What new Python features are the most useful for you?

Post by xkcd-sucks » Thu, Jan 19, 2023, 5:56 PM UTC

randomluck040 wrote: post #4

pizza wrote: post #2

- einops (probably by far the most useful Python dsl for me as of late, esp. if you deal a lot with tensors) - numpy __array__ interface standardization (easy compatibility across a lot of numerical libraries, e.g. torch, awkward-array, zarr, jax.numpy, dask, xarray) - the various jit compilers (e.g. numba, jax.jit, torch.compile) - hw-accelerated functional programming style (e.g. functorch.vmap) - pybind11 for easy…

Matplotlib is really killing me. I never have learned it properly and switch between methods to address the same things in Matplotlib. Every time I have to plot something I have to Google how to do it or I have a template ready already. In my opinion it’s the most annoying library I have to use.

Post by mianos » Thu, Jan 19, 2023, 8:59 PM UTC

mharig wrote: post #38

mianos wrote: post #32

Earlier quoted context omitted.

OK, it's 100% likely I saw it, just so rarely I can't remember. Testing. Over that long time I have used many test frameworks. My normal workflow is to try and exercise all the code I write as I go. On the other hand I am 100% pedantic about using 'schema' ( https://pypi.org/project/schema/ ) to validate all the types, structure and domains when I read json.

Nearly the same here. The only type error I make a little bit more than I would like, is expecting an iterable and passing a string instead, because I forgot parentheses around arguments. But I made way more type errors when programming in C/++. Why do you prefer schema to (fast)jsonschema?