Function noll_index 

pub const fn noll_index(n: u32, m: i32) -> u32

The Noll single index of mode (n, m), from 1.

$j = n(n+1)/2 + |m| + c$, where the parity correction c alternates which of the $\pm m$ pair comes first with n mod 4. That alternation is the whole reason Noll indexing cannot be computed from |m| alone, and the usual place conversions go wrong.