search

BooDoo / ebooks_example

Heroku-ready turn-key _ebooks implementation
22 stars 16 forks source link

TODO: Support (externally hosted) corpus consumption #2

Open BooDoo opened 9 years ago

BooDoo commented 9 years ago

If user has a file they'd like to use to seed the corpus, and a place to host it in plaintext, we should support that!

BooDoo commented 9 years ago

This is coming along nicely using Cloudinary's free tier at Heroku as the file store. There's a web interface and we can push/pull from the app itself.

ETA: Monday Feb 23 or sooner.

therourke commented 9 years ago

Hello... I just started using your bot on the advice of Brett o'Connor. Wondering whether you ever deployed the externally hosted CSV for these bots? I'd love to use that as a feature.

(my bot is @zebrapedia by the way)

BooDoo commented 9 years ago

There is a branch (persist-cloudinary) that I used for my @0xB00D00 account for about a month. Here's the Deploy to Heroku link.

It misses caching Cloudinary assets often, which means if you have a large corpus you'll run into problems with it re-fetching too often and running well over your allotted bandwidth.

I can't recommend it as stable, but ¯_(ツ)_/¯

BooDoo commented 9 years ago

So for example I ended up getting an angry email from Cloudinary after I let a deployment with my twitter archive (~18MB) run for a month because I was at 469% of the allotted usage for a free account.

therourke commented 9 years ago

Thanks. Have tried this, but having trouble. The Cloudinary url is in this format?

http://res.cloudinary.com/USERNAME

I keep getting 'configure environment' errors like this: "growing-nimbly-1303 will overwrite CLOUDINARY_URL on pkdbotcloud"

therourke commented 9 years ago

Any idea what might be wrong with it? Am I missing something? Thanks

BooDoo commented 9 years ago

Sorry about the delay - you leave CLOUDINARY_URL blank, it's automatically populated by the Cloudinary add-on.