Rename this package to jwst_datamodels

[braingram]

Maybe something like:

• rename the repo (github will redirect to the new repo)
• update development installs (jwst, regtests, asdf downstream, jdaviz) to point to the new repo
• make a pypi release with the new name
• release an empty last version of stdatamodels that has 1 requirement jwst_datamodels

[emolter]

@nden what's your take on this, is it worth doing and should we go ahead and do it?

[braingram]

As a point of reference webbpsf is being/is renamed as stpsf. We could follow a similar procedure as in:
spacetelescope/webbpsf#951
(although I suspect we could rename the repository instead of creating a new one).

[melanieclarke]

I'm not sure I see the value in this. It would be a big, disruptive change for all downstream software as well as users. Since JWST datamodels are still "st" datamodels, the current name is not inaccurate. Going from a general name to a more specific one also means that it would be harder to add any shared functionality here if we did eventually want to do so.

[braingram]

The more general "stdatamodels" name has led to confusion many times about the usage of this package with folks thinking that it is used by roman (which it is not). Renaming the package would avoid this confusion.

I think this can be handled in a way that is no more disruptive than any API deprecation using the process defined above. We can consider keeping the "stdatamodels" namespace package (with a single dependency of "jwst-datamodels") active for an extended period of time (if not forever) to allow all existing usage to continue to work (with a warning that "jwst-datamodels" is the new preferred package).

[melanieclarke]

I suspect that confusion stems more from the fact that we originally intended to use this package in romancal, then decided not to, than from the name itself. Either way, I don't see that lingering confusion over the use of this package sufficiently justifies the effort to change its name. Even if it is not more disruptive than any other API deprecation, an API deprecation for a widely used package is very disruptive!

My vote would be to work on updating the documentation for this package instead, to clarify its current use and intent.

[braingram]

I was curious how widely used this might be. Looking at pypi stats there were ~3k downloads last week and considering there are ~4.7k downloads of jwst most of those were likely due to this being a dependency of jwst.