You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I like that idea, and I even looked briefly at doing this originally, but I was working on a time limit.
The main hurdle involved in implementing this is that the Tumblr API only allows filtering by tag (exact match), so we'd need to implement the following:
Enumerate all posts on first startup (Tumblr's API returns max 50 per call)
Store data in an external cache like Redis (Heroku's has a free tier that's almost certainly good enough).
Occasionally poll for new posts (Tumblr doesn't support webhooks)
Add text matching goodness
Ultimately, this would make the bot better, as it'd respond a lot more quickly (no lag for calling Tumblr's API), and there'd be no danger of large Slack teams hitting Tumblr's API rate limit (though 1k/hour and 5k/day won't be a problem for most).
So I imagine an easier way of going about this will be to make use of the Swear Trek (and Effin' Birds) database that is under production and access is to patreon supporters w/ $1/mo or more. Presumably pass any access tokens into environment variables for direct use, which would eliminate the need to directly use Tumblr's API, store a cached index of posts, and polling.
Instead of a random GIF, how about one that matches text?
This would preferably be in the style of the
/giphy <search text>
followed by shuffle/select/cancel behavior with the Giphy slash command.The text was updated successfully, but these errors were encountered: