Kyle Barron

> this crate is almost 5 years old and has only ever had one insertion strategy, so I didn't see the benefit in adding another generic parameter to support genericity...

The hilbert values are `u32`, so it's just one extra buffer with `4 * numItems` (not `numNodes`) length

for ref https://docs.rs/kdtree/latest/kdtree/kdtree/struct.NearestIter.html

> Looking at polars, it seems they also have both, but I am not entirely sure about the connection between both. polars has a module-level `polars.from_arrow`. The main problem with...

The PyCapsule Interface is focused on use cases around _importing_ some foreign data to your library. I think the right way forward is not to specify a specific import API,...

> Is the python capsule available at runtime? Sorry, by "pycapsule objects" I meant to say "instances of classes that have Arrow PyCapsule Interface dunder methods and can export PyCapsules"....

Well, I'd say the point of arro3 is to handle cases like this. But at the same time _stable enough to be a required pandas dependency_ is a pretty high...

FWIW > ```py > df = pyarrow.Table.from_arrays( > odf.column_arrays(), names=odf.column_names() > ).to_pandas() > ``` That can be just ```py pyarrow.table(odf).to_pandas() ``` as long as `odf` implements the pycapsule interface

> So when implementing one protocol, it should be `__arrow_c_stream__` I think. Yes that makes sense and I'd agree. I'd suggest always exporting a stream.

I think the question is: what should data consumers be able to infer based on the presence of one or more of these methods. See https://github.com/apache/arrow/issues/40648 - If an object...