Preparing qualified corpus table

[stakats]

The qualification step might only be this:

1. Select "content" column
2. There is no second step

Why so simple? Because we cannot be certain that we will have any meaningful metadata about a corpus, nor anything that can be extracted from filenames. But that's OK, we don't need any of that to do some NLP. We just need to know which column has the text.