Add ability to customize examples for custom AI platform models

[jameswex]

• Motivation for features / changes

Cloud AI Platform models can have custom prediction code that takes in examples in any format the modeler wants. But WIT requires examples as lists, dicts, or example protos by design. So add the ability to transform examples to the custom format for cloud users that need this capability.

• Technical description of changes

• Add adjust_example function argument to set_ai_platform_model methods, similar to existing adjust_prediction argument.
• Call adjust_example before sending to the model if that argument is passed.

• Screenshots of UI changes

No UI changes.

• Detailed steps to verify changes work correctly (as executed by you)

Ran colab with custom cloud model that needs this capability and verified it works.

[jameswex]

@tolga-b please review

[tolga-b]

It looks good James. One comment for future, now we have many modes (set_ai_platform_model, set_custom_predict_fn, set_estimator etc.), I think it would be good to implement a clear_config function. For example, if someone sets aip then sets estimator it may cause unintended behavior depending on the order of if statements where config is being parsed. This is currently done is some functions (e.g. self.delete('estimator_and_spec') in set_custom_predict_fn).

[jameswex]

@tolga-b Agree we can approach that in a future PR

@stephanwlee please review :)