If a vocab refresh has happened, and if the app is currently re-building the networkx graph, TermHub may possibly be unusable while the new graph is being generated.
If so, we should set a manage var that says it's in progress, and whenever TermHub fetches on the backend, there needs to be a hook that checks this and returns a message to the user letting them know that the app is on pause until it completes.
To consider
[ ] Notification to user that update is occurring
[ ] Handle multiple requests (only allow 1 initiation; don't interrupt refresh of the pickle on another GET)
Workflowy notes dump
Options:
- a. Notification
- Let the user know that the refresh is happening and to come back in a few minutes
- b. Check independent of hitting the home route
- Some kind of cronjob process that's set up on the server that checks early in the morning, after a refresh would have been completed, or perhaps checks once an hour, and sees if there's been an update, and rebuilds the graph
- This probably requires or is best done with setting up a separate worker
- But that is also time consuming to set up since we haven't done that yet
Why this is low priority
This requires a lot of work / async requests
This happens rarely, like every 6 months
We likely won't have people actively using when this happens
Overview
If a vocab refresh has happened, and if the app is currently re-building the
networkxgraph, TermHub may possibly be unusable while the new graph is being generated.If so, we should set a
managevar that says it's in progress, and whenever TermHub fetches on the backend, there needs to be a hook that checks this and returns a message to the user letting them know that the app is on pause until it completes.To consider
Workflowy notes dump
Options: - a. Notification - Let the user know that the refresh is happening and to come back in a few minutes - b. Check independent of hitting the home route - Some kind of cronjob process that's set up on the server that checks early in the morning, after a refresh would have been completed, or perhaps checks once an hour, and sees if there's been an update, and rebuilds the graph - This probably requires or is best done with setting up a separate worker - But that is also time consuming to set up since we haven't done that yet
Why this is low priority