Just a wild idea: How hard would it be to pre-compute word sketches like this in the database and offer them somehow somewhere in the interface on demand?
I realise this may be a bad fit for the PML-TQ interface and it may belong more into Kontext search engine or somewhere else, but can we compute them for any treebank (it will be different relations for each POS probably), store them in the database and offer them via API easily?
Just a wild idea: How hard would it be to pre-compute word sketches like this in the database and offer them somehow somewhere in the interface on demand?
I realise this may be a bad fit for the PML-TQ interface and it may belong more into Kontext search engine or somewhere else, but can we compute them for any treebank (it will be different relations for each POS probably), store them in the database and offer them via API easily?