2020-12-21: Development Meeting

Kacper, Tommy, Mike H, Craig, Tim

Agenda

• Updates
• Image versions [name=Craig]
  • New Matlab IDE will only work with version >=R2020b
  • Jupyter Kernel/Xpra required for < R2020b
• Next meeting: 01/11/21

Updates

• Kacper
  • PR reviews
  • Fixed mran handling in r2d_wt buildpacks
  • Minor UI “improvement”
• Craig
  • Fixed r2d PR w.r.t. tests
  • PR reviews
  • Matlab redux
• Tommy
  • Misc PR reviews
  • Refactored SDTL/RDF stuff to address feedback from SDTL Team & Tim
  • Focusing some time on getting the current release out (PR review + docs + testing effors)
    • Would like to start work on some reproducible run stuff soon (SDTL pipleline depends on it and I only have ~2 months left)
• Tim
  • Order of SDTL elements do matter! Thanks, Tommy.
  • Fortunately, SPARQL queries still can be made to work.
  • Unfortunately, it looks like sequence is insufficient to resolve variable dependencies. Hope I’m wrong.
  • Will need to make a table of general SDTL program structures that can be expected and the type capabilities/limitations on querying for each.
  • https://github.com/ThomasThelen/sdtl-rdf-examples/blob/master/examples/03-compute_save_compute/script.py
  • From Craig:
    • https://gitlab.com/craig-willis/c2metadata-bd/-/tree/master/examples
• Mike H
• Mike L
  • Updated Tale Versions PR to work with newest versioning API changes
  • Noting Craig/Tommy’s feedback about the “name exists” validation behavior - not sure how easy it will be to intercept the modal before it submits or reopening it on failure. This may be a shortcoming in Angular/Google’s MatDialog.