Add a test that will run with every build (/merge to master) and will check

• Token alignment between UD and SPMRL dataframes
• Check for duplicate sentences against the known duplicates