leftist-quotes

Update 2022-08-07: I recently redid this as part of a general push towards more self-hosted architectures. Instead of being hosted on Heroku, as this post describes, all requests to this API are answered by my home server, which is actually just an old laptop. The DNS configuration and SSL certificates are managed by Cloudflare. Despite wanting to completely self-host and no longer rely on Internet giants, they were too much of a headache to configure locally.

Intro#🠑

As a celebration for the end of summer (joke), I threw together a super, super simple API to interact with 778 communist/leftist quotes, all from marxists.org.

There are three endpoints:

/ (the root): returns a random quote

/daily : returns a single quote every day, matches with the marxists.org quote-a-day page.

/all-quotes : returns all quotes.

Quotes are returned as objects in the form:

• body: the text of the quote
• link: the link to a page on marxists.org with the associated reading
• attribution: the name of the author and often the date or title of the original text

The first two routes return a single object, and the last route returns an array of objects.

Source code is hosted at github.com/18kimn/leftist-quotes

Demo#🠑

on learning#🠑

This tool was built through express.js, with the data preprocessed through R and the entire site hosted on Heroku. It’s incredibly simple, so I just threw everything in the root directory instead of creating src/ or static/ folders and so on.

This is my first API that I’ve built, which was exciting and definitely a learning experience. Two basic those things I learned:

• backend architecture: how do we set up GET requests? Where does this deploy?
• Web architecture in general: what are DNS requests and forwarding?

The second point was especially confusing and tedious, as I had to reroute https://leftist-quotes.com, www.leftist-quotes.com, and so on all to the same root site.

The other thing I realized while scraping the marxists.org webpage was that the marxists.org website was built in the stone ages. Every single quote was placed inside of a document.write(<p>...</p>) statement, and the entire page was essentially a single switch statement that was nearly eight hundred clauses long. On one hand, that must have taken a ridiculously long time to make since every quote had to be manually coded and formatted. On the other hand, it’s not performant as the page has to load a few megabytes more than it needs and anyone editing it in the future would probably be extremely frustrated.

Next steps#🠑

I hope to add more quotes eventually and tag the API with the source I got these from. Marxists.org is great, but there’s a little too much Hegel and Trotsky in on their source for my taste. @rwgilmoregirls on Twitter and redfish media always tweet great things, so grabbing the alt text on those tweets could be one option.