awkward VS django-ninja

there's a whole ecosystem in Python originally developed for high energy physics data processing: https://github.com/scikit-hep/awkward all because Numpy demands square N-dimensional array

Same technique used everywhere, here's a simple Julia pkg for the same thing: https://github.com/JuliaArrays/ArraysOfArrays.jl/blob/3a6f5b...

But Julia at least has the decency to just support ragged Vector{Vector} out of the box, and it's not that slow

Awkward Array.

Numba's @vectorize decorator (https://numba.pydata.org/numba-doc/latest/user/vectorize.htm...) makes a ufunc, and Awkward Array knows how to implicitly map ufuncs. (It is necessary to specify the signature in the @vectorize argument; otherwise, it won't be a true ufunc and Awkward won't recognize it.)

When Numba's JIT encounters a ctypes function, it goes to the ABI source and inserts a function pointer in the LLVM IR that it's generating. Unfortunately, that means that there is function-pointer indirection on each call, and whether that matters depends on how long-running the function is. If you mean that your assembly function is 0.1 ns per call or something, then yes, that function-pointer indirection is going to be the bottleneck. If you mean that your assembly function is 1 μs per call and that's fast, given what it does, then I think it would be alright.

If you need to remove the function-pointer indirection and still run on Awkward Arrays, there are other things we can do, but they're more involved. Ping me in a GitHub Issue or Discussion on https://github.com/scikit-hep/awkward-1.0

Django Ninja [1], it forever changed how I write Django project, in a way so elegant and productive.

[1]: https://django-ninja.dev/

Django Ninja is a RESTful wrapper for Django, while UtilMeta Python Framework uses a more concise declarative ORM Schema for Django and other future-supporting ORMs like sqlachemy and Peewee to build RESTful APIs more efficiently, and supports not only Django but all Python mainstream frameworks like Django, Flask, Starlette, FastAPI, Sanic, Tornado, etc.

I can't recommend [django-ninja](https://github.com/vitalik/django-ninja) enough. It's an easy to use, extremely fast, typed API for django. I've found it to be better in almost all aspects when compared to djangorestframework.

It's gaining popularity but is still widely unknown.

> The only place I really see Django at large companies is as an api using DRF or something.

This is not a bad thing. Using Django as an API backend is amazingly fast in terms of development time, especially with modern frameworks such as django-ninja [1].

Just use the built-in ORM to create models, write your endpoints, and use the built-in admin interface to play with the database if you don't have endpoints for everything.

There is also a less known feature of Django called admindocs [2], which automatically generates a human readable, hyperlinked documentation for your models and relations between them.

[1] https://django-ninja.rest-framework.com/

[2] https://docs.djangoproject.com/en/4.2/ref/contrib/admin/admi...

Personally, I also prefer django-ninja to DRF.

Actually that's not fully true. If you mix async and sync codes in django-ninja there will be some errors. Where's the proof ? django-ninja doesn't support async auth

Subscribe to this issue :D

Also recommend Django-Ninja. It basically reimplements fastapi's type and decorator-based API construction, but embedded directly in django so you have access to django's ORM and middleware library.