- default configuration files generated on first run
- database self-installation on first run
- searching
- search scope selections
- search scope exclusions
- near/not-near syntax
- progress polling
- automatic rewrites of searches to optimize for speed
- can cancel long in-progress searches (and/or vectorizations)
- dictionaries
- text browsing
- text maker
- indexing (+ flag unique forms)
- vocabulary list generation (+ flag unique headwords)
- morphology tables
- embedded help files
- semantic vectors
- general
- configurable model parameters
- configurable lists of stop-words
- choice of text building strategies
- nearest neighbors
- search for terms and yield a table and an interactive graph of nearest neighbors
- choice of graphing styles: "neighbors" or "neighbors + neighbors of neighbors"
- configure how many neighbors to track
- choice of modeling algorithms
- topic models
- configure number of topics to track
- option to build an interactive graph the LDA model via t-SNE reduction (graphs can be 2d or 3d)
- general
- for installation, see the
INSTRUCTIONS
folder - when running: hover over UI elements to get quick tips
- when running: click on
?
at bottom corner for detailed help - see also
antisigma
for video demos
- c. 83MB resident RAM on launch (note that a simple application like iTerm2 uses 354MB and Firefox uses 855MB)
- add about 125MB of RAM per postgres client connection in the pool
- a very heavy search like
all forms of πόλιϲ near...
can kick you up into 660MB
- go 1.22 because
slices.Concat
is called