Re-importing a NOFO should create 'archived' past version of NOFO

[pcraig3]

Summary

When someone re-imports a NOFO document, all of the previous sections and subsections are deleted and replaced with completely new sections and subsections.

This has the potential to cause problems — and one time it did, leading us to restore from a database backup to get the old version of the NOFO again.

We don't allow (non-admin) users to completely delete a NOFO, which is why we introduced the concept of "archiving" a NOFO. By the same token, we shouldn't allow people to wipe out all of a NOFO's content, which is akin to what deleting a NOFO actually means.

Functional requirements

We want this all to be invisible from a user's perspective, which means that we should not archive the current NOFO they are using and give them a new URL. People link to NOFOs by id number, so we should keep their existing URLs working. Instead, let's clone their NOFO, archive it, and then overwrite sections & subsections as we do now.

As context: When re-import is successful, we create an internal dictionary of sections and then call overwrite_nofo. This wipes out all existing sections (and subsections).

Instead, let's do something like this:

• Before we overwrite_nofo, duplicate our original NOFO
• The duplicated NOFO should immediately be set to "archived"
• Then we call overwrite_nofo

Also, we should introduce a new NOFO model field called ancestor or something (int field, default null), which would contain the ID of the archived NOFO. The idea here is that a NOFO should know how to find its previous version. If there are multiple previous versions, we can just pull them sequentially.

This puts us in a way better spot so that we can prevent users from messing things up unintentionally or intentionally.