search

NCATSTranslator / Feedback

A repo for tracking gaps in Translator data and finding ways to fill them.
7 stars 0 forks source link

Facelift of Filters for Path Type #851

Open sstemann opened 2 months ago

sstemann commented 2 months ago

In the few queries that I have run - one-connection is the total number of results. seems minor.

i believe there will always be a one-connection path:

lookups - by design are one-connection inferred - the inferred path is one-connection

https://ui.test.transltr.io/main/results?l=CYP3A4%20(Human)&i=NCBIGene:1576&t=1&r=0&q=9bb65a47-375c-4851-ae7c-050b55e05b33 image

https://ui.test.transltr.io/main/results?l=Common%20Cold&i=MONDO:0005709&t=0&r=0&q=fcd657be-e1e0-4aff-9fcd-3c8d16ab16c7 image

dnsmith124 commented 2 months ago

@sstemann this is absolutely true, and something we discussed.

Currently all of the other facets function such that if a facet is absent from the list (Metabolite is not among the ChEBI roles, for example) it means that there are no results for which that facet applies. It is odd to include 1 hops as an option considering it should always apply to all results, but having it not be present implies, based on how all the other facets work, that none of the results have one hops.

Here's an example where there are no 3-hop paths in a result set, so that facet doesn't display: https://ui.test.transltr.io/demo/results?l=Breast%20Cancer&i=MONDO:0007254&t=0&r=0&q=723a8922-0aae-4e1b-aca2-8706bb902619 Screenshot 2024-07-09 at 1 21 22 PM

If we don't show the 3 hop facet because none exist in the set, and then also don't show the 1 hop facet for a totally different reason it could be confusing for users.

Ultimately we decided to leave the one hop facet for this reason, despite it not being very useful itself as a facet (enabling it includes everything, disabling it excludes everything). Definitely open to being convinced otherwise, but that was our thought behind including it!

sierra-moxon commented 2 months ago

@sstemann @dnsmith124 - maybe this is closable? Or is there another thing to try here to make the one-hop counts in the facet values more intuitive?

gprice1129 commented 2 months ago

@sierra-moxon we can leave this open for now. The UI team has plans to discuss facets (in general) and this specific facet for the Fugu release.

sstemann commented 1 month ago

@gprice1129 this facet seems to be the same as described above in CI and Test, should we change the label or close?

dnsmith124 commented 1 month ago

@sstemann this change is part of the general UI facelift that's currently in progress, just updated the label.

sstemann commented 3 weeks ago

Changes should be:

Filter Group > Paths

Specific Filters Path Type - Direct, Indirect Objects within Paths - Chemical Entity, Disease,etc Support Path Length - 1+, 2+, 3+