Struct TaskRecord 

#[non_exhaustive]
pub struct TaskRecord {

    pub task_id: String,
    pub parent: Option<String>,
    pub depth: u8,
    pub kind: String,
    pub prompt_ref: Option<PromptRef>,
    pub provider_key: Option<ProviderKey>,
    pub model_key: Option<ModelKey>,
    pub params: TaskParams,
    pub input_digest: Option<Digest>,
    pub rendered: Option<Value>,
    pub raw_output: Option<String>,
    pub parsed: Option<Value>,
    pub verdict: TaskVerdict,
    pub input_tokens: Option<u64>,
    pub output_tokens: Option<u64>,
    pub latency_ms: Option<u64>,
}

One call of one model task: what it was asked under, and what became of the answer.

Identifiers are paths (u2/extract, u2/extract#repair1), so the calls of one turn read as the graph they ran as. The rendered prompt and the raw answer are kept only when the deployment asks for them.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

task_id: String

Where the call sits in the turn.

parent: Option<String>

The task that asked for this one.

depth: u8

How many calls ran before it in its chain.

kind: String

The task kind, as its routing purpose names it.

prompt_ref: Option<PromptRef>

The instructions it ran under.

provider_key: Option<ProviderKey>

The provider that answered, when one did.

model_key: Option<ModelKey>

The model that answered, when one did.

params: TaskParams

The settings the call was sent with.

input_digest: Option<Digest>

Digest of the request as sent.

rendered: Option<Value>

The request as sent, when the deployment keeps prompts.

raw_output: Option<String>

The answer as received, when the deployment keeps it.

parsed: Option<Value>

The answer as parsed, when it parsed.

verdict: TaskVerdict

What the runtime made of it.

input_tokens: Option<u64>

Prompt tokens the provider reported.

output_tokens: Option<u64>

Output tokens the provider reported.

latency_ms: Option<u64>

Wall-clock time of the call.

Implementations

impl TaskRecord

pub fn new( task_id: impl Into<String>, kind: impl Into<String>, verdict: TaskVerdict, ) -> Self

A record of kind at task_id, with nothing learnt yet.

Trait Implementations

impl Clone for TaskRecord

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for TaskRecord

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for TaskRecord

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for TaskRecord

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

Equality operator ==.

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

Inequality operator !=.

impl Serialize for TaskRecord

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for TaskRecord