closes #6

This directly compares the output of gwinc() and gwinc_matlab().  Any
noises that differ by more than 1% are noted and the return value of the
test is non-zero.  Differences can be plotted with the -p option.

This makes output directly comparable with pygwinc output

with --fom option

The return values are now commensurate with gwinc_matlab, making them
essentially interoperable.

addresses #2

code cleanup, IFO as yaml, CLI, test

See merge request !1

don't look for FilterCavity if "Freq Dependent" not defined.

This is now commensurate with what matlab gwinc does, and makes the IFO
structures directly comparable without any hacky conversions.

new gwinc_matlab module with gwinc_matlab() function that uses the
matlab.engine to calcualte noises with matlab gwinc.

Can be used as a hear drop-in replacement for the gwinc() function.

Made availabe to use via CLI with --matlab option

This makes it more compatible with matlab, allowing the resultant object
to be saved to a .mat and imported in matlab in the same format as matlab
gwinc ifo structures.

For now each noise just gets it's own trace.

direct port from MATLAB gwinc

This utilizes matlab.engine to run MATLAB gwinc on it's default IFOModel.
The resultant data (ifo and noises) are saved to gwinc.mat file, which is
then loaded by the test command.  The loaded ifo parameters are used to
create a pygwinc ifo model, from which pygwinc noises are calculated.

The difference between the matgwinc and pygwinc noises are then plotted
and displayed.b

This CLI produces a gwinc plot for the specified IFO model.  Includes options
to save to file, and interactive shell.

This adds a Struct() class that mimics a MATLAB struct by storing values as
class attributes, but also includes methods for converting to/from dicts,
yaml, and mat_struct objects.

A convenience function load_ifo() is included to load an ifo definition
either from a .yaml or .mat file or from an included <ifo>.yaml.

An 'aLIGO.yaml' definition is included.

no functional change

All noise calculation is moved to noise_calc() function.  Plotting is moved
to plot_noise() function.  The gwinc() function is now just a wrapper on
these two functions with some additional status reporting.