What is responder

A web framework created by kennethreitz, the creator of requests and pipenv, which many people know if you are using python.

Python hasn't had a framework that defines its own until Django, Flask, etc... appeared,

And the responder suddenly appeared in 2018 was a surprise.

Please refer to the official document for more details, but in short:

• A nice and cool framework for Flask and Falcon
• Supports functions such as GraphQL, async, etc. (as of 2019)
• API easy to use (easy to understand)

I think this is a top candidate for the Python Web Framework world.

What is tortoise-orm

You already know the term ORM made by Python SQLAlchemy. However, since this is also an old stuff (the first release in 2006), I searched for something new, and found the Tortoise ORM created by respecting Django's ORM.

I personally like Django's design philosophy, and I was able to simplify Model and Scheme in the same code, and support asyncio, so I decided to try it once

I think that compatibility with the responder feels pretty good for the above reasons.

Sample

https://github.com/shuto-S/sample-responder-tortoise

Technique used

• Python 3.7
• pipenv
• responder
• tortoise-orm

Install and run

Once you have pipenv installed, it should work if you hit the command as described in the sample in README file. It seems that tortoise-orm doesn't support pipinv, so it takes it directly from GitHub, but it works for just fine.

Since pipenv can add commands to scripts, it is also good when dealing with products.

[scripts]
migrate="python migrate.py"
start="python main.py"

Migration

Prepare the following file and execute it in Python.

from tortoise import Tortoise, run_async

async def migrate():
   # connect DB
    await Tortoise.init(
        db_url="sqlite://db.sqlite3",  # DB URL
        modules={"models": ["models"]}  # Set the file that Model is written
    )

    # run migrate
    await Tortoise.generate_schemas()

run_async(migrate())

Main

There are many ways to write, but you can set connect/disconnect as well as DB connection routing, and create main.py to launch the application.

When executed in Python, the responder starts up.

import responder
from tortoise import Tortoise
from router import add_routers

api = responder.API()

# Establish a connection to DB at startup
@api.on_event("startup")
async def start_db_connection():
    await Tortoise.init(
        db_url="sqlite://db.sqlite3",
        modules={"models": ["models"]}
    )

# Disconnect DB connection
@api.on_event("shutdown")
async def close_db_connection():
    await Tortoise.close_connections()

# Routing is set in router.py
add_routers(api)

if __name__ == '__main__':
    api.run(debug=True)

Routing

It was fine to write in main, but I created it because I wanted a routing-only file like Django's urls.py.

I want to refactor the source code without receiving the api: responder.API

import responder
from controllers.account import Account

def add_routers(api: responder.API):
    api.add_route("/accounts", Account)

Models

Of course implemented by tortoise-orm.

It is quite easy to use if you are used to work with Django.

from tortoise.models import Model
from tortoise import fields

class User(Model):
    id = fields.IntField(pk=True)
    name = fields.TextField()

    def __str__(self):
        return self.name

Controller

It corresponds to views.py in Django.

I was able to write quite cleanly because I can establish a routing configuration in router.py and a DB connection in main.py.

Also, since we only set the response (status code etc.) to resp, it is very pleased to the eye.

from models import User

class Account:
    async def on_get(self, req, resp):
        # Create a test account (user) for now
        await User.create(name="Test User")

        user = await User.first()
        resp.text = f"Hello, {user.name}"

Conclusion

An easy-to-use set of APIs that can be reached easily with a responder.

Django tick and clear ORM features from tortoise-orm.

It's just coming out (both first released in 2018), so I think it looks rough while using it, but because the compatibility is huge, this combination has the potential to aim for supremacy well enough.

There are a lot of functions that I haven't written about both in terms of samples, so you should try the idea of the Python Web Framework world.