This adds proper unit testing to the python interface. There are tests
for all image fusion algorithms that test invalid options and also make
a test prediction. There is also a test file for geo info and geo
transformation and one for the helper functions like for merging masks.

Unit testing is based on unittest module and the readme explains how to
run them. Alternatively, the unit tests are run during the conda build
process and in VS Code there is GUI support to run the tests. To capture
stdout during a test run (and receive print messages), the tests can
also be run with `nosetests tests/FILENAMEPART_test.py`. This might be
useful for debugging, but requires to install nose.

During testing some bugs could be spotted and fixed. This includes
`_get_prediction_jobs` and `_combine_mask_inplace`. Unit tests confirm
that these are resolved now. Additionally some input parameter checks
have been added that are also included in the unit tests.