search

jkomoros / card-web

The web app behind thecompendium.cards
Apache License 2.0
46 stars 8 forks source link

Written-out numbers and numerals should normalize to the same thing #526

Open jkomoros opened 3 years ago

jkomoros commented 3 years ago

E.g. "0 to 1" and "zero to one" should be equivalent in the nlp machinery. Some places I use written out numerals and some

Probably fine to only handle up to 100 (as well as thousand, million, hundred, etc)

jkomoros commented 2 years ago

Probably just a thing in the NLP machinery to, before stemming, convert e.g. "one hundred" to "100".

Whenever one of the words is found, it should also look ahead one word, and if it finds another word, combine them.

Should do tests to make sure common cases are handled well