A Manuscript Integration for Twitter

Hey there! You've landed on the page for the Manuscript/Twitter integration. Feel free to look around!

How Does This Thing Work?

Here's a high level overview. This app talks to three three different APIs - the Manuscript API (via the manuscript-api npm package, the Twitter API (via the twitter npm package, and a MongoDB instance (via the mongodb npm package).

Each time we hit the /publish endpoint, we pull a subset of the accounts, iterate over them, and publish any new tweets matching the configured search. That process looks something like this.

1. Check to make sure the accounts got a configured query and a working Manuscript API token (and bail if not)
2. Run the query against the Twitter API
3. If response comes back with next_results key, store the tweets, parse the query Twitter gives us as the next_results, and run that search. Keep doing that until either Twitter rate limits us or we get to the last page, and then we'll return the tweets.
4. If Twitter has rate limted us, back off for 20 minutes and create a case in the manuscript account letting suggesting the account holder reconfigure search (this will mean we've sent something like 15,000 tweets their way in the last 15 minutes)
5. If Twitter sent us an authentication error, delete the account's Twitter token and secret

If you're interested in the nitty gritty, we've tried to comment the code extensively.

Get Help

Have questions? Ask for help! If you are working on this project in Glitch, you can highlight any line of code that you'd like another set of eyes on. You'll see a "handraise" icon appear near the line numbers. Click that, and you can ask a question of the Glitch community. It just so happens that the creators of Manuscript are the creators of Glitch, so there's a pretty good chance you'll end up with someone who knows a little something about working with the Manuscript API.

Made by Fog Creek

\ ゜o゜)ノ