Struct OllamaRuntime 

pub struct OllamaRuntime { /* private fields */ }

Coordinates a shared Ollama client, model lifecycle helpers, and execution guards.

Implementations

impl OllamaRuntime

pub async fn new(config: RuntimeConfig) -> Result<Self>

Validates config, builds an HTTP client (timeouts + optional auth), then constructs Ollama via Ollama::new_with_client.

pub fn client(&self) -> &Ollama

pub fn guard(&self) -> &ExecutionGuard

pub fn models(&self) -> &ModelManager

pub async fn ensure_model(&self, model: &str) -> Result<()>

Shortcut for ModelManager::ensure. Equivalent to self.models().ensure(model).await.

pub async fn run<F, Fut, T>(&self, f: F) -> Result<T>

where F: Fn() -> Fut, Fut: Future<Output = Result<T>>,

Shortcut for ExecutionGuard::run. Equivalent to self.guard().run(f).await.

Prefer this at call sites when you only need guarded execution once; keep using Self::guard when you want a reusable handle (for example attaching metrics to the guard or holding ExecutionGuard::max_retries in scope).