Skip to content

The fastest, cheapest, most efficient open-source agent harness. Run more than 500 agents on a $10 VPS.

chore: files changed crates/openhuman-embed/src/complete.rs

Steven Enamakel016f1aa112ef
diff --git a/crates/openhuman-embed/src/complete.rs b/crates/openhuman-embed/src/complete.rs
index 2c01898791..a0b75a60bb 100644
--- a/crates/openhuman-embed/src/complete.rs
+++ b/crates/openhuman-embed/src/complete.rs
@@ -326,6 +326,13 @@ pub struct CompletionResponse {
pub raw: Option<Value>,
}
+#[derive(Clone)]
+struct CompletionEndpoint {
+ base_url: String,
+ api_key: String,
+ headers: Vec<(String, String)>,
+}
+
impl CompletionResponse {
fn from_wire(response: ModelResponse, format: Option<&ResponseFormat>) -> Self {
let text = response.text();
@@ -635,9 +642,9 @@ impl Completer {
async fn dispatch(&self, request: CompletionRequest) -> Result<CompletionResponse, CoreError> {
let endpoint = self.checked_endpoint(&request)?;
let format = request.response_format.clone();
- let model = request.model.clone();
+ let model_id = request.model.clone();
log::debug!(
- "[embed] complete start method={COMPLETE} model={model} messages={} timeout_ms={:?}",
+ "[embed] complete start method={COMPLETE} model={model_id} messages={} timeout_ms={:?}",
request.messages.len(),
request.timeout_ms
);
@@ -645,7 +652,7 @@ impl Completer {
let model = endpoint.headers.iter().fold(
tinyinference_llm::providers::openai::OpenAiModel::new(endpoint.api_key.clone())
.with_base_url(endpoint.base_url.clone())
- .with_model(model.clone()),
+ .with_model(model_id.clone()),
|model, (name, value)| model.with_header(name.clone(), value.clone()),
);
let model: Arc<dyn tinyinference_llm::model::ChatModel<()>> = Arc::new(model);
@@ -698,7 +705,7 @@ impl Completer {
};
}
log::warn!(
- "[embed] complete failed method={COMPLETE} model={model} elapsed_ms={}",
+ "[embed] complete failed method={COMPLETE} model={model_id} elapsed_ms={}",
started.elapsed().as_millis()
);
CoreError::Rpc {
@@ -707,7 +714,7 @@ impl Completer {
}
})?;
log::debug!(
- "[embed] complete ok method={COMPLETE} model={model} elapsed_ms={}",
+ "[embed] complete ok method={COMPLETE} model={model_id} elapsed_ms={}",
started.elapsed().as_millis()
);
Ok(CompletionResponse::from_wire(response, format.as_ref()))