Ml pipeline update

[hannahbrown7]

This pull request covers:

• incorporating new sampling functions into prd_pipeline.py to save out a test dataset (need to optimise where this is saved to!), then return train and validate datasets.
• demonstrate the use of hyperdrive for distributed hyper parameter tuning, including removing hard coded hyper parameters in prd_cluster_train_demo.py

I have made minor changes to the prd_mlflow_on_azure_demo and prd_mlops_azml_cluster_demo notebooks which make use of prd_pipeline so that they still work with the updated sampling functions

closes #43 closes #46

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[stevehadd]

<Ignore the comment that was here, all the comment should now be visible in one place>

[stevehadd]

I can also confirm I have successfully rerun the hyperdrive notebook, so all looks to be working OK.