Hello, Artem Chirkin. I have some question for easytensor.

[achirkin]

Hello, Artem Chirkin. I have some question for easytensor.

[achirkin]

Hi, thanks for your interest!

First of all, Vec3f is just a synonym for DataFrame Float '[3]. So you have here a nested DataFrame, normally, a preferred solution would be to have something like DataFrame Float '[XN k, N 3] (reads as "an array of 3D vectors, that has length of at least k"), or more general DataFrame Float (xns +: N 3). Though, you can always flatten nested DataFrames using joinDataFrame. But of course, you can surely use a nested DataFrame too.

Regarding your question, I don't remember if easytensor has this exact function, but you can make one using folds. For example, you can use sewfoldr, which reads as "simple element-wise right fold". That is, just like with a normal Foldable, you can write dataFrameToList = sewfoldr (:) [] to construct a lazy list (or sewfoldr (\x xs -> unScalar x : xs) [] for a nested DataFrame).
The Numeric.DataFrame.SubSpace module has a set of very flexible folds, maps, and traversals, which work with arbitrary dimensionalities (known or unknown at compile time). However, you may need to use lots of type annotations to fix the types (which makes it rather difficult use in the beginning).

[ubuntunux]

Thank you very much.