AI OnAI Off
I "solved" this by making an epi job that will delete content based on an integer ID of the content. This way I could delete the false root node and then the code will use the correct one. It does require you to retag the content brought over in the import though.
I'm currently checking to see if this pull request is the fix for this issue:
https://github.com/Geta/EpiCategories/pull/37
So, yesterday, I believe I found a bug in geta categories and epi import/export. We exported/imported some pages from pre-production to production for a DXP client and clicked the checkbox to include sub-items and to export files that the content links to.
Now, when we imported this to prod, it appears that it exported some geta categories, and somehow made a new root node for categories
This made us think they had missing categories, or someone had accidentally deleted the bulk of our tags. However, the "old" ones are still there, just not recognized as the category root
Here is the export screen from our production environment:
Now, on prod, we can see what appears to be 3 different "category root" nodes. The "wrong" one is currently used what clicking the Categories tab in the assets pane and when interacting with category related properties in edit mode. The "right" one, can only be seen here and is not used in assigning geta categories. The third one with just a dash appears to be from another import where this happened, except it was ordered after the right one, so it didn't overwrite the category root. The content id of the "right" one is 14 and the content id of the "wrong" one is 29688. Maybe there is a simple database query I can write to correct this?
In addition to the above photo, I can also find the "right" categories by referencing pre-production to get the category id (content id) and then manipulating the edit-mode url on prod.