Groupby adjacent values in continuous timeseries and get max value

Question

• I would like to group-by "grouping" column and get the max() value in said period

• The grouping should be based on whether adjacent values