Doubled speed of OBS2 read, 30% speedup of NAV2

eliminating numpy.genfromtxt() in nav2.py and obs2.py brought further substantial speed increase. Opens door for future JIT, MyPy, etc.
Default is now NOT verbose due to speed of reading. Use -v to show updates, plots.

10x overall speedup in RINEX 2 OBS reading

File that took one minute to read now takes only 5 seconds.
This accomplished by using 3-D Numpy array preallocated by 10 ms fast prescan of OBS file.

BUGFIX OBS2 with more than 10 measurements.

Note: this module would previously issue an error, not just silently give bad results. Now it works correctly.

Thanks @serioca fixes #25

cleanup SV renaming

Some RINEX writers leave gaps in SV names. instead of G07 they write G 7, which making indexing messy.
Now they are renamed in xarray.Dataset for easy, efficient indexing.

Hatanaka read, API simplification, specify measurements to read

• Enable Hatanaka compressed RINEX .crx and .crx.gz reading using stream processing via Hatanaka's crx2rnx--no temporary files are used
• simplify API with georinex.load convenience method
• speed reading by optionally specifying meas=

NAV file improvements

Add time reading from NAV files TimeRinex
Add several tests for handling of large NAV files
Add "use" option for RINEX 3 NAV files, to select specific satellite systems.

More RINEX3 NAV formats

Improved multi-system RINEX 3 NAV files (Beidou, GNSS, GLONASS, Galileo, QZSS, SBAS)

Increased test coverage of NetCDF4 (HDF5) read/write

Failsoft more parts of code where user may not have every prereq

Globbing, robust, .Z LZW read

• file globbing (input multiple files)
• read only data (nav, obs) specified by --tlim
• tolerate mangled parts of files (skips to next readable time step)
• increase speed by less eager reading of file
• using unlzw optionally, read .Z RINEX files, commonly used for RINEX 2.

Times, functionalize, read .Z files, optional indicators

Break up functions into logically grouped files for better maintainabiility

Expose time selection API (read parts of a file)

TimeRinex.py print times in OBS file

Add ability to read .Z lzw files via unlzw

Make Indicators (SSI, LLI) optional -useindicators over 2x reading speedup, and nearly 50% space savings (disk, RAM)

Rename, modernize, functionalize

Rename: PyRinex=>GeoRinex

Functionalize self-tests for easier diagnosis

Further functionalize/expose nice API, for example rinexheader to return dict() of header, with some mandatory values converted, and all others as str

pep8/type checking