2020-10-26: Development Meeting

Kacper, Tommy, Craig, Mike L, Mike H.

Agenda

  • Updates
  • No dev call next week (11/02) ([name=Kacper])
  • Default ORCID behavior ([name=Tommy])
    • Related to last week’s discussion of this PR
    • Behavior is looking more like…
      • If the user is connected to an ORCID and creates a new Tale, add their ORCID
      • If the user creates a Tale while not connected to ORCID, leave it blank
      • If the user has a Tale and connects an ORCID, do not modify the Tale author fields
        • First problem is determining which Tale to modify
        • Second is determining which author to modify
  • FAIR Convergence Symposium 2020
    • WholeTale: Implementation of Provenance Queries (TBD) ([name=Tommy])
      • Any ideas about what this might be?
  • Should we plan a release? ([name=Craig])
    • YES!
    • Mid-Decemberish
    • Kacper’s update from Exec call last week:
      • Applied for a new allocation on Jetstream for 2021.
      • Minor release: WT v0.9.1 was redeployed at TACC/Jetstream with beefier nodes (especially more HDD)
        • Notable features wrt v0.9:
          • Most recent R/Rstudio support
          • Updated DataONE client stack
          • Migrated to py3.7+
          • Ability to select the r2d version used to build a Tale.
      • Work on v0.10 is ongoing and will include:
        • Git Integration (details)
        • Significant overhaul of workspace/home backend (virtual_resources)
        • Versioning (details)
        • New, rewritten Dashboard
        • Full publish/ingest cycle for DataONE
        • MATLAB/Stata support
      • Tentative release date: 12/15/2020
      • Remaining features t.b.d: Reproducible Runs

Updates

  • Tommy
    • More C2Metadata
      • I don’t think that the SDTL owl file is correct/has ever been used (see incorrect ‘range’ for lists in the owl file)
        • Issues where lists (mergesDataFrames:[”dframe1, dframe2”]) have no defined representation in RDF
        • Proposing a way to do this with Dan
          • I’m not really concerned with what SDTL looks like in RDF so I’m creating an object for each string, with type schema:Thing and schema:name “dframe1”
    • PR reviews
  • Craig:
  • Kacper
    • no updates this week
  • Mike H
  • Mike L
    • Wired up the version timeline in the Versions side panel in the UI
    • Expanded Client API to handle get / getRoot / list / delete, in addition to create
    • Filed a bug about delete case, which throws a stack trace: https://github.com/whole-tale/wt_versioning/issues/6