pandas using widget without import

[scottdraves]

import pandas as pd
pd.DataFrame()

is giving me this stack trace:

---------------------------------------------------------------------------
UnboundLocalError                         Traceback (most recent call last)
~/anaconda/envs/beakerx/lib/python3.6/site-packages/IPython/core/formatters.py in __call__(self, obj)
    916             method = get_real_method(obj, self.print_method)
    917             if method is not None:
--> 918                 method()
    919                 return True
    920 

~/src/beakerx/beakerx/beakerx/runtime.py in f()
    421     def __get__(self, model_instance, model_class):
    422         def f():
--> 423             display_html(TableDisplay(model_instance))
    424         return f
    425 

~/src/beakerx/beakerx/beakerx/tabledisplay/tabledisplay.py in __init__(self, *args, **kwargs)
    191     def __init__(self, *args, **kwargs):
    192         super(TableDisplay, self).__init__(**kwargs)
--> 193         self.chart = Table(*args, **kwargs)
    194         self.model = self.chart.transform()
    195         self.on_msg(self.handle_msg)

~/src/beakerx/beakerx/beakerx/tabledisplay/tabledisplay.py in __init__(self, *args, **kwargs)
     33         self.hasIndex = None
     34         if isinstance(args[0], DataFrame):
---> 35             self.convert_from_pandas(args, types_map)
     36         elif isinstance(args[0], dict):
     37             self.convert_from_dict(args)

~/src/beakerx/beakerx/beakerx/tabledisplay/tabledisplay.py in convert_from_pandas(self, args, types_map)
    123             else:
    124                 self.columnNames[:0] = [args[0].index.name]
--> 125             self.types[:0] = [self.convert_type("", args[0][column].get_values()[0])]
    126 
    127     @staticmethod

UnboundLocalError: local variable 'column' referenced before assignment

there are 2 problems.

1. displaying pandas dataframes with the widget should only be enabled if you "import beakerx" or "from beakerx import *". it seems to be enabled by default, incorrectly.
2. an empty dataframe is crashing our displayer.