May 2, 2017, Tuesday

• New items

  • CyVerse Metadata hackathon
    • How is UK passing metadata of input to output?
    • API call to retrieve metadata without authentication (like Gramene, build our own metadata database)?
    • Example workflows for Association and Annotation
      • Can we add metadata to describe what tool will be used to visualize the outputs (JBrowser, Shiny then Manhattan-Plot or RNAseq)? The idea will be to generate the visualization link directly from the backend
      • How do we pass taxon_id from inputs to outputs
      • other metadata?
  • Authentication issue
    • 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)
  • Rename workflow description
  • 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?

• Old items

  • Working on setting archive as false (Postponed)
  • Need to figure out why PCA is not running (Resolved, somehow this private app is set to run on 01, the public system, also PCA.R has access limitation for running as a private app)
    • jobs-history 3438421142804753945-242ac115-0001-007
    • Job accepted and queued for submission.
    • Input staging is current paused waiting for a system containing input data to become available. If the system becomes available again within 7 days, this job will resume staging. After 7 days it will be killed.

• 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