moved fixes to sus in aLIGO.yaml to A+.yaml to fix sus noise curve

See merge request !24

modified quantum to ready for QN2mode

See merge request !22

This really doesn't need to be a part of precompIFO

matplotlib: use hex color names instead of the cool xkcd ones
When matplotlib >= 2.0 is ubiquitous, we can revert this.
numpy: use np.flipud in place of np.flip
np.flip only exists in numpy >= 1.12, and flipud is clearer anyway.

(addresses #23)

Suspension type BQuad changes the bottom stage material to silicon.
Reverse the SUS stage list in suspensionthermal.py to use the appropriate Temp.
Addresses #23

This is a port of the changes added to matgwinc in
gwinc/matgwinc@a83e5053b64e6e8d3176f8fc29df631972fec451

The new pygwinc quantum curve diverges slightly from the matgwinc
version at the highest frequencies, for reasons that are not clear
to me...

e.g. ifo['Sub.Element']

newtonian noise calculation was fixed in both py and mat gwinc, so
update matgwinc calculation cache in tests.

this was just a typo

use consistent naming

(Gitlab wants it to be called CONTRIBUTING.md)

Implements the changes recommended by Jan Harms in https://git.ligo.org/rana-adhikari/CryogenicLIGO/issues/45
Fixes #18

Add better description about IO, and clarify some of the argument help.

Add ability to load a noise budget previously saved in an HDF5 file.

If an HDF5 file is passed to the --save option (indicated by '.hdf5' extension)
then save the data to an HDF5 file without plotting.

Functions to save/load noise traces and IFO model parameters to/from HDF5
files.  Requires the h5py module.

free up the "util" namespace for other purposes not related to noise
calculation

namespace conflict with the gwinc plot keyword argument.