Release 2.0.3

• Integration with NeptuneML feature set in AWS Neptune
• Add helper library to perform Sigv4 signing for %neptune_ml export ..., we will move our other signing at a later date.
• Swap how credentials are obtained for ROLE iam credentials provider such that it uses a botocore session now instead of calling the ec2 metadata service. This should make the module more usable outside of Sagemaker.
• Add sub-configuration for sparql to allow specifying path to sparql endpoint

New Line magics:

• %neptune_ml export status
• %neptune_ml dataprocessing start
• %neptune_ml dataprocessing status
• %neptune_ml training start
• %neptune_ml training status
• %neptune_ml endpoint create
• %neptune_ml endpoint status

New Cell magics:

• %%neptune_ml export start
• %%neptune_ml dataprocessing start
• %%neptune_ml training start
• %%neptune_ml endpoint create

NOTE: If a cell magic is used, its line inputs for specifying parts of the command will be ignore such as --job-id as a line-param.

Inject variable as cell input:
Currently this will only work for our new cell magic commands details above. You can now specify a variable to use as the cell input received by our neptune_ml magics using the syntax ${var_name}. For example...

# in one notebook cell:
foo = {'foo', 'bar'}

# in another notebook cell:
%%neptune_ml export start

${foo}

NOTE: The above will only work if it is the sole content of the cell body. You cannot inline multiple variables at this time.