that did not change the functionality of the base classes. As far as I can tell they were not used except in VaspInteractive.interactive_close, so I have removed them but left the imports in place in case any HDF5 files still refer to them.

[coveralls]

Pull Request Test Coverage Report for Build 9225747625

Details

• 4 of 4 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.08%) to 71.101%

---

Totals
Change from base Build 9181543489:	0.08%
Covered Lines:	10626
Relevant Lines:	14945

---

💛 - Coveralls

[jan-janssen]

I think that is a very reasonable suggestion, originally there was:

    def to_hdf(self, hdf):
        """
        Save the object in a HDF5 file
        Args:
            hdf (pyiron_base.objects.generic.hdfio.ProjectHDFio): HDF path to which the object is to be saved
        """
        with hdf.open("dft") as hdf_dft:
            # hdf_go["description"] = self.description
            for key, val in self.log_dict.items():
                if isinstance(val, list) or isinstance(val, np.ndarray):
                    hdf_dft[key] = val[:-1]
                else:
                    hdf_dft[key] = val

This was then fixed around five years ago f2d6e66 and afterwards nobody cleaned it up.

[jan-janssen]

I feel this is not included in the to_dict() and from_dict() as the VASP base implementation stores the corresponding entries in Output.generic_output.dft_log_dict while the VASP interactive implementation uses Output.dft_output.log_dict.

[pmrv]

I'm not sure how it is handled, but the vasp.interactive.Output was literally not used as far as I could tell. When I create a test vasp job in a notebook (pr.create.job.Vasp) its job._output_parser object is the vasp.base.Output, only when VaspInteractive.interactive_close is called this was overwritten. I can test again with a "proper" interactive vasp job, though, in case there are any weird edge cases.

[ligerzero-ai]

LGTM, up to you to see if edge cases exist, but I don't think it's worth the time.

[pmrv]

Test failure discussed in #1444, merging anyway.