JSON should support using pivot.collections.Map directly

When supplying a pivot.collections.Map as the data object for a TableView, the data is fetched for each row using JSON.get(...). However, this method will default to using the BeanAdapter for anything other than a java.util.Map, which results in values like "class", "capacity", "count" and "comparator" fetching the values for the Map object instead of using the "get" method of the actual map to get the values.
This fix will likely speed things up (a bit) because the first "if" in the chain will now succeed for Pivot Map-based objects, instead of having to fail and do three other tests before getting to the "dictionary.get" for these objects.