Add pathways widget and GO-CAM landing to AmiGO pages

[kltm]

Project link

https://github.com/orgs/geneontology/projects/129

Project description

This project is to scoped to

• the addition of the pathways widget to AmiGO
  • gene product
• refreshing the build and deployment of AmiGO enough to support this.
• adding a simple landing page for GO-CAM models in amigo on the /model/ path (External links should go to new AmiGO GO-CAM pages web-gocam#23)

This is a follow-up to Patrick's previous work here: geneontology/amigo#650

We will work with the POs to make the (minimal) changes to AmiGO necessary to support this (this is not meant to be a major push in AmiGO) using current technology and stack.

PI

Chris

Product owner (PO)

Seth

Technical lead (TL)

Seth

Other personnel (OP)

Patrick, Seth (, Tremayne)

Technical specs

N/A
All remaining work is an extension of what we currently have and review is not necessary at this time.

Other comments

Noting previous geneontology/amigo#237

[kltm]

@kltm Bump start this; contact Patrick

[pgaudet]

Some test version can be presented at the GOC meeting

[kltm]

@vanaukenk @ukemi , @pkalita-lbl has made some very nice progress, which is now visible at: https://amigo-staging.geneontology.io .
@cmungall Has also suggested adding the tab to term pages, which I've added to "Discussion". I think we're at the point where we (you all at least) should meet and figure out next steps.

[vanaukenk]

Neato! Thanks @kltm and @pkalita-lbl
I just checked my favorite gene, pmk-1, which has two models.
At a quick glance, they look good, but it looks like we're using the older pathway preview rather than Tremayne's updated version?

[pkalita-lbl]

That's right, it is using a slightly older version of the widget. The latest version(s) have technical dependencies that make it incompatible with AmiGO. It's possible we could work around it in AmiGO, but it might be less effort to rework those dependencies in the widget for a future version.

[ukemi]

I just searched on one of my favorite genes Pkm Mus musculus (MGI:MGI:97591). It should at least bring back gomodel:5745387b00001516, but returns zero models.

[pkalita-lbl]

Yeah, another big caveat is that anything with an MGI ID won't work right now. The new API that backs this needs a fix so that it can resolve MGI:MGI:xxxxx identifiers (in addition to MGI:xxxxx as it currently expects).

[pkalita-lbl]

@ukemi The MGI: prefix has been worked out in the new API and the Pkm models should be showing up now.

[ukemi]

Beautiful! For now I also see the gene-centric models, but at some point we should not display those. A first step might be to use the same filtering criteria that is used other places to assign 'causal model' status. It's not ideal, but a good first step.

[pkalita-lbl]

Chris also pointed that out. I've already made the change, but it's not deployed on amigo-staging yet.

[kltm]

Talking with @pkalita-lbl yesterday, I believe that no tech doc is needed for the remaining parts of this project.

[pgaudet]

• Implementation of the Pathway viewer in term pages will be split in a different project
• @ukemi and @vanaukenk will spec out (soon) which metadata will be shown in the models pages for the current project - which we can get to from the gene product pages

[kltm]

Rescoped; now also see #78

[kltm]

Getting this out live will need to wait on the merged GO API hitting production: #14

[tmushayahama]

Just to note, when LP left I was told to create a GOCAM browser landing page as a web component and it was in working but got abandoned https://github.com/geneontology/wc-web-gocam

Tagging @kltm @pkalita-lbl

[pgaudet]

Hi Pascale,

David and I can do this, but we will need input from Seth (cc'ed here) and perhaps other GO software people.

On two previous manager's calls (8/30 and 9/13), we noted that finishing the pathways widget project was dependent on some of the API refactor work.

Seth, let us know if you have time before Wednesday to touch base about this. Next Monday or Tuesday would work for me; tomorrow is mostly booked.

Thanks,
--Kimberly

[kltm]

[Bringing in email thread]

---

[@pgaudet] As mentioned, as far as I'm concerned, this project has no active parts that can proceed without the production deployment of the new GO API. There is no more work here for anybody but myself and Sierra; the remaining work is probably on the order of half an hour.

While the new GO API--which is required for any further action--is /technically/ deployed, it has some data issues and may need to be withdrawn (geneontology/go-fastapi#77; GO API project currently, but may more to general bug later today) until Sierra can take a look at it. For this reason, we have not proceeded with updating everything to point at the new pages, etc.

As there is no more functional action from Kimberly and David here, it may make the communication loop shorter and more clear if I'm put on as PO, as there are only technical considerations remaining (delta the documentation ticket for you and Suzi geneontology/go-site#1929).

---

As an update to the email thread, it looks like we have the GO API fix in that we wanted, so I'll try wrapping up this project in the next day or so.

[pgaudet]

Hi @kltm I changed the PO and assignee to reflect your comment above.

[kltm]

@pgaudet No more active tickets here; closing out.