Naked blogging platform.
We have a mailing list at ~sirodoht/mataroa-community@lists.sr.ht for the mataroa community to introduce themselves, their blogs, and discuss anything that’s on their mind!
Archives at lists.sr.ht/~sirodoht/mataroa-community
Victor Freire has created mata, a CLI tool for Mataroa.
Feel free to open a PR on GitHub or send an email patch to ~sirodoht/public-inbox@lists.sr.ht.
On how to contribute using email patches see git-send-email.io.
Also checkout our docs on:
- Coding Conventions
- Git Commit Message Guidelines
- File Structure Walkthrough
- Dependencies
- Server Playbook
- Admin and Moderation
This is a Django codebase. Check out the Django docs for general technical documentation.
The Django project is mataroa
. There is one Django app,
main
, with all business logic. Application CLI commands are generally
divided into two categories, those under python manage.py
and those under
make
.
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements_dev.txt
pip install -r requirements.txt
A file named .envrc
is used to define the environment variables required for
this project to function. One can either export it directly or use
direnv. There is an example environment
file one can copy as base:
cp .envrc.example .envrc
.envrc
should contain the following variables:
export DEBUG=1
export SECRET_KEY=some-secret-key
export DATABASE_URL=postgres://mataroa:db-password@db:5432/mataroa
export EMAIL_HOST_USER=smtp-user
export EMAIL_HOST_PASSWORD=smtp-password
When on production, also include/update the following variables (see Deployment and Backup):
export DEBUG=0
export PGPASSWORD=db-password
This project is using one PostreSQL database for persistence.
One can use the make pginit
command to initialise a database in the
postgres-data/
directory.
After setting the DATABASE_URL
(see above), create
the database schema with:
python manage.py migrate
Initialising the database with some sample development data is possible with:
python manage.py loaddata dev-data
dev-data
is defined inmain/fixtures/dev-data.json
- Credentials of the fixtured user are
admin
/admin
.
To develop locally with subdomains, one needs something like this in their
/etc/hosts
:
127.0.0.1 mataroalocal.blog
127.0.0.1 random.mataroalocal.blog
127.0.0.1 test.mataroalocal.blog
127.0.0.1 mylocalusername.mataroalocal.blog
/etc/hosts
does not support wildcard entries, thus there needs to be one entry
per mataroa user/blog.
To run the Django development server:
python manage.py runserver
If Docker and docker-compose are preferred, then:
- Set
DATABASE_URL
in.envrc
topostgres://postgres:postgres@db:5432/postgres
- Run
docker-compose up -d
.
The database data will be saved in the git-ignored directory / Docker volume
docker-postgres-data
, located in the root of the project.
Using the Django test runner:
python manage.py test
For coverage, run:
make cov
The following tools are used for code linting and formatting:
- black for code formatting
- isort for imports order consistency
- flake8 for code linting
- shellcheck for shell scripts
To use:
make format
make lint
Deployment is configured using uWSGI and Caddy.
A server playbook document is also available, based on Ubuntu 20.04 LTS.
Environment variables for production are defined both in uwsgi.ini
(for uWSGI)
and in .envrc
(for manage.py commands such as migrations and cron management
commands).
cp uwsgi.example.ini uwsgi.ini # edit environment variables in uwsgi.ini
uwsgi uwsgi.ini # start djago app
caddy start --config /home/roa/mataroa/Caddyfile # start caddy server
To reload or stop the uWSGI process:
uwsgi --reload mataroa.pid
uwsgi --stop mataroa.pid
# or find the PID and kill that directly
ps aux|grep uwsgi
kill -9 <PID>
To reload or store the Caddy webserver:
caddy reload --config /home/roa/mataroa/Caddyfile
caddy stop
Also, two cronjobs (used by the email newsletter feature) need to be installed. The schedule is subject to the administrator’s preference. Indicatively:
*/5 * * * * bash -c 'cd /home/roa/mataroa && source ./venv/bin/activate && source .envrc && python manage.py enqueue_notifications'
*/10 * * * * bash -c 'cd /home/roa/mataroa && source ./venv/bin/activate && source .envrc && python manage.py process_notifications'
0 0 1 * * bash -c 'cd /home/roa/mataroa && source ./venv/bin/activate && source .envrc && python manage.py mail_exports'
Documentation about these commands can be found in section Management.
Finally, certain setting variables may need to be redefined:
ADMINS
CANONICAL_HOST
EMAIL_HOST
andEMAIL_HOST_BROADCAST
To automate backup, one can run backup-database.sh
which
dumps the database and uploads it into any S3-compatible object storage cloud
using the MinIO client. This script needs the database password
as an environment variable. The key must be PGPASSWORD
. The variable can live
in .envrc
as such:
export PGPASSWORD=db-password
To restore a database dump:
pg_restore -v -h localhost -cO --if-exists -d mataroa -U mataroa -W mataroa.dump
To add on cron:
0 */6 * * * /home/roa/mataroa/backup-database.sh
In addition to the standard Django management commands, there are also:
enqueue_notifications
: create records for notification emails to be sent.process_notifications
: sends notification emails for new blog posts of existing records.mail_exports
: emails users of their blog exports.
They are triggered using the standard manage.py
Django way; eg:
python manage.py enqueue_notifications
One can deploy mataroa without setting up billing functionalities. This is the default case. To handle payments and subscriptions this project uses Stripe. To enable Stripe and payments, one needs to have a Stripe account with a single Product (eg. "Mataroa Premium Plan").
To configure, add the following variables from your Stripe account to your .envrc
:
export STRIPE_API_KEY="sk_test_XXX"
export STRIPE_PUBLIC_KEY="pk_test_XXX"
export STRIPE_PRICE_ID="price_XXX"
This software is licensed under the MIT license. For more information, read the LICENSE file.