We need to be able to return text metadata in a structured way, i.e. getting the desired string based on text_key and text_property, e.g. DataSet.get_text(name, text_property=[], text_key='').
text_property needs to be a list to support getting text versions for different contexts, e.g. plain question texts vs. axis-edits.
@nitsrek
@roxananeagu / @RoxanaMariaNeagu