Raising warning/error when no data columns are found

[srcrowl]

Is your feature request related to a problem? Please describe.
When performing thresholding on a dataset (such as through the 'kinact.test_threshold function), it currently does not raise an error message indicating that specific data columns are lost or the reason. As a result, the output of the test threshold function when either no data columns are inputted or all data columns are removed for one reason or another is NaN, without any explanation as to why this is the case.

Describe the solution you'd like

• In the test_threshold function, if there are no data columns available, print at the summary statistics (maximum, minimum, and median) to be equal to 0, and add a warning that explains that no data columns were present.
• During binarization of an experiment, check to see if there are data columns present. If not, raise an error and provide possible explanations for why this is (no data columns provided and no columns with 'data:' in front, or data columns removed for one reason or another, such as not being numeric).