Peculiar behavior- Descriptions

[ecurrano]

I am working on Bighorn Basin data entry and noticed that a bunch of Description nodes have become separated from their specimens. I believe that I correctly entered the Description nodes, and also, it should not be possible to enter description nodes without linking to a specimen (required field). There are not too many descriptions here, and so I am manually reconnecting things, but curious if anyone knows what might have happened so that this does not happen again.

BHB Meh MLA Description in the Bighorn Basin P-E Boundary group is an example of a description with no specimen:

Also, when I am working through changing things, the Description edit search has changed. Here are screenshots of what I see:

and then when I click on groups:

(all the groups I belong to have been selected.

[NoisyFlowers]

@ecurrano

The Description search changed with #359. This was to make the dialog's search behavior consistent with that of the outer Description select box, which shows you all of your Descriptions across your Groups.

Selecting a Description, whether from the select box or the search dialog, just populates the fields in the Description edit form with data from the db. Nothing about this would separate a Specimen from a Description. Nothing is written until you click Submit.

Can you tell me more about what you were doing when the Specimen and Description became separated? The more I know, the better chance I can reproduce the problem.

For instance, are the Specimens that are separated from their Descriptions still in the db or have they been deleted? Can you also provide an example of a Specimen that previously had a Description? What action(s) were you taking prior to when you noticed the separation?

[NoisyFlowers]

Never mind. Found it!

There was a misconfiguration in the API regarding the name of the field through which descriptions are passed on Specimen mutation. In addition, the API was not sufficiently hardened regarding the requirement that Descriptions have a Specimen. This combination allowed the API server to delete DESCRIBED_BY relationships on Specimen mutation.

This is fixed in paleobot/pbot-api@f9ab84a.

On ddm-dev branch and running on dev.

Note: Editing a Specimen on prod will continue to orphan it from its Description until we deploy this fix.

Now, once deployed, what to do about the existing victims?... You can manually reattach them via the Description edit form. If there are too many for this to be viable, let me know.

[aazaff]

Can you make a list of the detached ones for @ecurrano

[NoisyFlowers]

Here is a file of Descriptions without Specimens. This is not a valid configuration, so these are all victims of the bug and need to have their Specimens reattached.
DescriptionsWithoutSpecimens.csv

Here is a file of Specimens without Descriptions. This is a valid configuration, so there is no way for me to discern which of these are actual victims as opposed to just Specimens that don't have Descriptions.
SpecimensWithoutDescriptions.csv

[ecurrano]

Doug should deploy. Julian, Ellen, and Dori will work on reuniting the orphaned Descriptions with Specimens.

[NoisyFlowers]

Bug fix merged to master and running on stage/prod.

I'm leaving this open to track db corrections.

[ecurrano]

Julian, Dori, and Ellen will all work on correcting these and log what has been reunited here: https://docs.google.com/spreadsheets/d/12ht-MzsdYJs1kMAl6_w6Z6DAqhscBBnTXDRdnqQq2Z4/edit?usp=sharing

[ecurrano]

There are only a couple of these left, and Ellen needs to deal with them.