Enable thinking mode in full chat view, keep disabled in widget/panel

stream_chat_with_tools now accepts a think parameter. run_generation
forwards it to Ollama. The message POST route reads think from the
request body. ChatView passes think=true so qwen3 uses chain-of-thought
reasoning for full conversations; the dashboard widget and ChatPanel
omit it, staying fast. Dashboard button updated to "Think it through
in Chat →" to signal the deeper capability.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

src/fabledassistant/services/llm.py

@@ -114,12 +114,17 @@ async def stream_chat_with_tools(
     messages: list[dict],
     model: str,
     tools: list[dict] | None = None,
+    think: bool = False,
 ) -> AsyncGenerator[ChatChunk, None]:
     """Stream chat completion from Ollama with tool support.
 
     Yields ChatChunk objects. If the model returns tool_calls, a
     ChatChunk(type="tool_calls") is yielded. Always ends with
     ChatChunk(type="done").
+
+    Set think=True to enable the model's chain-of-thought reasoning (qwen3+).
+    Thinking tokens are consumed by Ollama and not forwarded to the caller;
+    only the final response content is yielded. Expect higher TTFT when enabled.
     """
     options: dict = {"num_ctx": 32768}
     if tools:
@@ -129,7 +134,7 @@ async def stream_chat_with_tools(
         "messages": messages,
         "stream": True,
         "options": options,
-        "think": False,
+        "think": think,
     }
     if tools:
         payload["tools"] = tools