Tuber needs a database, a session store, and a job store as well as a web server. For testing and development you can use a single sqlite database as the database, session store and job store while using the build-in python and node webservers for a very simple deployment. Larger systems will benefit greatly from using a dedicated database instance (generally postgres), a redis server for sessions and jobs, and nginx or apache as a webserver.
There are numerous ways to configure things, but we try to make the most common was as simple as possible to implement.
The latest version of Tuber is published to GitHub Packages as ghcr.io/magfest/tuber-frontend:latest and ghcr.io/magfest/tuber-backend:latest. You can view them here.
To deploy using docker first install docker on your platform, as described here.
With the docker daemon running, you can now pull and run tuber:
docker-compose upThis will set up a small production-style stack of containers, using postgres for the database, nginx as a reverse proxy, and redis as the session and job store. Once it finishes starting you should be able to access your instance at http://localhost:8081
Note: The sample docker-compose file does not currently configure SSL. You should either set up a reverse proxy to handle SSL, or edit contrib/nginx.conf to use your certificates and edit docker-compose.yml to allow access to port 443.
After cloning this repository you will need the following dependencies:
dnf install npm python3 python3-devel python3-pip # Fedora/RHEL/CentOS
apt install npm python3 python3-dev python3-pip # Debian/Ubuntu
brew install npm python postgresql # MacOSOn Windows you'll have to install nodejs, Python3 and postgreSQL. Make sure to add both npm and python to your PATH during installation.
Once the dependencies are installed you can start up the backend and frontend development servers:
Tuber uses environment variables to configure some basic settings.
DATABASE_URL=sqlite:///database.db
FLASK_ENV=production
REDIS_URL=
WORKERS=2
CIRCUITBREAKER_THREADS=2
CIRCUITBREAKER_TIMEOUT=5
ENABLE_CIRCUITBREAKER=truepython -m venv venv
source venv/bin/activate
cd backend
python -m pip install -e .
tuber
# Windows
python -m venv venv
venv\Scripts\activate.bat
cd backend
python -m pip install -e .
..\venv\Scripts\tuber.exeThe server should now start up and begin listening on port 8080 for API requests.
In a separate terminal from the backend, install and serve the vue frontend:
npm install --global yarn # Yarn is recommended for the frontend
cd frontend
yarn install
yarn run serveThis will start the frontend on port 8081. You can connect your browser to http://localhost:8081 and complete the initial setup page to begin using tuber.
Both the frontend and backend will hot-reload as you change code.
If you want to create a new table or modify an existing one you will need to create an alembic migration. Most of the time, you can do this by autogenerating it.
First, create the table definition in tuber/models/.py, and make sure it is imported in tuber/models/init.py.
Next, use alembic to create the migration file:
venv/bin/alembic -c backend/tuber/alembic.ini revision --autogenerate -m "Added widget column to the whatsit table"This should create a migration file in migrations/versions. Read through it and adjust the steps as necessary. The next time you restart your dev instance it will run the migration.
You can also trigger the database update manually:
venv/bin/alembic -c backend/tuber/alembic.ini upgrade headMake sure to commit the migration along with the code that uses it!
If you receive the following ambiguous error message while running python setup.py develop: ld: library not found for -lssl
Try setting the link path for openssl and running it again: export LDFLAGS="-L/usr/local/opt/openssl/lib"
Sometimes when merging a branch that has its own new migrations into your own branch you'll have to tell alembic what to do. If you see alembic complaining about multiple heads check here: https://blog.jerrycodes.com/multiple-heads-in-alembic-migrations/