E. Lynette Rayle will describe current state of wikidata work (as identified on Prioritization page) and then ask Christine Fernsebner Eslao what we might consider "done" for the next pass on wikidata, then perhaps create new and more specific issues for later improvements
2019-10-18 Steven has had some discussion, question remaining is what might be "good enough" for this round given Sinopia's ability to let the user go out to the native environment and bring in a URI. Lynette comments that no response from wikidata folks who seemed interested in the feedback we gave. Will discuss in a future QA "study hall"
2019-10-25 On agenda for study hall 1 Nov
2019-11-08 Discussion in study hall, suggestion is that we need to explore all wikidata APIs to make the fetch faster, sense is that the search speed is adequate for now. Currently known APIs return huge amounts of data and are thus very slow; using the SPARQL endpoint will require some extra code in QA to POST the query and parse result → Investigation DONE, will follow up with fetch on https://github.com/LD4P/qa_server/issues/15 (comment in issue for next steps)
John Skiles Skinner , Huda Khan and others working on white paper on knowledge panels as followup from Blacklight LD @ Stanford
2019-10-25 Will be discussed in DAG next Tuesday
2019-11-08 Discussed in last week's DAG meeting, to be revisited by next DAG meeting 11/26
Huda Khan to discuss with Astrid and David possible collaboration with U Chicago over usability (and maybe others in DOG team)
Meeting with Astrid, Lynette, Steven and David and Emma set for 11/21 (Tim is out that week) to discuss potential usability work for discovery and Sinopia
Huda Khan working with EBSCO on scheduling vocabularies and data tool demo
STARTING with items not in DIscogs but AWAITING more work in Sinopia to import data. Sinopia work cycle 2 (through December 6) will we hope include the ability to read in RDF back from Trellis. We hope that we can leverage this to import RDF from a lookup in Discogs or ShareVDE.
Tim worked on call number browse with classifications. Some issues with cases where more than one book with the same call number, it gets hung up – won't fix for current demo. Also, pulling in classifications the LC call number classification path is not consistent between works. Also id.loc.gov have entities for only about half of the call numbers classifications → Jason will write to Nate and Kevin to ask about whether this work can be completed
Huda has been working on subject browse using additional index
Huda has worked on timeline, also using additional index with start/birth/death dates,
John has been working with DB with open syllabus info (CIP codes). Has carousel of related books based on syllabus data
Huda will do some screen shots for the partner meeting, will do video after meetings
Wrap up BAM! and do demo next week
SMASH! (to start real soon, run through end of January)
Sinopia work cycle will end December 6. After then we can only expect big fixes (perhaps somewhat expansively defined) through the rest of this grant
Understood difference between DIscogs and QA context results (hash or array of hashes) and have decided that Sinopia will always expect the more complex form. Tim has modified the discogs authority to match this pattern. Output from QA will be unified when these changes have been merge. Lynette will reviews Tim's PR for this today.
Lynette will also complete another PR (refactor with some other things) for community review today
Will release new QA with these when approved. Will then need Sinopia team to make the corresponding adjustments, Jeremy says this should be straightforward
Question of when Sinopia shows extended context – currently this happens in term lookup but not in the larger search ==> we expect that context will be necessary to select search result, should discuss on Wednesday. Jason will find some good examples in SHARE-VDE data to illustrate the point
Given limited time remaining in Sinopia work cycle, items in https://github.com/LD4P/qa_server/issues/238 that require both QA and Sinopia work (except pagination which is high priority in both) probably should be de-prioritized