Add embedding model API with vector storage and management UI

[Copilot]

Implements provider-based model architecture with embedding API endpoint and vector visualization.

Backend

• Database schema: providers, models (chat/embedding types), embeddings tables with foreign key relationships
• Embeddings API: OpenAI-compatible /embeddings endpoint that proxies to upstream providers, stores vectors as JSONB, tracks token usage and duration
• Model selection: Runtime model resolution by type with provider routing

Frontend

• Embeddings page: Data table with expandable detail panel showing vector values, dimensions, and metadata
• Provider/Model management: CRUD operations in settings for configuring upstream API providers and model mappings
• i18n: Full EN/CN translations for all new UI components

Example Usage

// Backend stores embedding vectors in database
POST /v1/embeddings
{
  "model": "text-embedding-3-small",
  "input": "Sample text"
}

// Frontend displays vector data
[0.123, -0.456, 0.789, ...] // 1536 dimensions

Test coverage includes unit tests for embedding storage, model selection logic, and UI components.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.