streaming: add chat_with_tools_raw_streaming to LM Studio + LoggingChatbotClient

LM Studio: delegate streaming to the inner OpenAI client.
LoggingChatbotClient: extract shared log_and_return helper, add
streaming override so deltas are forwarded and metadata still logged.

crates/llm-client/src/lm_studio.rs

@@ -1,7 +1,8 @@
 use async_trait::async_trait;
 use serde_json::Value;
+use tokio::sync::mpsc;
 
-use crate::{ChatOptions, ChatResponse, ChatbotClient, LlmRawMeta, LlmTurn, Message, openai::OpenAiClient};
+use crate::{ChatOptions, ChatResponse, ChatbotClient, LlmRawMeta, LlmTurn, Message, StreamDelta, openai::OpenAiClient};
 
 /// LM Studio client.
 ///
@@ -48,4 +49,17 @@ impl ChatbotClient for LmStudioClient {
     ) -> anyhow::Result<(LlmTurn, Option<LlmRawMeta>)> {
         self.inner.chat_with_tools_raw(messages, tools, options).await
     }
+
+    /// LM Studio is OpenAI-compatible: streaming forwards to the inner client.
+    /// If a local build rejects `stream_options`, the inner pre-delta buffered
+    /// retry covers it transparently.
+    async fn chat_with_tools_raw_streaming(
+        &self,
+        messages: &[Value],
+        tools:    &[Value],
+        options:  &ChatOptions,
+        delta_tx: mpsc::Sender<StreamDelta>,
+    ) -> anyhow::Result<(LlmTurn, Option<LlmRawMeta>)> {
+        self.inner.chat_with_tools_raw_streaming(messages, tools, options, delta_tx).await
+    }
 }