Datapunt Authentication & Authorization service

Future work:

Move to an async framework; apart from encoding/decoding tokens (HMAC + SHA256) this service does nothing significant, it may as well do I/O async so it can handle more incoming connections. A possible stack could be (asyncio uvloop aiohttp) | tornado + asyncpg + requests-futures. OR GEVENT.

Contributing

Found a bug or want to work on the code? You can branch the repository on GitHub or file an issue at its issue tracker.

1. Install the dev dependencies in your virtual environment

$ pip install -e .[dev]

Note: the [dev] refers to extra requirements for development and is specified in setup.py. You don't need to install them as the make test and make coverage targets work fine without having them in the virtualenv. The only reason you might want to install them is so for example pytest and responses can be resolved in you IDE.

2. Create a configuration or environment file

In config.yml you can see an example configuration. You can either:

make a copy of config.yml, adjust it to your needs and point to it using export CONFIG=`pwd`/my_config.yml
or export values for the environment variables referenced in config.yml.

3. Make sure you have a database running and that the tables exist

If you have docker installed then you can start a Postgres instance with:

$ docker-compose up -d database

You can use the authorization CLI to set up a user. Note that this will also silently create the needed tables, if they don't exist.

$ authz user [username] assign [DEFAULT | EMPLOYEE | EMPLOYEE_PLUS]

Now you can develop, run and test code!