Bridge dataclasses with SQLModel, pydantic and sqlalchemy

[adsharma]

In the last few years, dataclasses have become the standard way of creating Python objects.
They're convenient and fast because they don't validate at runtime and rely on static checkers.

On the other hand, pydantic and sqlalchemy are popular because they validate at runtime, do extensive meta programming.

This is an effort to have the cake and eat it too. You can use your dataclasses as usual. But with the new decorator, you get:

user.sqlmodel() # to generate a SQLModel object on the fly
user.__sqlmodel__ # the class corresponding to the object 

These provide the same functionality as SQLModel. Note that this doesn't include support for more advanced features like controlling the type used for the column in the database, foreign keys.

But hopefully it's a good start and we can keep adding support for more metadata in the future.

Benchmarks:

Creating 100000 User objects took 0.019142 seconds
Average time per object creation: 191.42 nanoseconds
Creating 100000 User SQL Model objects took 2.121499 seconds
Average time per object creation: 21214.99 nanoseconds