Struct TokenCalibration 

pub struct TokenCalibration {
    pub provider_tokens: u64,
    pub estimated_tokens: u64,
}

Provider-reported vs conservatively estimated footprint of ONE request. The user-visible context number is the provider’s; the policy’s is the estimate. Their ratio converts a target into estimator units.

Fields

provider_tokens: u64

estimated_tokens: u64

Implementations

impl TokenCalibration

pub const MIN_RATIO: (u64, u64)

Accepted estimate/provider ratios. The conservative estimator should read at or above the provider; outside this band a sample is more likely a transport/accounting artifact (omitted cache fields, a proxy, a mismatched request) than a tokenizer difference, so the plain estimator is used instead.

pub const MAX_RATIO: (u64, u64)

pub fn plausible(&self) -> bool

Trait Implementations

impl Clone for TokenCalibration

fn clone(&self) -> TokenCalibration

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for TokenCalibration

impl Debug for TokenCalibration

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

Formats the value using the given formatter.

impl Eq for TokenCalibration

impl PartialEq for TokenCalibration

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for TokenCalibration