Switch to using reference/secondary for insar_isce_burst_job

[forrestfwilliams]

Should be released after the updates to HyP3. Notably, I included no validation of the reference/secondary age order because HyP3 will be checking this for us. Happy to hear feedback that we want this though.

[jhkennedy]

I think you can be a bit simpler here; if users are using positional arguments, they can just use reference and secondary since it's type hinted as Union[str, Iterable[str]] and should be functionally equivalent to using granule1, granule2. So we only need to maintain using granule1 and granule2 as keyword arguments, which we can do with this:

    def submit_insar_isce_burst_job(self,
                                    reference: Union[str, Iterable[str]] = None,
                                    secondary: Union[str, Iterable[str]] = None,
                                    name: Optional[str] = None,
                                    apply_water_mask: bool = False,
                                    looks: Literal['20x4', '10x2', '5x1'] = '20x4',
                                    *,
                                    granule1: Optional[str] = None,
                                    granule2: Optional[str] = None,
                                    ) -> Batch:

Notably, this keeps the same number of positional arguments and keeps name, apply_water_mask and looks in the same position.

Then after the function definition, you would check granules and reference/secondary; something like:

if `reference` is None:
    if `granule1` is None:
        raise ValueError('Either reference and secondary must be provided, or granule1 and granule2')
    reference = granule1

if `secondary` is None:
    if `granule2` is None:
        raise ValueError('Either reference and secondary must be provided, or granule1 and granule2')
    secondary = granule2

and you can drop all of the arguments = locals().copy() stuff (there's no harm in a granule1 and granule2 variable existing).

---

Maybe not this PR, but in the future, it might be worth moving the *, before name, forcing the options to come in as keyword arguments.

[forrestfwilliams]

@jhkennedy I believe we're already doing what you're suggesting (at least in part) here (allowing people to use granules but raising a warning if they do. I like raising an error if neither are specified and will add that. Also correct that we're not breaking yet, just prepping to.

[jhkennedy]

@forrestfwilliams you're breaking the interface here, which you do not need to do with my suggestions. Examples, 👇

[jhkennedy]

You're breaking the interface by not maintaining the same number and order of positional arguments. For example, this works with the current SDK and my suggestion but not with this PR:

job = hyp3.submit_insar_job('S1_...-BURST', 'S1_...-BURST`, 'my_job_name')

[jhkennedy]

I'll just open a PR

[forrestfwilliams]

Thanks!

[forrestfwilliams]

For some reason, mamba/conda cannot locate the build package even though it's on conda-forge. This is causing the tests to fail.

[forrestfwilliams]

This PR is good-to-go from my perspective, but we should hold off on merging it for the time being. We need to wait to update HyP3 and the SDK until Disco can support the breaking change in Vertex.