Flask REST

This library is a tiny REST toolkit intending to simplify your life when you want to create a REST API for your flask apps.

Install it

Well, that's really simple, it's packaged and on PyPI, so:

$ pip install flask-rest

Use it

Handlers

Create your classes with specific methods (namely add, get, delete and update), register it with an url, and you're good.

Here is a simple example on how to use it:

from flask import Blueprint
from flask_rest import RESTResource, need_auth

# Subclass a RestResource and configure it

api = Blueprint("api", __name__, url_prefix="/api")

# You can define a authenfier if you want to.

class ProjectHandler(object):

    def add(self): #This maps on "post /"
        form = ProjectForm(csrf_enabled=False) # just for the example
        if form.validate():
            project = form.save()
            db.session.add(project)
            db.session.commit()
            return 201, project.id
        return 400, form.errors # returns a status code and the data

    def get(self, project_id): # maps on GET /<id>
        # do your stuff here
        return 200, project

    # you can use the "need_auth" decorator to do things for you
    @need_auth(authentifier_callable, "project") # injects the "project" argument if authorised
    def delete(self, project):
        # do your stuff
        return 200, "DELETED"

Once your handlers defined, you just have to register them with the app or the blueprint:

project_resource = RESTResource(
    name="project", # name of the var to inject to the methods
    route="/projects",  # will be availble at /api/projects/*
    app=api, # the app which should handle this
    actions=["add", "update", "delete", "get"], #authorised actions
    handler=ProjectHandler()) # the handler of the request

If everything should be protected, you can use the authentifier argument:

authentifier=check_project

Where check_project is a callable that returns either the project or False if the acces is not authorized.

Serialisation / Deserialisation

When you are returning python objects, they can be serialized, which could be useful in most of the cases. The only serialisation format supported so far is JSON.

To serialise normal python objects, they should have a _to_serialize attribute, containing all the names of the attributes to serialize. Here is an example:

class Member():

    _to_serialize = ("id", "name", "email")

    def __init__(self, **kwargs):
        for name, value in kwargs.items():
            setattr(self, name, value)

If you want to have a look at a real use for this, please head to https://github.com/spiral-project/ihatemoney/blob/master/budget/api.py