Struct OpenAiProvider 

pub struct OpenAiProvider { /* private fields */ }

An OpenAI-compatible HTTP provider. The composition layer supplies its endpoint, credentials, and headers from runtime configuration.

Implementations

impl OpenAiProvider

pub fn new( base_url: impl Into<String>, api_key: impl Into<String>, extra_headers: HashMap<String, String>, ) -> OpenAiProvider

Construct a provider for the given endpoint and key.

pub fn with_retry_log(self, log: Arc<RetryLog>) -> OpenAiProvider

BP-7: report every retry this provider makes into log, which the caller drains after each completion (see RetryLog).

Trait Implementations

impl Provider for OpenAiProvider

fn complete<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, req: &'life1 ChatRequest, on_delta: &'life2 (dyn for<'a> Fn(&'a str) + Sync + Send), ) -> Pin<Box<dyn Future<Output = Result<(ChatMessage, Usage), RuntimeError>> + Send + 'async_trait>>

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

Run one completion. on_delta is called with each text chunk as it streams in. Returns the fully assembled assistant message and usage.

impl Provider for OpenAiProvider

fn complete<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, request: &'life1 ChatRequest, on_delta: &'life2 (dyn for<'a> Fn(&'a str) + Send + Sync), ) -> Pin<Box<dyn Future<Output = Result<(ChatMessage, Usage)>> + Send + 'async_trait>>

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

Run one completion and return its canonical assistant message and usage.