Social reading and reviewing, decentralized with ActivityPub
Find a file
2020-05-10 16:45:03 -07:00
fedireads Re-thinks connector mappings 2020-05-10 16:45:03 -07:00
fr_celery First draft async imports. 2020-04-22 12:43:55 +01:00
.env.example Move to redis and fix a bunch of things 2020-03-31 16:31:33 -07:00
.gitignore Only .gitignore base image dir (not static/images) 2020-02-19 19:19:20 -08:00
CODE_OF_CONDUCT.md Create CODE_OF_CONDUCT.md 2020-02-19 00:41:20 -08:00
docker-compose.yml Set flower broker 2020-03-31 16:50:58 -07:00
Dockerfile Move to redis and fix a bunch of things 2020-03-31 16:31:33 -07:00
fr-dev Add a command to fr-dev to restart celery. 2020-04-22 12:43:55 +01:00
init_db.py Removes default books from initdb 2020-05-10 15:18:40 -07:00
LICENSE Create LICENSE 2020-01-28 12:45:31 -08:00
manage.py move some configuration into environment vars 2020-02-15 13:45:44 -08:00
README.md Update README.md 2020-04-24 12:19:19 -07:00
rebuilddb.sh Create initial example db from a separate python file. 2020-02-24 16:25:47 +00:00
requirements.txt Move to redis and fix a bunch of things 2020-03-31 16:31:33 -07:00

Book Wyrm (aka FediReads)

Social reading and reviewing, decentralized with ActivityPub

Contents

The overall idea

What it is and isn't

FediReads is a platform for social reading! You can use it to track what you're reading, review books, and follow your friends. It isn't primarily meant for cataloguing or as a datasource for books, but it does do both of those things to some degree.

The role of federation

FediReads is built on ActivityPub. With ActivityPub, it inter-operates with different instances of FediReads, and other ActivityPub compliant services, like Mastodon and Pixelfed. This means you can run an instance for your book club, and still follow your friend who posts on a server devoted to 20th century Russian speculative fiction. It also means that your friend on mastodon can read and comment on a book review that you post on your Fedireads instance.

Federation makes it possible to have small, self-determining communities, in contrast to the monolithic service you find on GoodReads or Twitter. An instance can be focused on a particular type of literature, be just for use by people who are in a book club together, or anything else that brings people together. Each community can choose which other instances they want to federate with, and moderate and run their community autonomously. Check out https://runyourown.social/ to get a sense of the philosophy and logistics behind small, high-trust social networks.

Features

Since the project is still in its early stages, not everything here is fully implemented. There is plenty of room for suggestions and ideas. Open an issue to get the conversation going!

  • Posting about books
    • Compose reviews, with or without ratings, which are aggregated in the book page
    • Compose other kinds of statuses about books, such as:
    • Comments on a book
    • Quotes or excerpts
    • Recommenations of other books
    • Reply to statuses
    • Aggregate reviews of a book across connected FediReads instances
    • Differentiate local and federated reviews and rating
  • Track reading activity
    • Shelve books on default "to-read," "currently reading," and "read" shelves
    • Create custom shleves
    • Store started reading/finished reading dates
    • Update followers about reading activity (optionally, and with granular privacy controls)
  • Federation with ActivityPub
    • Broadcast and receive user statuses and activity
    • Broadcast copies of books that can be used as canonical data sources
    • Identify shared books across instances and aggregate related content
    • Follow and interact with users across FediReads instances
    • Inter-operate with non-FediReads ActivityPub services
  • Granular privacy controls
    • Local-only, followers-only, and public posting
    • Option for users to manually approve followers
    • Allow blocking and flagging for moderation
    • Control which instances you want to federate with

Setting up the developer environment

For most testing, you'll want to use ngrok. Remember to set the DOMAIN in .env to your ngrok domain.

Without Docker

You will need postgres installed and running on your computer.

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt
createdb fedireads

Create the psql user in psql fedireads:

CREATE ROLE fedireads WITH LOGIN PASSWORD 'fedireads';
GRANT ALL PRIVILEGES ON DATABASE fedireads TO fedireads;

Initialize the database (or, more specifically, delete the existing database, run migrations, and start fresh):

./rebuilddb.sh

This creates two users, mouse with password password123 and rat with password ratword.

And go to the app at localhost:8000

With Docker

You'll have to install the Docker and docker-compose:

docker-compose build
docker-compose up
docker-compose exec web python manage.py migrate
docker-compose exec web python manage.py shell -c 'import init_db'

Project structure

All the url routing is in fedireads/urls.py. This includes the application views (your home page, user page, book page, etc), application endpoints (things that happen when you click buttons), and federation api endpoints (inboxes, outboxes, webfinger, etc).

The application views and actions are in fedireads/views.py. The internal actions call api handlers which deal with federating content. Outgoing messages (any action done by a user that is federated out), as well as outboxes, live in fedireads/outgoing.py, and all handlers for incoming messages, as well as inboxes and webfinger, live in fedireads/incoming.py. Connection to openlibrary.org to get book data is handled in fedireads/connectors/openlibrary.py. ActivityPub serialization is handled in the activitypub/ directory.

Celery is used for background tasks, which includes receiving incoming ActivityPub activities, ActivityPub broadcasting, and external data import.

The UI is all django templates because that is the default. You can replace it with a complex javascript framework over my dead body mild objections.

Book data

The application is set up to get book data from arbitrary outside sources -- right now, it's only able to connect to OpenLibrary, but other connectors could be written. By default, a book is non-canonical copy of an OpenLibrary book, and will be updated with OpenLibrary if the data there changes. However, a book can edited and decoupled from its original data source, or added locally with no external data source.

There are three concepts in the book data model:

  • Book, an abstract, high-level concept that could mean either a Work or an Edition. No data is saved as a Book, it serves as shared model for Work and Edition
  • Work, the theoretical umbrella concept of a book that encompasses every edition of the book, and
  • Edition, a concrete, actually published version of a book

Whenever a user interacts with a book, they are interacting with a specific edition. Every work has a default edition, but the user can select other editions. Reviews aggregated for all editions of a work when you view an edition's page.

Contributing

There are many ways you can contribute to this project! You are welcome and encouraged to create or contribute an issue to report a bug, request a feature, make a usability suggestion, or express a nebulous desire.

If you'd like to add to the codebase, that's super rad and you should do it! At this point, there isn't a formalized process, but you can take a look at the open issues, or contact me directly and chat about it.