bign8 / chive-show

theChive wallboard

Hi there 👋

chive-show's Issues

Board: Move loading indicator to be a custom element

The loading indicator is pretty unique and might be better implemented as a self contained component.

Links

• https://micro-frontends.org/

Fix Logging

After looking at the last cron run, it appears logging isn't making it into appengine 😢

Here is the blog post for setting it up: https://cloud.google.com/logging/docs/setup/go

The package docs: https://pkg.go.dev/cloud.google.com/go/logging

Also, side note that the go libraries I'm currently using are for the "flexable" runtime, while the deployements are the "standard" runtime. Personally, I like the standard library's SDK a little better, maybe this should be a driving factor to port the package back to that. nvm, see #11

TODO: figure out which direction APPENGINE is actually going and report back.

Virtualized Content Pane

Currently the infinite scroller is... infinite, meaning eventually, the browser will run out of vertical space. See if we can virtualize the posts scrolling around. Keeping the previous, current and next post elements in the DOM. Note, having image dimensions pre-fetched may be fairly beneficial when configuring this.

GHA - cron to archive un-used staging deployments

Sometimes, we run out of active versions on the app-engine project and we can't create new staging deploys. It doesn't happen often, but when it does it's pretty annoying.

https://console.cloud.google.com/appengine/versions?project=crucial-alpha-706

It would be kinda cool if we could run a task on some fairly infrequent frequency, to archive old and unused app-engine versions.

[Security] Potential Secret Leak

It has been noticed that while using codecov/[email protected] your token 5c59aefa-2f68-41fe-b438-8c0ec2918830 is present in plaintext. Please ensure that secrets are encrypted or not passed as plain text in github workflows.

Server: Simplify and Cover

Add the appropriate level of abstraction to enable unit testing and increase coverage in the following areas. This issue is a placeholder that can be broken into further components as necessary.

• UI
• API
• Cron (this guy might have some hints on logical separations #2)

Percentile based tag culling

Culling the list of tags to send to the client is currently by a fixed number. As the amount of posts grows, this can grow stale quickly. Instead, use the 75% percentile or a half average (the mathematician in me is cringing right now) as the lower bound instead of 3. An example of this can be found: #2

While we are culling this list, we might try and remove the tags that the chive folks use to control their system. They appear to use the full tag to denote a post is full of images, and the *staple tags to pin posts to the home page. I haven't noticed many tags that are all lower-case that would want to be shared. Anyway, do some noodling on this, either a blacklist or a heuristic function would be sufficient. Maybe store the tags always and just cull when returning posts and tags from the store so we can change the logic later and not have to migrate the data.

Improve Tag Saving Performance

With the tags feature, it appears that saving tags could use some improvement.
I suggest starting a SINGLE transaction; using a GetMulti for ALL the keys involved, and a PutMulti once the counts have been updated.

Tag Selector

Add the ability to select from all the tags provided by the /api/v1/tags endpoint. Possibly put a hovering button in the lower right (similar to gmail compose button) that opens a menu or something. Some mock ups before contributing would be nice.

Gradual Asset Loading

Assets are currently loaded all at once. Currently doesn't really matter on reasonable hardware, but it would be good to reduce the load on theChives servers.

Add deployment tests to appengine

Problem	Automatic deployments can really damage the datastore
Solution	Build structure necessary to ensure production datastore cleanliness
Status	Staging
Build

TODO

• Ensure proper DB versioning + prevent db migration errors
  • Wrap request contexts with a deployed version identifier as a key namespace
• Tap a DB rollover script
  • Update db with current version from older namespaces
  • Allow incremental (cron based) updates to ensure not exceeding GAE limits
  • Make api executable so integration tests can run on them

Another possibility is to abstract datastore and not write, but softly persist modifying changes to development stores

Mobile: do not repeat posts

Use cursor based paging on the "random" endpoint. This logic could generate a unique shuffling seed if no cursor is provided, and re-use the seed for subsequent shuffles. Additionally logic will need to be added to include "offset" and "limit" parameters encoded in this "cursor". The following form should be simple enough to implement: <offset>:<limit>:<seed>.

Links

• https://www.moesif.com/blog/technical/api-design/REST-API-Design-Filtering-Sorting-and-Pagination/

Move Cron to B1 Instance

Why expose it on the frontend? Old version of this can be found in #2

Add deployment tests to appengine

Problem	Continuous Deployment is a neat idea
Solution	Use Travis CI to help automate deployment
Status	Staging
Build

TODO

• Deploy to App Engine as release candidate
• Run deploy tests on App Engine release candidate
  • API request /api/v1/random
  • Index serving
  • Static assets (referenced by index) serving
  • Crawl site (ensure no 404s)

GHA - update to use latest versions of go

A combination of dependabot and/or actions/setup-go changes might be needed here to get this working, but this should help unblock #57