Refactor HTMX Implementation to avoid duplicate create forms

[caseyhans]

Refactors our HTMX views to avoid having multiple create forms at once, since this introduced bugginess.

The general pattern is this (refer to lit workflows as a good example):

• In the template where you want the create form to exist, add a div with the classes create-row hidden
• On the create button, add hx-target=".create-row" hx-swap="outerHTML"
• In the edit form template, add the class create-row if this is a create and not an update (so that duplicate button presses will replace this item). Don't delete the form if a user hits cancel, just hide.

[caseyhans]

@shapiromatron ready for a draft review--still have UDF binding and mgmt task viewsets left

[rabstejnek]

There were a few bugs I witnessed when testing this out:

• The cancel button doesn't work on label create forms in the HAWC Prime flavor. More specifically it looks like the hidden css is being overridden by flex css after conferring with Casey.

Other optional changes:

• I noticed usage of classes for the create row htmx trigger. In practice this should only be a single element, so I'd personally prefer using an id, but it doesn't really make a difference with functionality. This change may be a bit branching so I'd be fine if it wasn't addressed.

Edited this comment after some bugs turned out to be local.

[caseyhans]

@rabstejnek I can't recreate any of these bugs. The first two are css-related, so a hard refresh might fix it. Not sure about the last bug. I'll keep trying to replicate the bugs and reach out for a meeting otherwise. I'll also look into using IDs for the create rows, that's a good call

[caseyhans]

OK the label cancel bug was some weird issue with the 'hidden' class being overwritten by 'd-flex' class with PRIME style. Good catch! Glad we're fixing this, not sure where else this bug might have been present...

[caseyhans]

Ok, I tried swapping the create-row htmx target to IDs instead of classes, but it was introducing some weird bugginess to the autoscroll script that I couldn't pin down. Since it's not really a functional change anyways, I think we're OK to leave it as a class. If the change is worth the time, I can keep working on it.

[shapiromatron]

Ok, I tried swapping the create-row htmx target to IDs instead of classes, but it was introducing some weird bugginess to the autoscroll script that I couldn't pin down. Since it's not really a functional change anyways, I think we're OK to leave it as a class. If the change is worth the time, I can keep working on it.

Seems reasonable to me, I think this looks like it's ready for my review. Thanks yall!!

[shapiromatron]

Functionally, it works great, and fixes the bugs I introduced with my alternative approach. Since the selectors on pages with multiple creation forms use a dual selector, .foo.bar, I think this should work fine in practice. https://stackoverflow.com/a/3772305/906385

Nice job!