Towards Fancy Indexing

[CarlKCarlK]

I just wanted to share my efforts toward offering fancy indexing in a project that uses ndarray.

Python Numpy supports fancy indexing -- that is indexing an array by

• a number (including negative to count from the end)
• a list-like collection of numbers
• a list-like collection of booleans
• range-like things, including steps (and negatives)

When I ported our Bed-Reader genomics package from Python to Rust, I wanted to continue supporting fancy indexing. In the context of reading DNA data from a file to an ndarray, I got it working.
See: https://docs.rs/bed-reader/latest/bed_reader/index.html#index-expressions &
Towards Data Science Nine Rules for Elegant Rust Library APIs

Is their work or interest in supporting fancy indexing generally in ndarray?
(I'll be at RustConf '22 in Portland, August 5th, if any other ndarray users or developers well be around.)

• Carl