Look into systematizing the noctua-models migration SOP into a pipeline

[kltm]

Currently, I'm cutting and pasting from the SOP doc. This is error-prone and an all around not-so-great idea. Take a look at getting the meat of the noctua-models-migration process encoded in a pipeline file:

1. So I don't have to run in locally and transfer large files in and out
2. Get access to a larger machine == goes faster
3. Anybody can run it
4. (Opens the door for easily testing changes before we apply them)

[kltm]

Oh yeah--the i/o issue on our RAIDed servers...

[kltm]

Still poking at it and the GH commands seem to gum up a little locally. Trying to reset the state into the known: all other work gone, upstream master at HEAD.

[kltm]

Okay, some progress.
We probably want to experiment with this with on a branch so we can semi-manually complete the process and see how it works.
At an hour, that takes a while; but it is mechanical, so that hopefully makes up for it taking out the guesswork.

[kltm]

@vanaukenk Moving forward, I believe I've found a way of generating the potential diffs for a migration within a pipeline. I've set it up to drop the reports here: http://skyhook.berkeleybop.org/noctua-models-migrations/reports/
Assuming we are able to run it end-to-end, we should be able to a test run in about an hour with the push of a button.
geneontology/noctua#782

[kltm]

Wondering if I can just commit straight from the repo after running this? Might be the best of both worlds...

[kltm]

As an intermediate step, I could pause the pipeline and copy the resulting models out. That gets us a little closer.