- Change computation fallback result to return the empty dataframe(s)
instead of None
- Handle special case where only 1 TF is present
- Make the retry output more clear that it is a warning and not an error.

- Move package requirements to requirement.txt
- Update version, urls

Update algo.py (compatibility with latest pandas release)

BUGFIX for sparse expression matrices

In the pandas 0.25.3 and later version, the method of DataFrames, "as_matrix()", has been removed. And the substitution of the method is "to_numpy()".

- len() doesnt work on sparse matrices
- sklearn.GradientBoostingRegressor can handle sparse matrices for X,
but y has to be dense!

pygments style=borland

manni pygment style

sphinx theme, cfr. code highlighting problem