waldronlab / BugSigDB

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

Some trouble with Study 562 again #209

Open lgeistlinger opened 7 months ago

lgeistlinger commented 7 months ago

Two issues with https://bugsigdb.org/Study_562:

  1. The bottom of the page displays an error (see below)
Screenshot 2023-12-01 at 5 23 30 PM
  1. Maybe as a result of this error (?) the signatures.csv export from https://bugsigdb.org/Help:Export is malformatted in L626
Screenshot 2023-12-01 at 5 24 51 PM
lwaldron commented 7 months ago

This is a real outlier study that we created to represent the "typical, healthy" microbiomes of various body sites, and it has a lot of large signatures. If accommodating it may compromise performance of the rest of the wiki, we could split it into multiple studies or something.

lgeistlinger commented 7 months ago

Looks like 1. has resolved on its own, and 2. was actually an artifact from inspecting the file in Excel. I don't see such a malformed line when opening the file in vim or reading into R line by line.

jwokaty commented 7 months ago

It sounds like this issue has been resolved and can be closed?

lgeistlinger commented 7 months ago

Yes this can be closed

lgeistlinger commented 7 months ago

Although I do see the error at the bottom of the page again, so we might wanna check whether we want to do something about it. This issue is separate from the new zenodo release which is not affected by this issue.

tosfos commented 7 months ago

It probably depends on how loaded the server is when the page is requested. We'll check on how we can fix this.

tosfos commented 6 months ago

We're going to defer the loading of all the hierarchy information in the tooltip that is displayed on hover until the hover is actually performed. This should cut down on server load and fix timeout issues. The downside is that sometimes it may take a bit of time to load the information depending on the cache status. We think this is a worthwhile tradeoff. We'll experiment and let you know how it goes.

cmirzayi commented 5 months ago

Relatedly, experiment 10 on study 562 shows no signatures at all but I see an extremely long signature when I go to: https://bugsigdb.org/Study_562/Experiment_10/Signature_1

This is perhaps a bit more troubling because there isn't even an error message that something should be there but isn't.

image

cmirzayi commented 5 months ago

Just confirming that Study 562/Experiment 10/Signature 1 does not exist in the CSV exports or in any named version (devel, Zenodo) that I have checked.

tosfos commented 4 months ago

We're going to bump priority on this