Create Python wrapper for plot_tcmpr.R

[bikegeek]

plot_tcmpr.R currently supports generating the mean intensity error plot (with confidence intervals), which is similar to the hurricane track error plots found on Fang Ling's web page (click on any one of the Hurricane 201x links) :
http://www.emc.ncep.noaa.gov/gmb/wx24fy/vsdb/gfs2016/

plot_tcmpr.R is written in R and would require a significant effort to reproduce this script's behavior using Python. As an alternative, create a Python wrapper that can be accessed through the existing METplus code base.

Possible approach:

Rpy2 is a Python package that allows one to access R methods within a Python script. Unfortunately, this requires users to download and install Rpy2. Initially implement in a "cludgy" way by building up the requested input options and passing these to plot_tcmpr.R similar to what is done in METplus when invoking the MET tools.

[bikegeek]

Completed. Now users can also indicate a directory in the -lookin option and plots will be generated for each file in the directory. The file name (without extension) is used to create a subdirectory under the specified output directory to differentiate the plots.

[bikegeek]

Tagged the latest NCAR/METplus repo as METplus_alpha_20170424

[bikegeek]

Design notes are found on UCAR wiki:
https://wiki.ucar.edu/pages/viewpage.action?pageId=333022957