Complete the extension of the Dioptra REST API to fully proxy all existing MLFlow functionality.

[keithmanville]

Identified MLFlow functions that need to be proxied:

• Mlflow.keras.log_model - create artifact, create new model version snapshot. Artifact files need to conform to mlflow definition of a model
• mlflow.log_param - this is information that is redundant of information we are already capturing
• Mlflow.keras.load_model - Takes model URI or artifacts URI.
• Create_registered_model - should happen in POST to ml model
• Create_model_version - should happen in PUT/POST to ml model version
• Mlflow.active_run - gives you context of ml flow run. Do not need this functionality.