Adapting to Input Token Limits

[BrettlyCD]

Using the Inference API with Hugging face has a limit of input tokens. Many of the schemas have more table information than can be put into a prompt. So you can either pay OR come up with a way to be more efficient with the space and target only relevant tables.