Feature request: Add support for custom embedding base URL

[lesshaste]

I apologise if I have missed how to do this, but as far as I can tell the feature doesn't exist yet.

It would be great to add support for a custom base URL for embedding models (like OPENAI_EMBEDDING_BASE_URL or embedding_base_url config option).
Currently, the EmbeddingClient in openevolve/embedding.py only supports:

• OpenAI embedding models (text-embedding-3-small, text-embedding-3-large) using api.openai.com
• Azure embedding models
  This makes it impossible to use other providers for embeddings like:
• OpenRouter (which offers qwen/qwen3-embedding-8b and other embedding models)
• Local embedding servers
• Other OpenAI-compatible APIs
  The fix would involve:

1. Adding a new config option (e.g., embedding_base_url) to DatabaseConfig in openevolve/config.py
2. Modifying EmbeddingClient.init to accept a base_url parameter
3. Passing base_url to the OpenAI client initialization
   This would allow users to set OPENAI_EMBEDDING_BASE_URL env var or embedding_base_url in their config to use any OpenAI-compatible endpoint for embeddings.
   Example use case: Using OpenRouter for embeddings while using a different provider for LLMs.

[codelion]

This is a good idea to add and related to #323 which asks for Gemini embedding support. Your proposal is the more general solution since an embedding_base_url config option would cover any OpenAI-compatible endpoint.