Ml2p

------------------

* Replace Sagefaker by moto sagemaker client to mock sagemaker.
* Refactor CLI commands.
* Add ml2p dataset generate command to generate datasets using the CLI.
* Add ModelDatasetGenerator class.
* Simplify how environment variables are passed to training job.
* Add ml2p-docker generate-dataset command.
* Update documentation.

------------------

* Unpin all dependencies to stay up to date with Pallets packages.

------------------

* Added the on-create feature to notebook servers.

------------------

* Pinned Flask to a stable version.

------------------

* Pinned Flask, Flask-API and click.

------------------

* Added reference documentation.
* Added a tutorial.
* Added tests for the ML2P command line utilities.
* Added support for attaching training and deployment instances to VPCs.
* Open sourced ML2P under the ISCL.