Add support for reasoning models with effort control and reasoning content display

[Copilot]

Adds support for OpenAI reasoning models (o1, o3-mini, etc.) with configurable reasoning effort and visible reasoning process.

Backend

• Added reasoning_effort parameter to ChatCompletionRequest (values: low, medium, high)
• Added reasoning_content field to ChatCompletionMessage and ChatCompletionDelta
• Extract reasoning content from model response_metadata in both streaming and non-streaming flows
• Pass reasoning effort to LLM via extra_body parameter

Frontend

State & Types:

• Extended ChatOptions, ChatRequest, and response types with optional reasoning fields
• Added reasoning effort to chat state with persistence

UI Components:

• Reasoning effort dropdown in chat settings (conditionally visible when model.supports_reasoning === true)
• Collapsible reasoning section in Answer component (blue-bordered, expandable)

Streaming:

• Separate buffer for reasoning content during streaming
• Reasoning displayed before main answer

Example

// Chat settings now include reasoning effort
const options: ChatOptions = {
  system: systemPrompt,
  temperature: 0.7,
  reasoning_effort: "medium"  // Optional, only for reasoning models
};

// Responses include reasoning content
interface ChatResponse {
  answer: string;
  reasoning_content?: string;  // Model's reasoning process
  // ...
}

All changes are backward compatible—non-reasoning models function unchanged.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.