That boilerplate can help you start a new REST appliaction based on Falcon Framewrok.
- Falcon - Web framework for building speedy web APIs
- redis-astra - Redis ORM provider
- Marshmallow - Library to simplified object serialization
- Python 3.7
- pipenv
- redis
Some solutions was inspired by Django, but this one has very simple implementation by design. The main idea gatchered from Django is: your project will have some dedicated logical blocks of code, called applications.
app.py it the only entrypoint. When you start it, it will try to load all packages from applications (see settings.yaml) and then it will call on_ready function on each application if that function existed there.
on_ready function receives reference to AppManager instance — this is applications common storage with one significant method: export. Applications can interact with each other through exported names.
For example, let's look at src/rest/__init__.py. On the on_ready function creates Falcon application and then that application exported under "rest" name. Then let's look at src/api/__init__.py. There is add_route method. This method will be called on Falcon application. Because app.rest is our application now.
Also there are system applications made by this way. System application do not mentioned in applications settings section. That applications will be run before the others during startup. One of them is Commands application. That application is avaliable for the others applications as app.commands. Thanks to this application, you can register your own commands to run throught the app.py. A good example contains in the rest application: app.commands.register('rest', handler). It means you can run app.py rest and the handler will be called.
Redis-astra is simpler than ORM in Django. You should just define model's fields. You also may define static method for object creation (see models/member.py). There is no save or commit methods: working with objects directly affects database. More info you can find on redis-astra page.
Marshmallow's serializers are used for objects representation (see serializers/member.py). It allows to use great features like custom fields in the answer.
git clone https://github.com/pilat/falcon-rest-boilerplate
cd falcon-rest-boilerplate
You can use Docker
docker-compose up --build
Or run it on your own system:
pipenv install && pipenv shell
python app.py rest
api/ directory contains an example of working as a REST service. Let's use httpie to communicate with our demo app:
# Create an member:
> http PUT http://localhost:8000/member/1 name='New member' [email protected] age=24
# Get member:
> http GET http://localhost:8000/member/1
# Get only several member's fields:
> http GET http://localhost:8000/member/1?fields=id,name
# Update member:
> http POST http://localhost:8000/member/1 name='Our member' age=25