Struct FoldedRun 

pub struct FoldedRun {
    pub identity: RunIdentity,
    pub meta: RunMeta,
    pub context: ContextSnapshot,
    pub messages: Vec<MessageRecord>,
    pub inference_count: usize,
    pub inference_usage: Vec<InferenceUsageRecord>,
    pub tool_call_count: usize,
    pub pending_batch: Option<PendingToolBatch>,
}

The state reconstructed from a run journal - enough to resume or inspect the run at its latest recorded point.

Fields

identity: RunIdentity

The run’s current owner/identity.

meta: RunMeta

The latest run metadata.

context: ContextSnapshot

The reconstructed current context window.

messages: Vec<MessageRecord>

The recorded inbound messages, in order.

inference_count: usize

Number of inferences recorded.

inference_usage: Vec<InferenceUsageRecord>

Per-call usage, in the order the calls landed.

The point of keeping every entry rather than a running sum: a sum is already available from RunMeta, and what it cannot answer is whether any single call exceeded the window, or which kind of call the spend went to.

tool_call_count: usize

Number of tool calls recorded.

pending_batch: Option<PendingToolBatch>

A dispatched tool batch whose results never made it into the context window (the run crashed mid-batch). None when the run has no batch in flight or the batch’s turn already landed in context.

Trait Implementations

impl Clone for FoldedRun

fn clone(&self) -> FoldedRun

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for FoldedRun

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

Formats the value using the given formatter.

impl PartialEq for FoldedRun

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for FoldedRun