feat(chat): redesign input bar, expand sidebar, add thinking toggle

- Restructure floating input: dominant textarea with compact bottom
  toolbar (model badge, thinking toggle, attach, send/stop).
- Model badge sizes to the current selection (not widest option) via
  a layered transparent select, with truncate-on-overflow fallback.
- Auto-expand the conversation sidebar on desktop and slide chat
  content right when open instead of overlaying.
- Add per-request thinking toggle (brain icon, default on, persisted
  in localStorage) sending chat_template_kwargs.enable_thinking.
- Always disable thinking for title summarization.
- Generate chat titles before the main response to keep the SSE
  stream from staying open past visible completion and to avoid
  busting the KV cache between turns.

backend/internal/client/client.go

@@ -66,7 +66,7 @@ func (c *Client) EditImage(ctx context.Context, body openai.ImageEditParams) ([]
 	return resp.Data, nil
 }
 
-func (c *Client) SendMessage(ctx context.Context, chatMessages []*store.Message, model string, cb StreamCallback) (string, error) {
+func (c *Client) SendMessage(ctx context.Context, chatMessages []*store.Message, model string, enableThinking bool, cb StreamCallback) (string, error) {
 	// Ensure Callback
 	if cb == nil {
 		cb = func(mc *MessageChunk) error { return nil }
@@ -89,7 +89,8 @@ func (c *Client) SendMessage(ctx context.Context, chatMessages []*store.Message,
 		},
 	}
 	chatReq.SetExtraFields(map[string]any{
-		"timings_per_token": true, // Llama.cpp
+		"timings_per_token":   true, // Llama.cpp
+		"chat_template_kwargs": map[string]any{"enable_thinking": enableThinking},
 	})
 
 	// Perform Request & Allocate Stats
@@ -172,7 +173,7 @@ func (c *Client) CreateTitle(ctx context.Context, userMessage, model string) (st
 	output, err := c.SendMessage(ctx, []*store.Message{{
 		Role:    "user",
 		Content: prompt,
-	}}, model, nil)
+	}}, model, false, nil)
 	if err != nil {
 		return "", fmt.Errorf("failed to sent message: %w", err)
 	}