oktjs v0.1.3

oktjs

Port of Open Korean Text to JavaScript; it has no external dependencies, and runs in the browser.

Note that a modern browser with support for ES2018 RegExp unicode escapes is necessary.

Try it online!

Building

To build Oktjs, the following must be installed:

• A JDK.
• sbt to compile the Scala code.
• yarn to fetch dependencies and bundle the JavaScript code.

Then, yarn can be used:

$ yarn build

Details

Oktjs uses Scala.js to compile Open Korean Text to JavaScript, so it is cloned as a submodule to use its sources. A few changes are required to make it work with JavaScript:

• open-korean-text/src/main/scala/org/openkoreantext/processor/util/KoreanDictionaryProvider.scala is replaced by a shim.
  • The shim uses resources.js instead of embedded resources to load dictionaries.
  • resources.js embeds resources.json.gz using esbuild.
  • resources.json.gz is generated by resources.json.gz.build.js, which reads resources in open-korean-text/src/main/resources/org/openkoreantext/processor/util and writes them to a JSON file, which is then gzipped.
• A minimal shim of Twitter Text is provided.
• A minimal shim of CharArraySet is provided.
• A Scala.js wrapper for the Open Korean Text API is written in Okt.scala and then re-exported with types by index.ts.