Struct StubBackend 

pub struct StubBackend {
    pub reply: String,
}

A deterministic fake that streams a fixed reply word by word.

Enough to exercise streaming, early-stop, and token accounting with no model present. Registered as the stub provider, so a spec can select it with model = Stub "anything" — which makes it possible to test a pipeline end to end without depending on a model’s wording, or on a model at all.

Fields

reply: String

What every request generates.

Trait Implementations

impl Default for StubBackend

fn default() -> Self

Returns the “default value” for a type.

impl InferBackend for StubBackend

fn supports_images(&self) -> bool

The stub accepts images so a multimodal pipeline is testable without a vision model — but it reports what it received rather than discarding it. Silently ignoring images is the failure this whole guard exists to prevent; a test backend that does it cannot catch anyone else doing it.

fn infer<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, req: InferRequest<'life1>, on_token: &'life2 mut (dyn for<'t> FnMut(&'t str) -> bool + Send), ) -> Pin<Box<dyn Future<Output = Result<InferResult>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Generate from req.prompt, invoking on_token once per token.

fn model_name(&self) -> String

Identifier recorded in the job’s usage accounting.