Struct TransferMetadata 

pub struct TransferMetadata {

    pub input_unit: &'static str,
    pub output_unit: &'static str,
    pub output_scale: u32,
    pub domain_min: u16,
    pub domain_max: u16,
    pub range_min: i32,
    pub range_max: i32,
    pub direction: MonotonicDirection,
    pub knot_count: usize,
    pub strictly_monotonic: bool,
    pub flat_segment_count: usize,
    pub requested_max_error: u32,
    pub achieved_max_error: u32,
    pub worst_case_input: u16,
    pub achieved_max_inverse_code_error: u16,
}

Compact facts recorded by the host generator for a transfer table.

The error fields describe numerical table and output-quantization error against the configured ideal source. They do not include sensor, component, ADC, model, self-heating, or calibration uncertainty.

Fields

input_unit: &'static str

Human-readable input unit, such as "adc_code" or "millivolt".

output_unit: &'static str

Human-readable physical output unit, such as "degree_celsius".

output_scale: u32

Integer output quanta per physical output unit.

domain_min: u16

Inclusive minimum input.

domain_max: u16

Inclusive maximum input.

range_min: i32

Inclusive minimum physical output (endpoint of the knot range).

range_max: i32

Inclusive maximum physical output.

direction: MonotonicDirection

Output monotonic direction.

knot_count: usize

Number of piecewise-linear knots.

strictly_monotonic: bool

True when every adjacent knot pair has unequal outputs.

flat_segment_count: usize

Count of adjacent knot pairs with equal outputs.

requested_max_error: u32

Requested maximum numerical error in output quanta.

achieved_max_error: u32

Exhaustively measured, conservatively rounded-up maximum error.

worst_case_input: u16

Input where the achieved maximum error first occurs.

achieved_max_inverse_code_error: u16

Exhaustively host-measured worst |invert(convert(x)) as i32 − x as i32| over the input domain, in codes, under the table’s default FlatResolution.

This is a measured round-trip bound, not a promise of identity: a nonzero value means some observations do not survive a convert-then-invert cycle exactly. Flat runs are resolved with FlatResolution::PreferLowInput; overriding the policy with PiecewiseLinearTransfer::with_flat_resolution can exceed this bound on tables where flat_segment_count is nonzero.

Trait Implementations

impl Clone for TransferMetadata

fn clone(&self) -> TransferMetadata

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for TransferMetadata

impl Debug for TransferMetadata

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

Formats the value using the given formatter.

impl Eq for TransferMetadata

impl PartialEq for TransferMetadata

fn eq(&self, other: &TransferMetadata) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for TransferMetadata