How do I add the catalog input to the SAR model?

The [documentation](https://github.com/Azure/mmlspark/blob/master/docs/SAR.md) for the SAR model says:

> The input to SAR consists of:
>
> * transaction (usage) data
> * catalog data

I am possibly being a little dim but...there's no way to add a dataframe containing catalog data to the SAR Model defined in the the Python package for this repository. I can't see any parameter or function on either instantiation or during the `.fit()` call that accommodates that.

How do I add the catalog data to the SAR algorithm through the Python API?

Start with docs/SAR.md and the Python SAR model API, then compare the documented transaction and catalog inputs with the model constructor and its .fit() call. The issue is complete when the relationship between catalog data and the Python API is clearly documented and consistent.

Written by the indexing model from the issue text.