obofoundry / obofoundry.github.io Goto Github PK

Currently we have a distinct script https://github.com/owlcollab/owltools/blob/master/OWLTools-Oort/bin/build-obo-ontologies.pl

Used in this job
http://build.berkeleybop.org/job/simple-build-obo-all/

That makes all the fallthru files on http://berkeleybop.org/ontologies/

Eventually everyone will take charge of builds but for now, the metadata hardwired in this script should be placed in the main md files

Each person (OBO team, or possibly also ontology developers) could have an entry in the config.yml file

E.g.

team:
  - name: Chris Mungall
    orcid: ORCIDURL
    github: cmungall
    roles: [twg,...]

Or alternatively, one md per person (that can be collected same way as ontologies)

This would be easy to render as html in jekyll

The YAML would have a json-ld context allowing direct translation to RDF

cc @ontoden

We could just output ontologies.txt from https://raw.githubusercontent.com/OBOFoundry/OBOFoundry.github.io/master/registry/ontologies.yml

But may be better if ontobee could take the native format directly. Note the new yaml offers better possibilities for ontobee users. For example, some ontologies have >1 namedGraph imported, but this is AFAICT done informally using an ad-hoc internal registry, which we can now centralize

I believe the new OLS takes in the yaml, cc @simonjupp

Suggested by Erick Antezana on obo-discuss:

• download icon redirecting to various formats: obo, owl, ...

cc @proccaserra, Peter McQuilton

notifications:
  email:
    - [email protected]

happy to do this during the teething phase, but should go to wider group. Others should add emails here https://github.com/OBOFoundry/OBOFoundry.github.io/edit/master/.travis.yml -- or should we just use obo-operations list?

Many of our pages have a domain field. This is somewhat arbitrary but could be normalized.

It would be useful to have a page that lists all ontologies within a domain. Even better if this can be customized. For example, different community members could take responsibility for different domains: health, anatomy, environment, ...

How should this repo be licensed?

Should different components be licensed differently? How do we indicate this with a LICENSE file?

Some suggestions:

1. ontology metadata CC-BY or CC-0
2. Site tooling and web interface: python script to make yml; jekyll HTML templates: BSD3
3. content authored primarily by OBO Foundry (FAQs, principles, id-policy, how-tos, etc): CC-BY

The authors of 2-3 are primarily members of the OBOF. For 1 could be considered joint authorship between individual ontology maintainers and OBOF.

See for examples PRs #61 #60 #59 #58 - I modified these so that the ftp site was a (cardinality 1) 'page' property.

No reason to restrict page to be a single valued; could be a list (but homepage should be functional, as it is mapped to foaf:homepage)

And/or: have new properties for different page types; e.g. ftp_site - todo - explore different vocabularies for this

The repository essentially forked at one point, @proccaserra's STATO went missing.

as per the discussion on OBO foundry operation call on August 11th, 2015.

Finally we have a field for licenses. Note ontologies without declared licenses (most) get visually highlighted on their page.

OBIB (ontology for biobanking) does not show on ontology page:
http://obofoundry.github.io/

How to add the OBIB link in?
http://obofoundry.github.io/ontology/obib.html

Suggested by Erick Antezana on obo-discuss:
broken link (star): http://obofoundry.github.io/link-to-foundry-principles

We are lacking this for ms, and it looks like the others use the old style sourceforge URLs

cc @germa @ontoden

$ grep code.goog ontology/*md
$ grep sourceforg ontology/*md

Some may be from obsolete ontologies, but even so best to have these migrated for posterity

FBcv (FlyBase controlled vocabulary) is tagged as deprecated.

http://obofoundry.github.io/ontology/fbcv.html

I would say this is not the case. Although the Drosophila Phenotype Ontology, which is included in FBcv, is released separately, the whole of FBcv is still updated regularly.

I think it would be useful to allow for multiple contacts for an ontology, including multiple people or mailing list(s).

• freeze metadata additions to legacy googlecode/sourceforge site
• plan to redirect obofoundry.org before googlecode dies
• announce in advance to obo-discuss list
• decide on protocol for making changes moving forward

The old http://obofoundry.org site had a lot of pages that haven't been touched since 2007ish

These are currently in the navbar under "LEGACY"

I suggest removing forever

Clicking on details for FBdv shows only this sentence at the top of the page

"A structured controlled vocabulary of the development of Drosophila melanogaster."

http://obofoundry.github.io/ontology/fbdv.html

I'd like to update the details, but I'm not able to.

Currently hardcoded to be the ancient obo format version I made. Should be made the complete ontology now it is in OWL. obo-basic can be a subproduct

Migrating from #18 cc @LENOV

kisao is still on SF - I strongly recommend migrating to GH. See https://docs.google.com/document/d/1iyVY8kDBJIEydoWFLG9j5BoO4VmdIXES95fVhIkFXfk/edit#

(there is no tracker enabled on SF, we used to have https://sourceforge.net/tracker/?group_id=293617 registered but this no longer works)

Is this correct?

Suggested by Erick Antezana on obo-discuss:

• add an extra button to retrieve metrics for a selected ontology: number of entries/classes, number of relationshios/properties, depth, ...

I really like the buttons, but would it be possible to have the target displayed a mouseover?

I set up

https://github.com/orgs/OBOFoundry/teams/community

this group doesn't have permission to directly commit or merge (I prefer this responsibility is taken on by a small circle, remember anyone is welcome to fork and PR). However, it does allow for assignment of tickets on trackers, which is useful if we need a non OBO Foundry person to handle something. It may also auto-subscribe to tickets? (which can be easily unsubscribed)

I recommend a policy whereby all main ontology contacts are encouraged to join the community team

Our documentation is spread in multiple places:

1. On the old obofoundry.org site, mostly stale. See: #25
2. On the wiki running at berkeley, http://wiki.obofoundry.org/wiki/index.php/Main_Page (the principles are here)
3. On the obo operations google code wiki. See #23
4. In the obo operations google drive https://drive.google.com/drive/u/0/folders/0B8x6Blq96uycalZvM1BLRjJmb2c - mostly meeting notes, but other stuff too
5. other google docs. E.g. I see lots of outreach meeting notes but no parent folder shows up for me

In addition to being spread out, much of it is confusing and stale.

I recommend consolidating as much as possible into github pages - meeting agendas and notes are best left in google docs, but for anything else, one place.

There is an argument to use the github wiki for some purposes, this has some advantages over jekyll/gh pages: easier to create new pages (but how often do we do that?), slightly easier to interlink. My vote would still be for pages

http://obofoundry.github.io/ontology/swo.html

James has left the EBI

@helenp should this be you?

We would like to implement better interoperation with ncbo - ideally they would pull our metadata

Did someone volunteer for reaching out on the obo op call? @proccaserra? Bjoern?

See also #28

When using jekyll 2.5.3

       Deprecation: The 'pygments' configuration option has been renamed to 'highlighter'. Please update your config file accordingly. The allowed values are 'rouge', 'pygments' or null.

Suggested by Erick Antezana on obo-discuss:

• it would be great to be able to sort the ontologies (e.g. by name, by modification date, by prefix, ...)

Note: see #2 - for now we just listed Foundry at the top.

Proposed recommendation:

Keep the description field concise. Don't write it like a paper abstract. Say immediately what the ontology contains.

Add more details as necessary in the markdown section

These ontologies don't resolve from the yaml file

http://www.berkeleybop.org/ontologies/dpo.owl
http://www.berkeleybop.org/ontologies/kisao.owl
http://www.berkeleybop.org/ontologies/lipro.owl
http://www.berkeleybop.org/ontologies/nmr.owl
http://www.berkeleybop.org/ontologies/ovao.owl
http://www.berkeleybop.org/ontologies/resid.owl
http://www.berkeleybop.org/ontologies/rnao.owl
http://www.berkeleybop.org/ontologies/xso.owl

Consider using angled text for the icon columns
https://css-tricks.com/rotated-table-column-headers/

See for example http://obofoundry.github.io/ontology/chebi.html

Is there a way to have wrap around or something similar that would display the full link instead of cutting with '...'

cc @zhengj2007

Is there a way to have headers following as you scroll down, or repeat them in the middle?

The website was seeded from the uberon one. Scoop all that out and replace with old content

We currently use depicted_by to allow people to include a logo, see http://obofoundry.github.io/ontology/obi.html

• Should a different property be used? Any semweb vocab have a property for logo?
• We should make it clear that the URL must resolve to actual image
• We should have guidelines on size (or dynamically resize?) - see #70

cc @alanruttenberg @kltm

Can we just use a CNAME?

No product in the YAML for the following ontologies
hao
miapa
obcs
opl
sibo
tads