Applied Linked Data 2017-01-10

Owned by Bill Branan
Jun 16, 2017
Time: 10:00 AM PST / 1:00 PM EST
Google Hangout: https://plus.google.com/hangouts/_/gv3gejk2am4xggnpntydvheh3aa (doesn't seem to be working at the moment)
Backup Phone Line If Hangout Doesn't Work: 712-775-7035 (Access Code: 960009) (Use this)


Attendees:
    
Agenda:
  1. Linked Data Fragments Caching Update
    1. Lynette progress (Marmotta):
      1. https://github.com/ActiveTriples/linked-data-fragments/tree/cache_config
        1. Notes on this: https://confluence.cornell.edu/display/~elr37/Linked+Data+Fragments+%28LDF%29+-+scratch+pad
    2. Steven progress (Blazegraph):
      1. Bullet point of shame (no progress).

  2. Small API Changes:
    1. Pull request by tamsin johnson: https://github.com/ActiveTriples/linked-data-fragments/pull/33
    2. That pull request only updated the in-memory RDF.rb. Steven is currently working to update Blazegraph and Marmotta to this spec.
      1. Question on contexts in Blazegraph and Marmotta: how exactly are they used? Blazegraph isn't likely using the setting correctly at the moment.

  3. Other Recent Pull Requests
    1. https://github.com/ActiveTriples/linked-data-fragments/pull/36
      1. Documentation; already approved.
    2. https://github.com/ActiveTriples/linked-data-fragments/pull/35 
      1. Dependency version update; needs work in RDF::Blazegraph gem.
         
  4. CHF / DCE update
    1. deliverables https://docs.google.com/document/d/1HHRlcK6e6Jz5qXUj7fXJ2Bccs9N8KXpXbwmk1ww9AEQ/edit?usp=sharing
    2. waffle board https://waffle.io/curationexperts/chf-sufia
       
  5. Resolving string literals in Linked Data Fragments (short discussion; experimentation presentation still pending)
    1. From last time, we would add a passthrough SPARQL endpoint to Linked Data Fragments.
      1. How configurable? Just at <server> : <port>/sparql ?
         
    2. QA would be configurable to know of this endpoint. It would be updated to allow the passing of correct SPARQL queries to Linked Data Fragments. An example SPARQL query for MESH might be:

      1. prefix nlm: <http://id.nlm.nih.gov/mesh/vocab#>
        SELECT DISTINCT ?term_ident ?plabel
        WHERE { { ?term_ident nlm:prefLabel ?plabel
        FILTER regex( str(?plabel), "^#{q}", "i" ) }
        UNION
        { ?term_ident nlm:prefLabel ?plabel
        FILTER regex( str(?plabel), " #{q}", "i" ) }
        }
         

    3. However, this goes against the principles of Linked Data Fragments API. Is there a better way to do matching? What if one wants to match against prefLabel, label, and description for terms?

  6. Changing from Working Group to top level Interest Group update
    1. Still working on this.

  7. <Add additional agenda items here>

Next Meeting:  January 23 at 12pst/3est