Winds is a beautiful open-source RSS and Podcast app created using React/Redux/Node. Use the free hosted version or run it on your own server and customize it as you see fit. Contributions are always appreciated. In fact, we're planning a series of tutorials to help you get started. Activity Feeds & Discovery in Winds are powered by Stream, the app leverages Algolia for search, AWS for hosting, MongoDB Atlas for a hosted database (DBaaS), and SendGrid for email. All of these services have a free tier.
To get started with Winds, please visit https://getstream.io/winds/
In addition to the desktop apps for macOS, Linux, and Windows, there's now a web version of Winds available at https://winds.getstream.io/
Stream is sponsoring 1 person to work full time on Winds. The job posting is available on AngelList.
Winds is packed full of awesome features behind a beautiful user interface and user experience. Here's a quick look at what the project has to offer:
Help us improve Winds and/or vote on the Roadmap for 2.1
- Search detail screen
- Playlist support (partially implemented)
- Team support (share an activity feed with colleagues or friends to discover and collaborate)
- Mobile application powered by React Native
- Folder Support
The full stack can be found on StackShare.io.
Stream is an API for building activity feeds. For Winds the follow suggestions and the list of articles from the feeds you follow is powered by Stream. Stream accounts are free for up to 3 million feed updates and handle personalization (machine learning) for up to 100 users.
Algolia is used for lightning fast and relevant search. We use their JavaScript search client to easily setup the Winds search implementation. Algolia, accounts are free up to 10k records and 100k operations.
MongoDB Atlas provides a Database as a Service, and serves as the backend datastore for Winds.
The following tutorials will not only help you start contributing to Winds, but also provide inspiration for your next app.
Note: We're actively working on this portion of the README. To stay up to date with the latest information, please signup for the hosted version at https://getstream.io/winds.
- Architecting a large React/Redux codebase
- Building activity streams and newsfeeds on top of Stream
- Implementing search with Algolia
- MongoDB database schema design
- Design thought process for Winds
- PM2 in production environments
- Creating a RESTful API design with Express.js
- Takeaways on Building a React Based App with Electron
- The Winds Stack
- Building Touch Bar Support for macOS
- Testing Node.js in 2018
To download Winds, visit https://getstream.io/winds/.
Commands:
brew install redis mongodb
brew services start mongodb
brew services start redis
cd winds/api
yarn
cd ../app
yarn
Sign up for both Stream and Algolia, and create the following .env
file in the app
directory, replacing the keys where indicated:
DATABASE_URI=mongodb://localhost/WINDS
CACHE_URI=redis://localhost:6379
JWT_SECRET=YOUR_JWT_SECRET
API_PORT=8080
REACT_APP_API_ENDPOINT=http://localhost:8080
STREAM_API_BASE_URL=https://windspersonalization.getstream.io/personalization/v1.0
STREAM_APP_ID=YOUR_STREAM_APP_ID
REACT_APP_STREAM_APP_ID=YOUR_STREAM_APP_ID #(this needs to be included twice, once for the backend, and once for the frontend to make realtime connections directly to Stream)
STREAM_API_KEY=YOUR_STREAM_API_KEY
STREAM_API_SECRET=YOUR_STREAM_API_SECRET
REACT_APP_ALGOLIA_APP_ID=YOUR_ALGOLIA_APP_ID
REACT_APP_ALGOLIA_SEARCH_KEY=YOUR_ALGOLIA_SEARCH_ONLY_API_KEY
ALGOLIA_WRITE_KEY=YOUR_ALGOLIA_ADMIN_API_KEY
MERCURY_KEY=YOUR_KEY_HERE
Note: If you are running the test suite, you will need to have a test version of the
.env
file inside of theapi/test
directory.
Then run:
pm2 start process_dev.json
cd app && yarn start
git clone git@github.com:GetStream/Winds.git
The following instructions are geared towards Mac users who can use brew
(Homebrew) to install most dependencies. Ubuntu users can use apt
, and Windows users will have to install directly from the dependency's site. Non-debian-based Linux users will probably be able to figure it out on their own :)
cd Winds/app
yarn
cd ../api
yarn
Winds uses MongoDB as the main datastore - it contains all users, rss feeds, podcasts, episodes, articles, and shares.
If you're on a Mac, you can install MongoDB through Homebrew by running:
brew install mongodb
(You can also install MongoDB from the official MongoDB site.)
You can also run MongoDB in the background by running:
brew services start mongodb
At Stream, we use Redis as an in-memory storage for the Winds podcast processing and RSS processing workers. It contains a list of podcasts and RSS feeds, which the workers pick up and process using the bull
messaging system.
If you're on a Mac, you can install Redis through Homebrew by running:
brew install redis
(You can also install Redis from the official Redis site.)
Then, start Redis by running:
redis-server
...which creates (by default) a dump.rdb
file in your current directory and stores a cached version of the database in that location.
You can also run Redis in the background by running:
brew services start redis
For testing purposes, you will want to use the test data located here.
Use mongoimport
or mongorestore
to import the data. There are two username and password combinations for testing:
Username: admin@admin.com
Password: admin
Username: test@test.com
Password: test
You will need to run the FLUSHALL
command in Redis to ensure that the new content is picked up.
Note: This will override any local data that you may have. Please be cautious! Also, this will not create Stream follows – please followo feeds manually to generate them.
To contribute to Winds, sign up for Stream to utilize the activity and timeline feeds.
(Reminder: Stream is free for applications with less than 3,000,000 updates per month.)
- Sign up for Stream here
- Create a new Stream app
- Find the App ID, API Key, and API Secret for your new Stream app
Append the Stream App ID, API Key, and API secret to your .env
file:
STREAM_APP_ID=YOUR_STREAM_APP_ID
STREAM_API_KEY=YOUR_STREAM_API_KEY
STREAM_API_SECRET=YOUR_STREAM_API_SECRET
Once you've signed in, create "feed groups" for your Stream app.
A "feed group" defines a certain type of feed within your application. Use the "Add Feed Group" button to create the following feeds:
feed group name | feed group type |
---|---|
podcast |
flat |
rss |
flat |
user |
flat |
timeline |
flat |
user_episode |
flat |
user_article |
flat |
It's fine to enable "Realtime Notifications" for each of these feed groups, though we won't be using those yet.
In addition to Stream, you also need to sign up for Algolia, to contribute to Winds, for the search functionality.
(Algolia is free for applications with up to 10k records.)
- Sign up for Algolia here
- From the Applications page, click "New Application" and create a new Algolia application. (We recommend something similar to
my-winds-app
)- (Select the datacenter that's closest to you.)
- From the application page, select "Indices" and click "Add New Index". (We recommend something similar to
winds-main-index
)
From your app, click the "API Keys" button on the left to find your app ID and API keys.
Append your Algolia application ID, search-only API Key and Admin API Key to your .env
file to look like this:
REACT_APP_ALGOLIA_APP_ID=YOUR_ALGOLIA_APP_ID
REACT_APP_ALGOLIA_SEARCH_KEY=YOUR_ALGOLIA_SEARCH_ONLY_API_KEY
ALGOLIA_WRITE_KEY=YOUR_ALGOLIA_ADMIN_API_KEY
From the root directory, run:
pm2 start process_dev.json
To see logging information for all processes, run:
pm2 logs
cd app && yarn start
Winds API server uses:
- Mocha as testing framework
- Chai as assertion library
- Sinon as mocking library
- nock as HTTP mocking library
- mock-require as module mocking library
Tests are located in api/test
folder.
File structure is intentionally mirroring files in api/src
to simplify matching tests to tested code.
To run tests:
cd api && yarn run test
To run tests with extended stack traces (useful when debugging async issues):
cd api && yarn run test_deep
Add your code to a file in api/test
folder (preferably mirroring existing file from api/src
folder).
Refer to Mocha documentation for examples of using BDD-like DSL for writing tests.
Modules are mocked in api/test-entry.js
as mocks have to be installed before any modules are loaded.
Fixtures are loaded via loadFixture
function from api/test/fixtures
folder
Various utility functions are provided in api/test/util.js
(See other tests for expamles of usage).
Build a production version of Winds by running:
./build.sh
This creates production-ready JS files in api/dist.
To run the production JS files:
pm2 start process_prod.json
Unfortunately there is no unified standard for RSS.
Go to the api
directory and run yarn link
to make these commands available:
winds rss https://techcrunch.com/feed/
For podcasts:
winds podcast https://www.npr.org/rss/podcast.php\?id\=510289
For Open Graph scraping
winds og http://www.planetary.org/multimedia/planetary-radio/show/2018/0509-amy-mainzer-neowise.html
RSS Discovery
winds discover mashable.com
Article parsing (using mercury)
winds article https://alexiskold.net/2018/04/12/meet-12-startups-from-techstars-nyc-winter-2018-program/
Pull requests for improved RSS compatibility are much appreciated.
Most of the parsing codebase is located in api/src/parsers/
.
All support is handled via GitHub Issues. If you're unfamiliar with creating an Issue on GitHub, please follow these instructions.
Thank you to all of the maintainers and contributors who've helped Winds become what it is today and help it stay up and running every day. We couldn't do it without you!
- Tommaso Barbugli
- Dwight Gunning
- Matt Gauger
- Max Klyga
- Zhomart Mukhamejanov
- Julian Xhokaxhiu
- Jonathon Belotti
- The Gitter Badger
- Meriadec Pillet
- Alex Sinnott
- Lawal Sauban
RSS is an amazing open standard. It is probably the most pleasant way to stay up to date with the sites and podcasts you care about. Our reasons for contributing to Winds are explained in the blogpost Winds 2.0 It's time to Revive RSS. In this section we will list other open source and commercial projects that are having an impact on Reviving RSS: