Struct Inputs 

pub(super) struct Inputs<const N: usize> {
    pub(super) dim: usize,
    pub(super) axes: [usize; 3],
    pub(super) k: Jet<N>,
    pub(super) r: [Jet<N>; 3],
    pub(super) kpar: [Jet<N>; 3],
    pub(super) direct: [[Jet<N>; 3]; 3],
    pub(super) reciprocal: [[Jet<N>; 3]; 3],
    pub(super) measure: Jet<N>,
}

The inputs of one sum as jets in the slots above (or constants for the split derivative), in lattice coordinates except for the Cartesian shift.

Fields

dim: usize

axes: [usize; 3]

Cartesian axis of each lattice coordinate.

k: Jet<N>

r: [Jet<N>; 3]

kpar: [Jet<N>; 3]

direct: [[Jet<N>; 3]; 3]

reciprocal: [[Jet<N>; 3]; 3]

Reciprocal rows; a_i · b_j = 2π δ_ij gives their vector derivatives.

measure: Jet<N>

Cell measure |det a|.

Implementations

impl<const N: usize> Inputs<N>

pub(super) fn new( wave: Family, k: Complex, lattice: &BlochLattice, r: [f64; 3], variable: bool, with_reciprocal: bool, ) -> Self

Without with_reciprocal the reciprocal rows and the measure, which only the Ewald parts read, stay zero: the direct shells skip their jets. Always inlined, so the caller builds the inputs in its own frame instead of copying a returned value.

pub(super) fn point(&self, rows: &[[Jet<N>; 3]; 3], n: [i64; 3]) -> [Jet<N>; 3]

Integer combination Σ_i n_i rows_i of lattice rows.

pub(super) fn combination( &self, rows: &[[Jet<N>; 3]; 3], c: [f64; 3], ) -> [Jet<N>; 3]

Combination Σ_i c_i rows_i of lattice rows with integer-valued c.

pub(super) fn reduce( &self, rows: &[[Jet<N>; 3]; 3], reduction: Option<&Reduction>, ) -> [[Jet<N>; 3]; 3]

The reduced rows U rows, or rows when they are reduced already.

pub(super) fn image(&self, point: &[Jet<N>; 3]) -> [Jet<N>; 3]

Cartesian displacement -r - R of the image at lattice point R.

pub(super) fn phase(&self, point: &[Jet<N>; 3]) -> Jet<N>

Bloch phase q · R of lattice point R.