normalisation in document/term frequency graphs

The document frequency visualisation offers the option to normalise by the total documents matching your query, and the term frequency element gives the option to normalise by the total documents in the corpus per bin (e.g. the total documents in that year).

This is inconsistent: there is no reason why either graph should not have the other type of normalisation. It can also cause confusion since these options can sound kind of similar, so if they are never presented next to each other, they might appear as inconsistent descriptions of the same thing.

We should just make both options available for both graphs.