Comparison and Motivations¶
APIFlask starts as a fork of APIFairy (which share similar APIs with flask-smorest) and is inspired by flask-smorest and FastAPI. So, what are the differences between APIFlask and APIFairy/flask-smorest/FastAPI?
In a word, I try to provide an elegant (act as a framework, no need to instantiate additional extension object) and simple (more automation support for OpenAPI/API documentation) solution for creating web APIs with Flask. Here is a summary of the differences between APIFlask and similar projects.
APIFlask vs APIFairy/flask-smorest¶
It's a framework (and why?)¶
Although APIFlask is a thin wrapper on top of Flask, it's actually a framework. Thus, there is no need to instantiate additional extension object:
from flask import Flask from flask_api_extension import APIExtension app = Flask(__name__) api = APIExtension(app)
You only need to use the
APIFlask class to replace the
from apiflask import APIFlask app = APIFlask(__name__)
The key reasons behind making APIFlask a framework instead of a Flask extension is:
- I have to rewrite the
Flaskclass to ensure the natural order of the arguments injected into the view function.
- I have to rewrite the
Blueprintclass to add route shortcuts.
See the following two sections for more details.
A natural order of view arguments¶
By acting as a framework on top of Flask, APIFlask can overwrite the way Flask passes path arguments to view functions from keyword arguments to positional arguments.
Assume a view like this:
@app.get('/<category>/articles/<int:article_id>') # category, article_id @input(ArticleQuerySchema, location='query') # query @input(ArticleInSchema) # data def get_article(category, article_id, query, data): pass
With APIFlask, you can accept the arguments in the view function in a natural way (from left to right, from top to bottom):
def get_article(category, article_id, query, data)
However, with APIFairy, Flask-Smorest, or webargs, the path variables
article_id) need to be declared after the input data:
def get_article(query, data, category, article_id)
APIFlask added some route shortcuts (
app.post() , etc) for
Instead of doing something like this:
@app.route('/pets', methods=['POST']) def create_pet(): pass
You can just use
@app.post('/pets') def create_pet(): pass
Flask will have original support for these route shortcuts in the 2.0 version.
More automation for OpenAPI generating¶
- Add an auto-summary for the view function based on the name of view functions.
- Add success response (200) for a bare view function that only uses route decorators.
- Add validation error response (400) for view functions that use
- Add authentication error response (401) for view functions that use
- Add 404 response for view functions that contain URL variables.
- Add response schema for potential error responses of view function passed with
docdecorator. For example,
These automation behaviors can be changed with related configuration variables.
More features compare with APIFairy¶
- Add a
docdecorator to allow set the OpenAPI spec for view functions in an explicit way.
- Support more OpenAPI fields (all fields from
- Support to customize the API documentation config and CDN URLs.
- Return JSON response for all HTTP errors and Auth errors as default.
- Class-based view support.
APIFlask vs FastAPI¶
- FastAPI is nearly three years old, while APIFlask is only three months old. The former is production-ready. The latter is still in the early stage.
- For the web part, FastAPI builds on top of Starlette, while APIFlask builts on top of Flask.
- For the data part (serialization/deserialization, OpenAPI support), FastAPI relies on Pydantic, while APIFlask uses marshmallow-code projects (marshmallow, webargs, apispec).
- APIFlask builds on top of Flask, so it's compatible with Flask extensions.
- FastAPI support async. APIFlask will have the basic async support with Flask 2.0.
- APIFlask provides more decorators to help organize things better.
- FastAPI injects the input data as an object, while APIFlask passes it as a dict.
- APIFlask has built-in class-based views support based on Flask's
APIFlask vs Flask-RESTful¶
Flask-RESTful's core components were deprecated, see this issue for more details.