VEuPathDB / web-monorepo

A monorepo that contains all frontend code for VEuPathDB websites
Apache License 2.0
2 stars 0 forks source link

EDA tree variable search: consider treating an underscore as white space #758

Open danicahelb opened 10 months ago

danicahelb commented 10 months ago

related to: https://github.com/VEuPathDB/web-monorepo/issues/759

@dpbisme noted differences in behavior for the 'browse and subset' tab when dealing with curated (public) vs user (private) datasets.

For curated datasets, a text query always returns taxa or pathways where the query matches any part of the data. For example, searching 'coli' returns E. coli (along with a lot of other stuff).

However, the same query with a user dataset has different behavior. See example screenshots below for searching 'cutibacterium' in a curated vs private dataset.

Initially, I though that maybe full length exact matches were needed for user datasets, but that's not it either. Searching 'g_' returns all genera, but adding any other character immediately returns zero results.

image image

See slack thread: https://epvb.slack.com/archives/CBLPK9ZD3/p1701492087527609

image
dmfalke commented 9 months ago

Note that those variable names start with a g__ (the letter "g", followed by two underscores "_"), so typing g__Cut should filter the list to just "g__Cutibacterium".