Ontology Working Group GitHub Meeting, Tuesday April 3rd, 2015

From Planteome.org
Jump to: navigation, search

Agenda and Notes:

Attending: LC, JE, CM, JP

1. Discussion of OWL Tools

  • CM- you can use OWL Tools now to convert OBO to OWL, is now on GitHub, https://github.com/owlcollab
    • help file is there "Command line examples"
  • OWL Tools is undergoing a rewrite, will be called "Robot" (Robot is an OBO tool)

2. Setting up the GitHub Repository

  • CM- Granularity discussion- do we want to have all the ontologies in one repo or have separate repos for each ontology?
    • Question of whether we want the same trackers or separate ones- but we are keeping the ontology trackers at SourceForge
  • If the ontologies are each in their own repo, we can have a separate mini-wiki in the repository
  • See the "GitHub_Possible_Format- gSlides
  • Permissions are controlled per repo, so we probably want to have the reference ontologies in one and the application ontologies in another
  • CM- Recommend not having the different formats in different folders, but think of the source files and the release files. OBO and OWL files should go in the same directory together.
  • source (src) directory- contains the files for the release
  • doc directory contains documentation files in markdown format
  • Commit messages are given specific ids
  • Possible way to set up the repos- have a repo for the external ontologies
  • It makes sense to do the final release of the PO for the PSDS paper, at Github, rather than putting it out and then moving everything

3. Existing Repositories:

  • plant-trait-ontology- was set up last spring 2014 for the TRY meeting,
    • Using the create repository script, has a number of extra features such as the image and tools folders
    • RW submitted a list of traits from the TRY database, but they have never been added. We should keep this list and eventually get them integrated.
    • This version of the TO is out of sync with the one on our svn.
  • planteome-ontologies (was Ontology-Files)
  • If we have all the ontologies in one repo, then they will be grouped here

4. Moving to Protege 5.0

  • CM: Recommend staying in OBO format for now to facilitate doing the difs, but can do the editing in OBO in Protege 5.0
  • LM: Looks like Protege 4.3 also allows opening and saving OBO files
  • Advantage of the upcoming Protege 5.0 is that you can import other ontologies
  • This will be very useful once we start working on the PSO
  • Need to specify which format the curators should be saving in.

5. Setting up repositories for other parts of the project on GitHub

  • JP- We will need to have a place to put the code for the image annotation developers source
  • Planteome web site- will be doing quite a bit customization, so that will need to be stored somewhere
  • data warehouse code
  • Probably each individual component should be their own repo.
  • Not sure where the association files are going to be stored, as iPlant has said they do not want to store them there- possibly at Github or leave on our SVN?

6. Next Ontology meeting April 7th

  • Elizabeth and Leo will be there
  • CM should be available
  • Need to get the collaborators into the concept of using the repositories.
  • Suggest uploading the Trait Template as a CSV file, as long as it does not have a lot of functions etc.

6. Plan of Action:

  • For ontology editing: Keep working in OBO format for now, to facilitate doing the difs, but start using Protege 5.0

LM: Looks like Protege 4.3 also allows opening and saving OBO files? Have installed both on my computer.

  • We will save the list of traits from the TRY database that Ramona created in a working/scratch folder. "Done- added the TRY list to the TO folder on SVN"
  • We can get rid of the existing "plant-trait-ontology-TRY" repository (note- I renamed it to avoid confusion)- Who is doing this?.
  • Set up the Plant Trait Ontology at Github and start testing it and using it- but need to import the history- See below and the [All-Hands meeting notes 4-16]
  • Set up 1 repo for the planteome-ref-ontologies (use lower case and dashes to be more "githubby")- Done, renamed the existing repo, but it needs to be reconfigured.
  • Future of SVN needs to be discussed
  • Justin E will make a cleaned-up version of the SVN, so that we can keep the history of the files. And then move them over to GitHub. May just make it as a test to see how it will work and then get rid of it. "What is the status on this??"

Follow-Up by email 4-21

  • Moving from SVN to Github:
    • Justin is working on getting the Trait Ontology loaded, along with the history- still working out a few bugs. JE is going to remove the existing 'planteome-ref-ontologies [[1]]' repo so we can start fresh
      • CM: OK, I still recommend doing this via a planteome.owl file we will make- A single point of configuration. We're going to be doing things like adding bridging axioms between ontologies; we don't want to have to change an amigo conf each time.
  • Moving SourceForge trackers to GitHub:
    • CM: Are you going to try and migrate either the SF history or the commit history? If you are interested we are doing something similar for GO (analogous situation: tickets in SF, SVN in our own hosted server): google doc
    • We need to keep our trackers on SourceForge- so Justin is working on moving the commit history only
      • CM:I really strongly advise to you move to using GH for tickets. SF will presumably go away eventually. It's very bad for searching tickets (the GO editors are burned constantly). Plus, you will really really really like having the tickets and VCS integrated, trust me. You will. You will not look back.
  • Ontology editing:
    • Keep working in OBO format for now, to facilitate doing the difs, but start using Protege 5.0.
    • LM: Looks like Protege 4.3 also allows opening and saving OBO files as well?
      • CM: It's broken. And still broken in 5beta [2], [3]
      • This means that if you want to work in obo and Protege you will need owltools on the command line to convert

Back to the Planteome_Ontology_Development_and_Coordination_Working_Group page