2018-10-22: Development meeting

Kacper, Adam, Bryce, Mihael, Tommy, Craig

Agenda

  • Updates
  • Discussing UX report:
    • KK: Do we have consensus on non-UI issues
      • Copy on launch
        • KK yes; MH yes; Bryce yes; Craig yes; TT yes; AB yes
        • Not present: SW, IT, DP
        • Questions:
          • TT: Are we copying the data. KK yes, it’s a “cp -r”
          • MH: It’s the equivalent of git checkout
      • Are we following repo2docker model? i.e. there are well-specified files to define environment changes
        • KK yes; MH yes; Bryce yes; TT yes; AB yes; Craig yes
      • The data directory is at the same level of the workspaces
        • MH: yes; KK yes; TT yes;
        • Bryce: not sure; AB not sure; CW not sure
        • Comments:
          • MH: Workspace contains code and configuration files
            • I’m in favor of input/output/scratch
          • MH: Wants to formalize the model a bit “on paper”
  • Follow-up call with Kacper, Craig, Tommy:
    • CW: Today a DataONE user could use WT to analyze data
      • “Analyze in Whole Tale” v “Launch in WT”
      • TT: “Launch in WT” with no tale
        • Register, create tale, launch it
        • Matt doesn’t want the button into Production until we have a sustainability plan
      • CW: What does the change look like?
      • CW: How does it roll out to member nodes?
        • TT: member nodes need to update metacat UI
        • Flag to show button will be off, but could be enabled
      • CW: How hard is it to get a dev stack up and running for a single member node?
      • TT:
        • Need mockup
        • Small loading screen with spinner
        • We already do some of this (spinner from compose)
        • KK: Can we use the job model? Endpoint uses Job and route uses Job status?
          • We can just run Jobs in Girder (e.g., no gwvolman)
          • TT: Registration system uses notification stream
        • KK: We can pipe the notification to the job status itself. The job has a status – poll it and display progress.
          • Endpoint gives you a job model
        • TT:
          • Can reuse progress bar, but will need to come up with a scheme for calculating progress.

Updates

  • Adam:
    • No updates from ND
  • Bryce:
    • No updates
  • Craig:
    • No updates for dev other than UI/UX review
    • Planning related to tasks that are not bound to new UI
    • https://github.com/whole-tale/whole-tale/issues/43
  • Kacper:
    • Reviewed Mikes code, added some comments
    • Attempted to write and run some tests, but ran out of time when spinning up new environemnt (next)
    • A PR/merge will be sent in asap
    • Would like to add functionality to ingest dataverse data
  • Mihael: *
  • Tommy:
    • Fixed a bug where packages were getting incorrect permissions (the orcid should not have been https)
    • Added an introjs to the publish modal
    • Transitioned (and added) some tasks to zenhub
    • Working on Tale Importing documentation (UI mockups are first, then backend tasks (eg what do we save, etc))
    • On the Stack:
      • More publishing testing
      • Send in DataONE branch with “Launch in Whole Tale” button
      • Commit the Tale import route
      • Fix bug in publishing modal where package URL only shows the identifier/pid