Datatype must provide an itemsize

Hey,

I'm stuck with a problem using bayesian_optimization with the ''direct'' algorithm in python 3.5

The stacktrace yields:

```
Call-back cb_fcn_in_direct__user__routines failed.
Traceback (most recent call last):
.
.
.
File "env/lib/python3.5/site-packages/RoBO-0.2.1-py3.5.egg/robo/maximizers/direct.py", line 58, in maximize
File "env/lib/python3.5/site-packages/DIRECT/__init__.py", line 206, in solve
cdata
ValueError: data type must provide an itemsize
```

I have checked the itemsize of iidata, ddata, cdata and the bounds, but all provide an itemsize.

Thank you, Paul

Start with the traceback in robo/maximizers/direct.py at line 58 and the failing call in DIRECT/__init__.py at line 206. Reproduce the failure using Python 3.5, bayesian_optimization, and the direct algorithm, then compare the itemsize handling for iidata, ddata, cdata, and the bounds. Done means the direct algorithm no longer raises this ValueError for the reported setup.

Written by the indexing model from the issue text.