Holoviews: import fails with pandas 0.20

[binaryaaron]

Hi,

Trying to follow some of the example notebooks led to this exception:

```{.python}
import holoviews as hv
import datashader as ds
from holoviews.operation.datashader import aggregate, shade, datashade, dynspread
from holoviews.operation import decimate
hv.notebook_extension('bokeh')
decimate.max_samples=1000
dynspread.max_px=20
dynspread.threshold=0.5

```{.python}
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-2-149b08508266> in <module>()
----> 1 import holoviews as hv
      2 import datashader as ds
      3 from holoviews.operation.datashader import aggregate, shade, datashade, dynspread
      4 from holoviews.operation import decimate
      5 hv.notebook_extension('bokeh')

/Users/agonzales/anaconda3/lib/python3.5/site-packages/holoviews/__init__.py in <module>()
     12                             commit="$Format:%h$", reponame='holoviews')
     13 
---> 14 from .core import archive                                # noqa (API import)
     15 from .core.dimension import OrderedDict, Dimension       # noqa (API import)
     16 from .core.boundingregion import BoundingBox             # noqa (API import)

/Users/agonzales/anaconda3/lib/python3.5/site-packages/holoviews/core/__init__.py in <module>()
      1 from .boundingregion import *  # noqa (API import)
----> 2 from .data import *            # noqa (API import)
      3 from .dimension import *       # noqa (API import)
      4 from .element import *         # noqa (API import)
      5 from .layout import *          # noqa (API import)

/Users/agonzales/anaconda3/lib/python3.5/site-packages/holoviews/core/data/__init__.py in <module>()
      9 import param
     10 
---> 11 from ..dimension import redim
     12 from .interface import Interface
     13 from .array import ArrayInterface

/Users/agonzales/anaconda3/lib/python3.5/site-packages/holoviews/core/dimension.py in <module>()
     12 import param
     13 
---> 14 from ..core.util import (basestring, sanitize_identifier,
     15                          group_sanitizer, label_sanitizer, max_range,
     16                          find_range, dimension_sanitizer, OrderedDict,

/Users/agonzales/anaconda3/lib/python3.5/site-packages/holoviews/core/util.py in <module>()
     30 
     31 try:
---> 32     import dask.dataframe as dd
     33 except ImportError:
     34     dd = None

/Users/agonzales/anaconda3/lib/python3.5/site-packages/dask-0.8.1-py3.5.egg/dask/dataframe/__init__.py in <module>()
----> 1 from .core import (DataFrame, Series, Index, _Frame, map_partitions,
      2                    repartition)
      3 from .io import (read_csv, from_array, from_bcolz, from_array, from_bcolz,
      4                  from_pandas, from_dask_array, from_castra, read_hdf,
      5                  from_imperative)

/Users/agonzales/anaconda3/lib/python3.5/site-packages/dask-0.8.1-py3.5.egg/dask/dataframe/core.py in <module>()
     34 return_scalar = '__return_scalar__'
     35 
---> 36 pd.computation.expressions.set_use_numexpr(False)
     37 
     38 

AttributeError: module 'pandas' has no attribute 'computation'

Downgrading to pandas 0.19 solves it; locale switching ala this fix does not.

[philippjfr]

This is an upstream issue with dask, you'll have to upgrade to dask 0.14.3. Let me know if you run into any more issues.