2019-04-10: Development Planning

Kacper, Craig

Planning

  • v0.7 testing/updates
    • Environment customization (repo2docker)
      • Reconcile with Yuvi’s latest proposal
      • Add bdbag dependency
      • Can we use the “Cloud” path method without patching Rstudio?
        • https://github.com/whole-tale/rstudio-base/blob/master/patches/custom_breadcrumbs.patch
    • Publish to DataONE
      • Styling changes, bug fixes (e.g., progress bar on import)
      • Multiple authors
        • PR to fix the creator with no ORCID problem
        • Dashboard change to render creator on card if no authors
    • Import from BagIt
      • https://github.com/whole-tale/girder_wholetale/pull/259
    • Bugs bugs bugs testing
      • Need rc2 to be built
      • https://github.com/whole-tale/girder_wt_data_manager/issues/39
      • Restart notification
      • Notification parity with Girder
    • Released by mid May?
  • v0.8+ planning
    • 3 party integration through oauth
      • We will use primarily for ORCID and possibly Agave or anything that requires authenticated access to data
    • UI refactor continued
      • Copy on Launch (PR issued)
      • Dataset citation PR issued
      • Compose > Create modal + select data widget
      • Run instance refactor to tale (ML)
      • Browse refactor (ML)
      • Notification handling refactor (ML)
      • Manage/catalog refactor (ML)
        • Note http://www.crc.nd.edu/~kfurse/wholetale-css-mock-up-sui_f2018/browse-catalog-meta.html
      • Killing all the dead code (ML)
      • Compute environments
    • Other UI
      • Globus logout
      • API keys support (for DVN and Zenodo publishing)
      • Changing the environment (remix)
    • Other already on plan
      • Migrate plugins from 2.x to 3.0
      • Usability test
      • Licenses
      • Terms of use
      • Publish to Dataverse
      • Accessing data with authentication (TDAR)
      • Import (from DataONE)
      • Remote zip files (WIP)
      • Kubernetes
      • User namespacing system (1.4.3) (overview)
      • Tale validation framework?
        • Run tale without home directory
      • Brown Dog integration (1.6.2)
      • iRODS = Agave ingest
      • Autodetecting changes
      • Tracking and storing provenance
        • Depends on prov task group input
      • TACC/Agave integration
        • Agave data
        • Remote execution w/ labels
  • For AEA/Odum outreach, what do we need?
    • Ability to run Stata, Matlab, SAS
    • Support for anonymous review (ACLs or tokens)
    • Publish to Dataverse
    • One-click run in Whole Tale (from DVN)
      • Unique URL for Tale – “whole tale instance” + PID
  • Things we want to/should do:
    • Image build infrastructure
    • Public registry + identifiers
    • Publish to Zenodo
    • One click (easy way to run from a single link)
    • Sharing/ACLS
    • Security audit/qualys scanning
    • Deploy to GCE/AWS
    • Git integration? Is there anything to be done?
    • UI for environment specification
    • Statistics/metrics
    • HPC and HTC Tales
      • Britton’s tale?
      • Zongyang’s tale?
      • ETK tale?
      • Support Condor + Pegasus + Parsl + Agave