If I want to handle all possible output from TRIT is there a full taxonomy that I could look at to understand more about Topics or Social Tags or Relationships say?
I like to assume 2.5 seconds per document on average, its the safe number it really depends on the size of documents in your document set. so depending on what you mean by large corpus. I would say an example size of 100k documents would take around 70 hours with 1 concurrency or around 17 hours if you are hosting your own TRIT server with 4 concurrent requests going at all times.
This question was first asked in How to Enhance Your Search Platformswebinar (8th of Feb). Please see here: How to Enhance Your Search Platforms