20180619-minutes.md

Metadata Templates Working Group Minutes

Tuesday, 2018/06/19

10:00 am EDT

Attendees:

Terrell Russell, Mike Conway, Jason Coposky, Kory Draughn, Ilari Korhonen, Alastair Smith, Christine Staiger, Nirav Merchant, Arthur Newton, Dave Fellinger, Deep Patel, Hetal Patel, Ray Idaszak, Jaspreet Gill, Justin James, Dave Fellinger

Agenda:

• Existing Metadata Template Work

  • TRiRODS 2017/05/17
    • https://www.youtube.com/watch?v=_b4AvhhG7mc
    • https://slides.com/irods/trirods_metadata_templates
  • repositories
    • https://github.com/irods/irods_schema_metadata_templates
    • https://github.com/DICE-UNC/metadata-template-service
    • https://github.com/DICE-UNC/jargon-extensions
    • https://github.com/DICE-UNC/jargon-extensions-if
    • https://github.com/irods-contrib/metalnx-web
    • https://github.com/DICE-UNC/irods-cloud-browser
    • https://github.com/irods-contrib/irods_rest_services
  • whitepaper march 2017
    • https://github.com/DICE-UNC/jargon-extensions/blob/development/SpecWhitePaper-MetadataTemplates-03_13_2017.pdf

• Discussion

  • metalnx currently stores template information in the AVUs directly
  • aim for first of july for metalnx v2.0.0
  • irods-ext database as repository for sidecar
  • swagger spec allows application of templates at any level (many-to-many)
  • current work is
    • define a template (json) public/private
    • made of elements, validation, possible grouping
    • so users don't see bag of AVUs - guidance
    • users see forms, not just empty boxes
    • using external SKOS metadata services, controlled vocabularies
  • maastricht has integrated with Disqover, we want a clean look too
    • compared notes at UGM2018
  • formbot is in the development branch of jargon-extensions repository
    • https://github.com/DICE-UNC/jargon-extensions/tree/development/metadata-templates/src/main/java/org/irods/jargon
  • some composability being requested from industry
    • definition coming from 'above'/administrator
    • application-specific capability to merge/add context to admin-defined templates
  • dataverse, cyverse already have ideas too
  • think about capabilities (aka search)
    • what should be grouped when pulled back to the interface
    • in the end, need to be displayed to a human
  • use cases
    • virtual collections could be defined by a set of search attributes informed by metadata template
    • publication
    • validation
  • need to remember to be concerned about scalability
    • exif everywhere, metadata catalog getting large
    • need to separate to other tables/services
    • also ability to limit extra metadata, to minimize cruft
  • QueryArrow… approach is to layer it atop indexing framework (atop 4.2+)
    • do testing for scalability and speed
  • required interfaces for javascript and python
  • iRODS rest services (Swagger definition) https://github.com/irods-contrib/irods_rest_services
  • code to the interface
    • swagger api (provide a service) in front of the json documents
    • insulate the implementation of saving json files in iRODS
    • metalnx will be the first implementation

• ACTION - new repo for this working group in irods-contrib namespace

  • will link back to existing work

• metalnx v2.0.0 progress (aiming for july 1)

  • rules across a zone
  • tickets mostly work
  • rnd working, can be packaged separately
  • microservices mostly working
  • placeholder that says 'NO ACCESS', and possibly request access