Custom mapping of pandas DataFrame columns to KNIME types

[applio]

knimepy should support custom mapping of pandas.DataFrame columns to particular KNIME types via an optional dict-like object.

I further suggest that we create an example for one of these dict-like objects, ideally suited for chemical data (especially SMILES data and other KNIME types typically coming from tools like RDKit) that can be used from knimepy out-of-the-box by lots of people without modification. This example would recognize a column name pattern of '*_smiles' and the like for other formats. Per @greglandrum suggestion: also match the plain form of each. So "^smiles$", "^smarts$", etc.