fastapi-users VS databases

I'm working with it now as a backend to svelte! It's very ergonomic especially since I'm using sqlalchemy + postgres (asyncpg) so I could just copy paste the example code. Heads up though that it doesn't implement any form of token refresh pattern so you'll have to do that yourself.

It might be worth checking out fastapi-users. It has a lot of increased functionality like JWT with refresh, OAuth social login, cookie sessions, and more. It’s pretty cool: https://github.com/fastapi-users/fastapi-users

1) There is https://github.com/fastapi-users/fastapi-users library for FastAPI, but social auth is not very clear documented, DB not included.

found this: https://github.com/fastapi-users/fastapi-users

Well there is already an existing package that works pretty well : https://github.com/frankie567/fastapi-users

In the Python world, Flask and FastAPI are more performant, lighter frameworks. The FastAPI Users module looks like a good option for authentication.

I'd recommend you check out FastAPI-Users if you haven't already: https://github.com/frankie567/fastapi-users

It provides a users framework with OAuth2 integration, support for different ORM backends, and routes for authentication, verification, etc.

Full disclosure: I am a contributor to this project.

In the code ,we are using SQLAlchemy as our ORM(Object Relational Mapper) and Databases as our query builder.

MySQL is a powerful and popular database, and it's well-suited for use with Python. In the following section, we are going through how to set up a connection to a MySQL database in an asynchronous manner using the create_asynchrouns_engine function from SQLAlchemy. We'll also use the databases package and the aiomysql extra dependency. We'll also adapt the SQLAlchemy declarative approach to defining our users' table.

To take an example: I switched a service over from doing synchronous (plain def everywhere) to async (async def and await everywhere, with async DB, and saw… basically no performance improvement.

No, as I said, async db connection is the key. You can do that with SQLAlchemy using the databases library, but since you are not using SQLAlchemy it makes more sense to use the underlying db driver, aiomysql, directly.

Using the encode/databases library and forgoing SQLAlchemy's ORM

ormar is a mini async ORM for python. It uses sqlalchemy for building queries, databases for asynchronous execution of queries, and pydantic for data validation. You can create an ormar model and generate pydantic models from it.

works but there seems to be a difference between the regular slqalchemy syntax from what https://github.com/encode/databases/ is using in the docs?