May 9, 2017, Tuesday

• Ongoing

  • CyVerse Metadata hackathon
    • Using COPO for entering metadata?
  • Authentication issue (change config.yml to 72 hrs)
    • Looks like user has to refresh browser to show they have been logged out
    • If don't, they can not save workflow, launch job? ...
    • To make it possible for user to construct workflow with long running jobs, we probably need to make it logged in as long as possible: 72 hours? We still need to log them out automatically to make things less confusing (looks like they can still launch job but history will stay gray)
  • Add output description from Agave to display below diagram when clicked
  • Workflow name needs to be unique (popup window for error if user enter a non unique name)
  • A popup window is shown when job is submitted (Job submitted successfully or failed!)
  • Support job history checking for debugging
  • Browse Data Store will default to public (to user if log in? and change user to home?)
  • Can only save a workflow that is newly built (can not save running workflow or loaded workflow)
    • Need to (temporarily) disable 'Save' for running workflow
      • Because a running workflow will be saved as 'my_workflow' (default), when loaded, it will not load job history, not show job status on diagram. This could be related to the building workflow bug
      • Ideally, we should allow saving a running workflow? Output folders are not yet built for un-submitted jobs!
      • Should we move 'workflow name' and 'workflow description' to a popup window when user click on save?
  • Working on setting archive as false (Postponed)

• Optimization

  • Need to support direct link, which will resolve the tutorial link above!
    • Not completely done, direct link works (workflow not check), can not utilize forward/backward browser button
    • Might need to split App component to leftbar, main, and rightbar (hold on for now)
  • Database rebuilt after updates
    • Need instructions to dump and reload old contents (if possible, depends on specific schema modifications)
  • Load workflow
    • Need to clear out current history panel to make sure steps are consistent between workflow forms and job histories
    • Should warn user that history panel will be cleared out and give them option to save it before loading
  • CSS modifications

• TACC Workflow

  • Use RNA-seq workflow as an example
  • All apps run at TACC Stampede
  • For job and workflow, outputs will be saved to sci_data
    • Working on setting archive as false and move data manually with agave import
    • Need to work with Tony to make sure sci_data is automatically linked to TACC's resource server
    • If not, set up a storage system (data.cyverse-cshl.org) at TACC (similar with data.sciapps.org)

• SciApps Web API

  • Code is already in place
  • Just need documentation and systematic design
  • Might need to add new API end points depending on community inputs
    • Take the JBrowse-Galaxy use case to see whether we need to add new end points