Struct ModelContextLimits 

pub struct ModelContextLimits {
    pub provider: String,
    pub model: String,
    pub context_window: u64,
    pub max_completion_tokens: u64,
    pub recommended_completion_tokens: u64,
    pub source: ModelLimitSource,
    pub confidence: ModelLimitConfidence,
}

Provider-neutral model context limits.

Adapters translate provider-specific metadata into this shape so the ledger stays free of provider-client types. See ModelContextLimits::resolve.

Fields

provider: String

Provider label (e.g. "umans", "opencode-go").

model: String

Model id as selected by the user (e.g. "umans-coder").

context_window: u64

Total context window in tokens (input + output).

max_completion_tokens: u64

Maximum completion tokens the model accepts.

recommended_completion_tokens: u64

Recommended completion tokens for typical turns.

source: ModelLimitSource

Where the limits came from.

confidence: ModelLimitConfidence

How trustworthy the limits are.

Implementations

impl ModelContextLimits

pub fn available_input_budget(&self) -> u64

Available input budget = context window minus reserved completion budget and provider overhead.

Selection targets TARGET_BUDGET_RATIO of this value.

Auto-compaction may trigger above AUTO_COMPACTION_RATIO.

pub fn target_budget(&self) -> u64

Token budget the selection policy targets (80% of available input).

pub fn auto_compaction_threshold(&self) -> u64

Token budget above which auto-compaction may trigger (92% of available input).

pub fn resolve( provider: &str, model: &str, override_entry: Option<ModelLimitOverride>, live: Option<&LiveModelMetadata>, ) -> (ModelContextLimits, Vec<ContextDiagnostic>)

Resolve limits from candidate sources in precedence order: user override → live metadata → static provider metadata → fallback.

Returns the chosen limits plus diagnostics for invalid overrides and fallback usage.

Trait Implementations

impl Clone for ModelContextLimits

fn clone(&self) -> ModelContextLimits

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ModelContextLimits

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

Formats the value using the given formatter.

impl Eq for ModelContextLimits

impl PartialEq for ModelContextLimits

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for ModelContextLimits