Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Joshua Gay
    @jgay_gitlab
    Hi, the BioCompute Object Schema gitlab repo is now ready for ballot. https://gitlab.com/IEEE-SA/BioCompute/bco-schema/
    I created a tag (v1.3.0-alpha) based on the information provided to me previously. This is the version that is the precise version approved to go out to ballot. The zip file to be attached to the balloted materials can be retrieved from the following link: https://gitlab.com/IEEE-SA/BioCompute/bco-schema/-/archive/v1.3.0-alpha/bco-schema-v1.3.0-alpha.zip
    Similarly, the web-view tree version of that version is here: https://gitlab.com/IEEE-SA/BioCompute/bco-schema/tree/v1.3.0-alpha
    Joshua Gay
    @jgay_gitlab
    @NeuroGenomics_twitter I am not sure if you need the above info or if IEEE-SA staff is the one that will send out the ballot materials.
    Jonathon Keeney
    @NeuroGenomics_twitter
    OK great, thank you @jgay_gitlab, I'll ask Tom.
    Hadley King
    @HadleyKing
    @jgay_gitlab Very exciting! Thank you!
    Michael R. Crusoe
    @mr-c
    When is the next CRG call?
    Jonathon Keeney
    @CaffeineSingularity_gitlab
    We don't have one scheduled. We're still working our way through document integration for the accepted comments. It's taking a lot longer than I'd anticipated.
    Jonas Almeida
    @jonasalmeida
    It might help if we add https://gitlab.com/IEEE-SA/BioCompute/bco-schema/tree/v1.3.0-alpha as an integration: https://gitter.im/biocompute-objects/community#integrations (only room admin will be able to do it).
    Hadley King
    @HadleyKing
    @jonasalmeida
    Good idea. I think that @jgay_gitlab would have to do that....
    Hadley King
    @HadleyKing
    I have created a dev branch for the BCO_Specification repo. I have moved the documentation to clean up the repo a little
    Moving forward we use Master only for release versions?
    Stian Soiland-Reyes
    @stain
    I think that makes sense, @HadleyKing, to keep master "stable" - considering the IEEE spec should not change elsewhere. So then you'll upload the new P2791 draft there as well on dev?
    Hadley King
    @HadleyKing
    @stain the dev branch is already staged to match 2791, except for a few lines (using the word BioCompute) and the w3id references.
    so considering that 2791 should not change anymore, I think maybe we could move the dev to master....
    Stian Soiland-Reyes
    @stain
    +1
    Jonas Almeida
    @jonasalmeida
    GitHub actions was just made public at their annual summit last week now enables native continuous delivery of workflows. I’ve been playing with it and suspect this could completely change the way we think about the FAIRness of workflow manifests like BCO. Anyone has thoughts on this? Github is the one place where everybody (40M and counting) gets together ...
    Jonas Almeida
    @jonasalmeida

    From: De Almeida, Jonas (NIH/NCI) [E] jonas.dealmeida@nih.gov
    Sent: Monday, November 25, 2019 9:27 AM
    To: Hadley King hadley_king@gwmail.gwu.edu @HadleyKing … (full list removed)
    Subject: Re: BioCompute Technical Steering Committee Meeting Notes - 08-22-19

    Hi Charles,

    As promised Thursday, the note on running workflows in github:
    40M developers on github woke up last week with a new tab, “actions”, and a workflow engine integrated in their developer environment – which already includes hosting web applications and authentication services, all of this with versioning and a marketplace where workflows can be published and found. So the conversation Thursday was about creating a BCO example in this environment to establish its FAIR use. I suspect most developers (definitely most students) have already migrated to serverless execution environments and having a one-click place where to validate, and disseminate, a workflow manifest/language, really is the place to be if you want to reach them. So the suggestion is to have a simple and compelling example of a BCO as a github action others can fork and investigate.

    References:

    1. Github documentation: https://help.github.com/en/actions/automating-your-workflow-with-github-actions
    2. Keynote at github universe explaining workflow actions: Here’s a link to a vide - https://youtu.be/9EoNqyxtSRM?t=693

    I left the same comment in gitter [this one :smile:] in case you want to discuss it further (not a fan of email conversations here …)

    Cheers,

    Jonas


    Jonas S Almeida, PhD, Senior Investigator
    NIH/NCI/DCEG Chief Data Scientist

    Jonas Almeida
    @jonasalmeida
    Hi Nicola @nsoranzo :-)
    Hadley King
    @HadleyKing
    @nsoranzo
    Jonas Almeida
    @jonasalmeida
    pls see email above - we can use it as a starting point
    Nicola Soranzo
    @nsoranzo
    @HadleyKing @jonasalmeida Are you looking at implementing GitHub workflows to: 1) do CI testing of the BioCompute Editor, i.e. that each PR/commit doesn't break its installation/functionality, or 2) validate BioCompute Objects stored in a GitHub repository ?
    Is it 1) or 2) or both?
    Jonas Almeida
    @jonasalmeida
    Mostly the latter, here’s a little inroad into a more inetractive editor that would test the executability of the manifest as you go.
    The library is at https://github.com/mathbiol/bco, the begginings of an interactive manifest at https://mathbiol.github.io/bco

    I guess the one line summary is that BCO manifests open the door to a web computing version that is particulalry timing for the real-time computational epidemiology of COVID

    Hadley King
    @HadleyKing
    @nsoranzo
    For now, Just 1 from my perspective. I c 2 as a more long term goal, but @jonasAlmeida may be more confident in how to do that now
    Jonas Almeida
    @jonasalmeida
    If it is the case that non-web versions of BCO are a particular case of the webcomputing generalization, then, showing that traditional BCOs can run as a github action will establish its FAIRness. At ISWC last year it was clear to many that reproducibility/portability comes with a time stamp - it has to be realisable with minimal effort in a few seconds. Otherwise people will just not do it. For example, there were papers with implementations as .jar files that were deamed irreproducible simply because it would compromise your machine to run them. Had they been delivered as a github action that woudn’t have been a problem ...
    Nicola Soranzo
    @nsoranzo
    OK, I think I can help with 1) at this stage
    Hadley King
    @HadleyKing
    @nsoranzo
    I have just committed some updates
    It is the start of what I think will work for the UI...
    I have a test app here: https://github.com/HadleyKing/bco-vue-app That I think can be inserted into the Galaxy Vue interface to edit BCOs. And I added a new method to allow downloads (but it still outputs to the browser)
    Jonas Almeida
    @jonasalmeida
    Goo job with API @p-m-ishra ! really enjoyed your presentation :smiley: :thumbsup:
    Jonas Almeida
    @jonasalmeida

    It is thought that a token-based system will be implemented, however, we are open to other methods of authenticating requests.

    my 3 cents:

    1. I think starting simple with an API key as the census does is indeed appropriate.
    2. Are you considering a public (no key) read-only option?
    3. For more advanced scoping (in the future) an OAUTH2 beaer token model may be advisable, particularly if you want the same resurce to support BCO reposition and editing as a mission critical resource for others such as the FDA.
    Hadley King
    @HadleyKing
    @carmstrong1gw Take a look at the comment above. Also more info available here: https://www.census.gov/data/developers.html
    Thanks @jonasalmeida !
    Jonas Almeida
    @jonasalmeida
    np @HadleyKing :thumbsup:
    Happy to help where I can, Nicola @carmstrong1gw
    Hadley King
    @HadleyKing

    @stain @carmstrong1gw made a fork for you for the BCO portal API.

    compare.py is the command line script
    he should run everything in a virtual environment
    he'll probably need to install jsonschema in the virtual environment
    he'll also need a folder to place the BCOs in and they have to follow a certain naming format