fix: consolidate streaming thought partials into single bubble

[ferponse]

Summary

When thought/reasoning chunks arrive via SSE streaming (e.g., Claude extended thinking, Gemini with thinking_config), each partial ReasoningChunk renders as a separate "Thought" bubble in the Dev UI instead of consolidating into one.

The consolidation logic in chat.component.ts correctly locates the existing partial thought event (lines ~1246/1255), but buildUiEventFromEvent creates a fresh UiEvent with only the current chunk's text. On replacement (line 1285), the previously accumulated text is lost.

Fix

Before replacing the existing thought event, prepend the existing thought text to the new event's text. This makes consecutive thought partials merge into a single progressively-growing bubble — matching the behavior of regular text streaming.

if (uiEvent.thought && existingEvent.thought) {
  uiEvent.text = (existingEvent.text || '') + (uiEvent.text || '');
}

Testing

• All 459 existing tests pass (ng test --no-watch --browsers=ChromeHeadless)
• Manually verified with vertex_ai/claude-sonnet-4-6 via LiteLLM with thinking={"type": "enabled", "budget_tokens": 10000}

Fixes #443

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.