time_pos index needs to be cast to int as of numpy 1.13.

[stochtastic]

Minor bugfix to address a problem I had with output_to_netcdf() using a clean Anaconda environment, reproducible via:

conda create -n GHCNpy python=2.7
source activate GHCNpy
conda install -c conda-forge geopy
conda install numpy netCDF4 requests matplotlib basemap
git clone https://github.com/jjrennie/GHCNpy.git
cd GHCNpy/
python setup.py install
python test.py

Testing output_to_netcdf(),

import ghcnpy as gp
gp.output_to_netcdf("USW00024233")

yields the following error from numpy:

('\nOUTPUTTING TO netCDF: ', 'USW00024233', '.nc')
('\nGETTING DATA FOR STATION: ', 'USW00024233')

GRABBING LATEST STATION METADATA FILE
Traceback (most recent call last):
  File "test_GHCNpy.py", line 2, in <module>
      gp.output_to_netcdf("USW00024233")
        File "/home/ar/Dropbox/Work/Projects/NXPhase/GHCNpy/ghcnpy/iotools.py", line 523, in output_to_netcdf
            ghcnd_data[time_pos,element_counter] = float(ghcnd_contents[counter][char:char+5]) / divisor
            IndexError: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices

Simple fix applied here is to cast time_pos to int.