async support for embeddings

[simonw]

Split from:

• Ability to execute prompts against an asyncio (non-blocking) API #507

[simonw]

Got o1-preview to attempt a first version of this: https://gist.github.com/simonw/47a9dd3a889001a147d5923f56a7f35c

files-to-prompt llm -e py -c | llm -m o1-preview "Implement asynchronous support for embedding models, inspired by how asynchronous support for regular models work. You'll need to build a new base class for these that shares some implementation with llm.models.EmbeddingModel - plus build a get_async_embedding_model() and _models and suchlike, plus implement async embeddings in the OpenAI example plugin using the async versio of client.embeddings.create. Output just the files that need changes plus new tests." -u

That prompt:

Implement asynchronous support for embedding models, inspired by how asynchronous support for regular models work. You'll need to build a new base class for these that shares some implementation with llm.models.EmbeddingModel - plus build a get_async_embedding_model() and _models and suchlike, plus implement async embeddings in the OpenAI example plugin using the async versio of client.embeddings.create. Output just the files that need changes plus new tests.