Struct CompressionOptions 

pub struct CompressionOptions {
    pub method: CompressionMethod,
    pub tolerance: f64,
    pub max_bond_dim: usize,
    pub normalize_error: bool,
}

Configuration for tensor train compression.

Controls the accuracy-vs-cost trade-off when reducing bond dimensions via TensorTrain::compress or TensorTrain::compressed.

Fields

Field	Default	Meaning
method	LU	Decomposition algorithm (see CompressionMethod)
tolerance	1e-12	Relative truncation threshold per bond
max_bond_dim	usize::MAX	Hard upper bound on any bond dimension
normalize_error	true	Whether error is measured relative to the norm

Choosing tolerance

• 1e-12 (default): near machine precision, almost lossless.
• 1e-8 to 1e-6: good for most scientific applications.
• 1e-4 to 1e-2: aggressive compression, useful for exploratory work.

Tighter tolerances produce larger bond dimensions and slower evaluation.

Examples

use tensor4all_simplett::{CompressionOptions, CompressionMethod};

// Default: LU with tolerance 1e-12
let opts = CompressionOptions::default();
assert_eq!(opts.method, CompressionMethod::LU);
assert!((opts.tolerance - 1e-12).abs() < 1e-15);
assert_eq!(opts.max_bond_dim, usize::MAX);

// Custom: SVD with moderate tolerance and bond dim cap
let opts = CompressionOptions {
    method: CompressionMethod::SVD,
    tolerance: 1e-6,
    max_bond_dim: 50,
    ..Default::default()
};
assert_eq!(opts.method, CompressionMethod::SVD);

Fields

method: CompressionMethod

Decomposition method (LU, CI, or SVD).

tolerance: f64

Relative truncation tolerance per bond.

Singular values (or pivots) smaller than tolerance * sigma_max are discarded. Smaller values preserve more accuracy but produce larger bond dimensions.

max_bond_dim: usize

Hard upper bound on any bond dimension.

Even if the tolerance would allow a larger rank, the bond dimension is capped at this value. Set to usize::MAX (default) for no limit.

normalize_error: bool

Whether to normalize the truncation error by the tensor norm.

Trait Implementations

impl Clone for CompressionOptions

fn clone(&self) -> CompressionOptions

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for CompressionOptions

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for CompressionOptions

fn default() -> Self

Returns the “default value” for a type.