Add openai embeddings

[Martok88]

This is working, but I haven't tested it extensively. I added a new optional parameter, maxBatchSize, which defaults to 100 if not set in the config. Apparently OpenAI's embedding endpoint can handle batches of 2048, but I also found reports of some users experiencing problems with large batches.
It still needs a documentation update, and should probably be reworked to use the OpenAI library.

TEXT_EMBEDDING_MODELS = `[
  {
    "name": "text-embedding-3-small",
    "displayName": "text-embedding-3-small",
    "description": "OpenAI hosted embedding model",
    "chunkCharLength": 768,
    "endpoints": [
      {
        "type": "openai",
        "url": "https://api.openai.com/v1/embeddings",
      }
    ]
  }
]`

[nsarrazin]

I pushed some minor changes here: 6fb16c9

I replaced authorization by apiKey so it matches the open ai endpoint and added a default for the URL.

Tested locally, it seems to work great! Thanks for adding this and if the changes look good to you we can merge this.

[nsarrazin]

I want to do a documentation pass in a later PR so I'll expand on this endpoint as well when doing it 😄

[Martok88]

Looks good!

[nsarrazin]

Thanks for the contribution! 🚀