Skip to content
New issue

Caching notebook results on a server

No due date 0% complete

We want to cache notebook results so users can share computational workload. This works by

  1. Specifying a git repo
  2. Checking central repo for most recent notebook; confirming with user if they'd like to use
    this notebook
  3. Cloning the git repo if the user would prefer to generate a new notebook
  4. Pushing the new notebook to the central git repo via a Gateway AP…

We want to cache notebook results so users can share computational workload. This works by

  1. Specifying a git repo
  2. Checking central repo for most recent notebook; confirming with user if they'd like to use
    this notebook
  3. Cloning the git repo if the user would prefer to generate a new notebook
  4. Pushing the new notebook to the central git repo via a Gateway API which makes sure users
    aren't up to shenanigans. This seems like a better approach then giving a bunch of people we don't know write access to a shared git repo. Prevents hard pushes, GitHub playing hardball, etc.
Loading