Add tests for endpoints code

[nsarrazin]

very WIP, will refactor to reuse most of the code between endpoints

plan is to test the following endpoints:

• HF Inference API
• Locally hosted TGI
• Locally hosted llama.cpp server
• OpenAI compatible (maybe openAI directly?)
• AWS SageMaker (will require spinning up an endpoint dynamically in a workflow and destroying it afterwards)
• AWS Lambda (same caveat as sagemaker)

And I will need to put all of this in a github action, testing the locally hosted services will be challenging, not sure yet how to spin them up inside the action.