I started looking into it

After reconsidering, this ( Bookshop data from snakmodel ) should not be added

• The data is very finely cut, seems almost split on newlines, with every entry consisting of not much more than a single sentence so it is not in an ideal format, individual documents are not separated.
• It is probably already included in cellar: using deduplicate(threshold=0.9) from semhash when deduplicating against cellar, ~25% of the texts are filtered out. BUT If ~1000 documents are removed from cellar, split on newline and cleaned in the same way, only ~37% is filtered out.
  So even with data that is 100% included in cellar, when cleaned the same way, only 37% is filtered out. So there is a chance bookshop's data is already in cellar, but the amount of overlap is unclear. The separate documents could be reorganized into larger texts for further investigation, but based on https://op.europa.eu/en/web/cellar/architecture, Cellar is the "Publications office's common repository" and Bookshop is a collection from the EU's publication office.

So (beside the data structure being not ideal) there is enough circumstanital evidence for it being in cellar, to not be included in dynaword.