Tx Rob. Much appreciated!
In testing it I concentrated on specific usage relevant to me ...
- analysis of a novel chunked into about 4,000 tiddlers each corresponding to a paragraph of the original text
- test one - "identify all Tiddlers with dog related content (puppy, puppies, dog, hound, Mastiff) -> Worked WELL.
- test two - "identify all Tiddlers that mention "Harold" -> Fine but an issue on ordering of results.
I'll make a more detailed comment on your GitHub.
Overall: I think this is very useful but perhaps limited to some use cases? Maybe when you writing more discursive content? Anyway, at least one market would definitely be fiction writers.
Many thanks
Josiah