Skip to content

Milestones

List view

  • Future Directions

    this milestone captures potential features or improvements to the application that are not related to current milestones but that we may want to tackle for a future version

    No due date
    •10/10 issues closed
    100% complete0 open10 closed

  • Added data Interoperability features

    Based on existing code, allow users to import a NDEx network, convert to our internal format, calculate, and store the embeddings and pre-trained training results for comparisons, and use this format to store metadata for nodes, edges, networks, and labels in a standard JSON format.

    No due date
    0% complete0 open0 closed

  • Per-user Geneset management

    • Allow users to manage uploaded genesets and jobs • For each run (i.e., multiple networks/parameters (runs) per geneset, and an index on pre-trained genesets), store the geneset in a local database in a new table, in addition to a table for results with parameters, linked to genesets. This redundancy allows for multiple results per user-supplied geneset • List genesets and link to results jobs run with statistics

    Overdue by 3 year(s)
    •
    Due by August 2, 2022
    •1/1 issues closed
    100% complete0 open1 closed

  • Incorporate user, data, and job management for enhanced useability into web server application

    • Allow sign-in with new user account. • Allow sign-ins with existing github accounts. Other logins will be new issues( for example ORCID, Gitlab, Google accounts) • Handling adding emails and notifications in form • Store jobs per-user when uses are logged in. • Store results per-user when the job is completed • Allow logged-in users to "star" pre-trained results to sort by popularity. • Generate per-user statistics. • Allow viewing pre-trained results or running jobs without accounts.

    Overdue by 3 year(s)
    •
    Due by July 29, 2022
    •10/10 issues closed
    100% complete0 open10 closed

  • Improved implementation of machine learning batch system

    The GPWeb v1 implementation was had several limitations including requiring manual deletion of cloud This requires a new cloud implementation to provision compute resource on demand run GPML v2 software via website click. This deliverable is for improvements to address limitations including Goals: • simplified cloud deployment scripts and procedures • re-design job compute architecture: o maintain low-cost, on-demand nature of job system o simplified in nature and deployment o robust, concurrent jobs (scalable) o remove current limitations (container instance quota) • automated management of cloud compute resources • incorporation of GPML v2 package into containerized GPML job running system

    Overdue by 3 year(s)
    •
    Due by July 25, 2022
    •2/2 issues closed
    100% complete0 open2 closed