waldronlab / BugSigDB

A microbial signatures database
https://bugsigdb.org
7 stars 6 forks source link

Strict checking of vocabulary-controlled fields #148

Closed lwaldron closed 1 year ago

lwaldron commented 1 year ago

Body site, Sequencing platform, and Statistical test currently allow free-form entry even though we have lists of allowed values. Here is an example: https://bugsigdb.org/Study_563/Experiment_2

Can we lock these down to only allow use of the allowed vocabulary? I saw a curator who had entered mixed-capitalization of the Condition field this summer but had already corrected it, and I wasn't able to free-form Condition, so I guess this has since been locked down?

tosfos commented 1 year ago

Will do

tosfos commented 1 year ago

We restricted these fields:

  1. Body site
  2. Condition
  3. Sequencing platform
  4. Statistical test

Please review and close if it looks good.

ftzohra22 commented 1 year ago

@lwaldron @tosfos It seems like the condition field is sometimes not auto-filling data from the ontology site/condition page (please check screenshot below from Study 181 that I'm curating). "Irritable bowel syndrome" is a valid EFO term as well as listed on our condition page but shows no results found. Given that this field is restricted, I'm also unable to enter it as free text. We've run into this issue once before with 'Huntington disease' but it was fixed the next day so I didn't bring it up.

image
lwaldron commented 1 year ago

I confirm I also see the same as @ftzohra22

tosfos commented 1 year ago

We've been struggling with this type of issue for a while. Ultimately it looks like a PageForms bug. And indeed there's a note in the code internals that mentions that the feature will struggle with a large set of autocomplete values. We're going to spend some time on a full solution.

lgeistlinger commented 1 year ago

Just noting that I experience the same for the body site field when eg trying to enter an existing term "buccal mucosa":

Screen Shot 2022-10-22 at 10 54 49 AM
lwaldron commented 1 year ago

I hear this is interfering with curation, so is there some temporary workaround if the proper solution will take a while?

tosfos commented 1 year ago

We'll check. I also bumped up the priority for this task.

lgeistlinger commented 1 year ago

Hi @tosfos just wondering what is the status of this one?

lgeistlinger commented 1 year ago

This would be a priority for us as it precedes other clean-up of eg body sites in #151

lwaldron commented 1 year ago

Agreed, this should be top priority because it is still interfering with curation.

tosfos commented 1 year ago

We've been waiting for the script that's resolving #111 to complete. That script does look done now, so we'll move forward on this. We'll have to perform a container restart. If all goes well, this should be resolved by tomorrow.

tosfos commented 1 year ago

We believe things should be much improved. Please test and let us know what you find.

lgeistlinger commented 1 year ago

Can confirm - tested the terms that gave issues above and autocomplete would show up fast with these terms. Can we remove the note "Due to a temporary issue, autocomplete may not load the full list of possible options." from the entry form @tosfos? This can then be closed.

ftzohra22 commented 1 year ago

Checked and can confirm its working

image
tosfos commented 1 year ago

That's great news! Removed the warning on both Experiments and Signatures.