ENH: Raster handling

[kmuehlbauer]

This PR handles the issues outline in #135

new functions

• georef.create_raster_dataset (data, coords, proj -> dataset)
• georef.extract_raster_dataset (dataset -> data, coords, proj)
• georef.set_raster_origin (flip data and coords, if necessary)
• georef.reproject_raster_dataset (and resampling)
• io.write_raster_dataset (write to disc)
• util.get_raster_origin (returns upper or lower)
• add unit-tests

deprecated functions

• deprecate io.to_AAIGrid and to_GeoTIFF
• deprecate io.read_raster_dataset (the magic needed in this function isn't worth the work)
• deprecate georef.resample_raster_dataset

other fixes

• fix zonalstats.DataSource.dump_raster to use new functionality
• fixed io.read_safnwc (proj4/wkt-handling)
• update notebooks/fileio/wradlib_gis_export_example.ipynb
• update georef.read_gdal_values to read multi band data
• fix util.find_bbox_indices to account for lower or upper origin data/coords
• fix ipol.cart2irregular_spline to account for data/coords origin

[codecov]

Codecov Report

Merging #137 into master will increase coverage by 0.08%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #137      +/-   ##
==========================================
+ Coverage    82.6%   82.69%   +0.08%     
==========================================
  Files          18       18              
  Lines        4341     4467     +126     
==========================================
+ Hits         3586     3694     +108     
- Misses        755      773      +18

Impacted Files	Coverage Δ
wradlib/georef.py	90.37% <ø> (+0.03%)	⬆️
wradlib/io.py	89.79% <ø> (-0.96%)	⬇️
wradlib/ipol.py	94.19% <100%> (+0.03%)	⬆️
wradlib/zonalstats.py	86.77% <100%> (+0.05%)	⬆️
wradlib/util.py	81.34% <100%> (+0.47%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 50e3571...20f7473. Read the comment docs.

[kmuehlbauer]

The py2.7 Travis build is "Killed", which means memory over consumption... Will try to investigate this. The py3.5 and py3.6 builds running fine.

[kmuehlbauer]

The py2.7 Travis build is "Killed", which means memory over consumption... Will try to investigate this. The py3.5 and py3.6 builds running fine.

See #139, which resolved this issue.

[kmuehlbauer]

@heistermann

This is ready to go! I tried to trace down all uses of the changed/deprecated functions and use the new implementation instead. Could you have a final review?

[heistermann]

Sorry for being so lazy and slow on this. This is really great - from tracking down the problem towards implementing a solution. Thanks, Kai! Any further issues can be follwoed up in master, if required.