few other cleanups

add automatic cacheing of various functions, and a cache_test() routine to test the runtimes with and without caching

set up json functions to not convert ascii (should be faster and preserve UTF8)

new gridtype: "discrete". This samples the grid points and does not set the probability.

attempted to harmoize the Moe and di Stefano (2017) variable names so they all use Moe2017 (or MOE2017)

added options to the Moe loader so we can change the JSON data and reload it after clearing the cache

the period distribution renormalization now gets the low and high limits from the loaded data

update_dicts() now copies the type of dict_1 for the new dict, rather than using OrderedDict (maybe merge_dicts should do this too?)

cleaned up some of the verbosity to be more like binary_grid and less annoying

Fixed some complaints that pylint had and fixed a correct mass ratio lower bound in the interpolators

Working on the M&S: almost done with everything. Just need to get the stuff right from multiple multiplicities in 1 run

Working on the M&S grid now: added alot of verbose output, and making sure all the interpolators are freed. Now I need to add the eccentricity as options and output for plots. Also reformatted