Struct CompletionRequest 

#[non_exhaustive]
pub struct CompletionRequest {
    pub rendered_prompt: String,
    pub output_schema: Option<Value>,
    pub execution_hints: ExecutionHints,
}

A single LLM completion request.

The caller is responsible for rendering the prompt — the trait does not own the template surface. output_schema is optional: leaving it None means “free-form text” and the adapter will treat the response as a raw string wrapped into crate::response::CompletionResponse::parsed_output.

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.

rendered_prompt: String

Fully-rendered prompt string the model should consume as user input. The trait deliberately carries a single string here rather than a system/user split: providers that distinguish roles can split the rendered prompt at a marker. If a future provider needs role-tagged or multi-part input, the trait must grow a structured input type rather than overloading this field — #[non_exhaustive] on this struct keeps that path open.

output_schema: Option<Value>

Optional JSON Schema 2020-12 document describing the expected output shape. When present, adapters pass it through to the provider’s structured-output API (Ollama format, OpenAI response_format.json_schema.schema, etc.). When absent the model is unconstrained and the adapter wraps any returned text into crate::response::CompletionResponse::parsed_output as a JSON string.

execution_hints: ExecutionHints

Generation hints. All fields are optional; adapters apply each hint only if the provider’s API supports it.

Implementations

impl CompletionRequest

pub fn new(rendered_prompt: impl Into<String>) -> Self

Build a request with just a rendered prompt; no schema, default hints. Use this for free-form text completion.

pub fn with_output_schema(self, schema: Value) -> Self

Builder-style setter for output_schema.

pub fn with_execution_hints(self, hints: ExecutionHints) -> Self

Builder-style setter for execution_hints.

Trait Implementations

impl Clone for CompletionRequest

fn clone(&self) -> CompletionRequest

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CompletionRequest

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

Formats the value using the given formatter.