Change summarize to remove grouping on output dataframe

[bleearmstrong]

This removes any grouping variables after a select call. Should it be decided that dplython should adopt dplyr's style of summarizing, this is easily adapted.