Updating lexicon at build rather than runtime

[donkindred]

My use case requires results returned in ideally less than 50ms. Is it possible to save a lexicon after adding words & tags instead of at runtime? I would ideally run it as part of the deployment pipeline.

In addition to the speed concerns, I'm computing the same thing every session, and it's a (minor, yet annoying) waste.

[spencermountain]

my curiosity is piqued - would love to know more about the project.

Many ways to do this, but it is a very minor computation, compared to the 10s-of-thousands of conjugations compromise does at runtime. It's just a javascript object (found here) that you can do whatever you'd like to, including storing it in localstorage or whatever. compromise/one comes with a (nearly) zero lexicon, so you could ship it once, or something.
You could also fork and add your terms to ./data/lexicon and run npm run pack && npm run build to have it included on init.
cheers!