Add support for programmatic manipulation of semantic model via sempy

The semantic-link library has huge potential. It would be really beneficial if semantic models could also be manipulated via code.

It would be great to see something like the following methods:

1. Add a new measure fabric.addMesaure(dataset,workspace, dax)
2. Add new table to semantic model (basically an extension of to_lakehouse_table() where i can also add it to a semantic model e.g. fabric.add_table_to_semantic_model(fabric_df, semantic_model, table_name)
3. in addition to 2. a function to create relationships in a semantic model: fabric.create_relationship(semantic_model, table1, table2, key_column1, key_column2, infer_cardinality)
4. Add an option to plot_relationship_metadata() to show all columns