scribe/README.md
2022-01-23 16:08:23 -05:00

5.8 KiB

Scribe - An Alternative Medium Frontend

This is a project written using Lucky. It's main website is scribe.rip.

Deploying Your Own

I'd love it if you deployed your own version of this app! A few others have already. To do so currently will take some knowledge of how a webserver runs. This app is built with the Lucky framework and there are a bunch of different ways to deploy. The main instance runs on Ubuntu but there are also directions for Heroku or Dokku.

One thing to note is that this app doesn't currently use a database. Any instructions around postgres can be safely ignored. However, Lucky (and it's dependency Avram) do require a DATABASE_URL formatted for postgres. It doesn't need to be the URL of an actual database server though. Here's mine: DATABASE_URL=postgres://does@not/mater

Hopefully a more comprehensive guide will be written at some point, but for now feel free to reach out to the mailing list if you have any questions.

Docker (Unsupported)

A Dockerfile is included to build and run your own OCI images. I don't use Docker personally so this is all community created and supported. If it breaks, please write to the mailing list.

To build:

$ docker build [--build-arg PUID=1000] [--build-arg PGID=1000] -t scribe:latest -f ./Dockerfile .

To run (generating a base config from environment variables):

$ docker run -it --rm -p 8080:8080 -e SCRIBE_PORT=8080 -e SCRIBE_HOST=0.0.0.0 -e SCRIBE_DB=postgres://does@not/matter scribe:latest

To run with mounted config from local fs:

$ docker run -it --rm -v `pwd`/config/watch.yml:/app/config/watch.yml -p 8080:8080 scribe:latest

Configuration

To allow your domain to show up on the homepage, the APP_DOMAIN environment variable must be set. Note that this only takes effect if the LUCKY_ENV environment variable is also set to production.

See the route_helper config for the code that powers this feature.

Other configuration needed when in production mode:

  • PORT: The port Scribe should run on
  • SECRET_KEY_BASE: A 32-bit string. Can be generated with lucky gen.secret_key
  • DATABASE_URL: May be any valid postgres url since Scribe doesn't use a database
    • Example: postgres://does@not/matter
  • GITHUB_PERSONAL_ACCESS_TOKEN: to proxy gists with authenticated GitHub API requests
  • GITHUB_USERNAME: to proxy gists with authenticated GitHub API requests

GitHub Gist Proxying

Scribe proxies GitHub gists. It does this by making API requests to GitHub's ReST API to get the gist file contents. GitHub limits API requests to 5000/hour with a valid api token and 60/hour without. 60 is pretty tight for the usage that scribe.rip gets, but 5000 is reasonable most of the time.

API credentials are in the form of a GitHub username and a personal access token attached to that username. To get a token, visit https://github.com/settings/tokens and create a new token. The only permission it needs is gist.

This token is set via the GITHUB_PERSONAL_ACCESS_TOKEN environment variable. The username also needs to be set via GITHUB_USERNAME. When developing locally, these can both be set in the .env file. Authentication is probably not necessary locally, but it's there if you want to test. If either token is missing, unauthenticated requests are made.

Project goals

I believe that Medium is a bad actor on the web. They offer a bad reading experience. Writing there benefits Medium more than the author. Counter to their promise of a wider reach, they offer worse SEO. They use extortionist business tactics. Finally, they want to centralize the currently decentralized world of blogging.

Since Scribe uses Medium content, I don't want to help people engage with it more than they must. My goal here is not to make a nicer Medium to engage with, but to make a less bad experience when people are forced to engage with it. I want Scribe to be a tool, not a platform.

It's intentional that there is no way to browse content from a user, see popular posts, consume via an RSS feed, or further engage with an article via comments or "claps". I want to spend my time encouraging writers to move to worthy platforms, not making a bad platform worthy.

Contributing

  1. Install required dependencies (see sub-sections below)
  2. Run script/setup
  3. Run lucky dev to start the app
  4. Send a patch to ~edwardloveall/Scribe@lists.sr.ht (it may not look like it at first, but that's an email address).

Installing dependencies

General instructions for installing Lucky and its dependencies can be found at https://luckyframework.org/guides/getting-started/installing#install-required-dependencies.

Installing dependencies with Nix

If you are using the Nix package manager, you can get a shell with all dependencies with the following command(s):

nix-shell

# Or if you are using the (still experimental) Nix Flakes feature
nix flake update # Update dependencies (optional)
nix develop

Learning Lucky

Lucky uses the Crystal programming language. You can learn about Lucky from the Lucky Guides.