Improve evaluation model params

[lu-ohai]

Improve evaluation model params

• Made all entities in evaluation as pydantic
• Extended AquaEvalParams to accept all model parameters from customer
• Reverted PR to remove this logic to service container side.

Notebook

• Create evaluation with model parameters

[github-actions]

📌 Cov diff with main:

📌 Overall coverage:

[mrDzurb]

I think pydantic by itself can give meaningful message if something wrong with the input params. We probably can have something like

except Exception as ex:
                raise AquaValueError(
                    "Invalid create evaluation parameters. "
                    f"Details: {ex}"

Can you check this?

In the evaluation framework i did this:

except ValidationError as err:
        custom_errors = {
            ".".join(map(str, e["loc"])): e["msg"] for e in json.loads(err.json())
        }
        raise InvalidEvaluationConfigValidationError(
            evaluation_config=config, errors=custom_errors
        )

[lu-ohai]

Just updated. Below are the results:

• Missing parameters

• Parameters with wrong type

[github-actions]

📌 Cov diff with main:

📌 Overall coverage: