Remove trailing empy columns in investigation sections

[mkuhring]

The investigation.txt is a tab-separated file with a varying number of columns in the different sections depending on the number of entries made there (e.g. number of declared ontologies, authors, assays, factors, ...).
I.e. per specification, different rows have a different number of fields.

When using Excel or LibreOffice Calc to adapt the investigation file, it will be saved with the same number of fields per rows afterwards.
This will result in a failed parsing attempt by altamisa, as the file doesn't meet the ISA-tab criteria.

By default, these files should always fail.
But as additional option, we could provide a forced cleanup of empty trailing columns per section.

The emphasis here is on "columns per sections".
Since the number of fields need to be consistent for all rows in a section.
I.e., if only one row of a section contains a value in an outer column, empty fields have to be retained in the other rows of the section.
So don't just delete empty trailing fields row by row but mind the context of the section.

Again, this should be optional / actively enabled.
But also come with an additional warning (since e.g. with SODAR it will be probably on by default).