Pol Febrer Calabozo

There should probably be some differentiation between nodes of different cells, so that for example the searching algorithms don't see it as the same node. I think in the end...

> From what I see in the example by Pol, I understand that the implementation you are thinking about would require each atom to be uniquely associated to one category,...

Why isn't it obvious? A category just determines whether an atom fulfills or not a certain criteria, and a classifier groups atoms based on properties. The grouping might even have...

Hmm yeah that's true, a category is kind of a classifier with just two groups. As I said in the other issue, I'm too busy to think about anything useful...

Typer is built on top of click and you can customize a typer CLI with click concepts. With "raw click" I find the syntax for defining arguments as verbose as...

We could start with the `sisl_toolbox` CLI, which is much simpler and not that crucial and see how it goes :) I can submit a PR with the `sisl_toolbox` CLI...

My idea is to create a `CLIargs` dataclass that contains all parameters that a CLI framework might use and add it as metadata in the type hint using `Annotated`: ```python...

The problem with that output is that the second atom should have `[0,1,0,0]` as a neighbor right? If you can't pass `pbc` to the neighbor computing routines there should be...

On your first example, I don't think both should be treated the same. There is a supercell shift in the second atom and that shift (but with opposite sign) should...

I don't agree with what you say from the first example. I think for `xyz1` the neighbors of the first atom should be: ``` [ [1, 0, 0, 0], [1,...