Code Coverage

[kmuehlbauer]

Currently we have roughly 90% code coverage.

While this seems good, much of that coverage is generated by reading files and checking some data of that files. This might give a wrong impression.

My suggestion is:

• [ ] split the current unit-tests into real unit-tests and filecheck-tests to separate these two wrt to code coverage
• [ ] bring unit-tests to 100% by adding tests (as much as possible)
• [x] fix notebook-tests to also create coverage report (currently broken: CoverageWarning: No data was collected. (no-data-collected)) #157
• [x] use codecov token #155