Investigate white space words.
When importing text there are unique words being imported, which only contain white space characters. This does not cause any problems, because they are not displayed as interactive words, but they should be filtered before they even get imported.
It is also important to keep it working for people, who already have text imported with white space words.
I was thinking that this might be tanjentially related to what happened with the issue we discussed in https://github.com/simjanos-dev/LinguaCafe/issues/227.
It's 99.9% not, because those whitespace words are being removed from the text either while importing, or when requesting the text for the text reader. This github issue does not cause any problem in functionality, there are just a few weird white space words on the vocabulary search page.