Bug Fix: Fix issue with trailing comma creating an empty node

[mialy-defelice]

Background

Fix for issue described in FDS-1392.

When a CSV data model has a trailing comma in a list, causing an empty string to be recorded as a node, this creates an empty value node, and errors when trying to create the JSONLD.

For this issue, delete any empty nodes from lists so this does not create the cascading issues.

To Reproduce

Using the current develop-refactor-schemas branch, attempt converting attached EL.data.model.csv (see below) to JSONLD with the following CLI command:

schematic schema convert path/to/EL.data.model.csv

It should error.

To Test

Use the fix branch with the same model and command, and it should pass.

For sanity, you can test the model where I deleted the problematic comma, and it goes through the conversion on the current branch.
EL_del_comma.data.model.csv
EL.data.model.csv

[mialy-defelice]

Will try updating load DF instead of the current fix.

[mialy-defelice]

Leave in previous fix, bc we want to ensure no empty strings go through, but deal with trailing comma in load_df.

[mialy-defelice]

@linglp @GiaJordan I attempted to do the fix where we let data models use the trim commas functionality in df_utils, however this resulted in many formatting changes not expected, meaning right now its not usable, for data models so I am going to stick with the fix developed in this PR.

[GiaJordan]

Could reproduce the error on develop-refactor-schemas and confirm that it was resolved on this branch!