New features and bug fixes

[javierdlrm]

• Filter deployments by status and/or model
• Get server logs of a running deployment
• Deleting a running deployment with .delete() will warn the user. You can delete running deployments with .delete(force=True)
• Updating the predictor or transformer script of a running deployment with .save() now waits until the running instances have been updated
• Guiding messages in different methods, such as:
• • model.deploy() -> "Before making predictions, start the deployment by using .start()"
• • model.start() -> "Start making predictions by using .predict()"
• Descriptive prints for the main objects: PredictorState(status: 'Running')
• Better handling of exceptions, such as running predictions on a stopped deployment