Will Hohyon Ryu

If you're building from the source, you can use whatever dictionary you want to use. The instruction is in https://github.com/open-korean-text/open-korean-text/blob/master/docs/contribution-guide.md. If you're using the Java/Scala API, you can use `.addNounsToDictionary`...

detokenize 라는 기능을 이용하시면 되는데요 완벽하지는 않습니다. Hohyon Ryu Software Engineer on Airbnb Payments. On Sat, Feb 2, 2019 at 10:04 PM 이중건 wrote: > 형태소를 다시 문장으로 복구하는 알고리즘은...

I will need to update some rules around the number + noun match up. Thanks for raising this! On Mon, Jul 30, 2018 at 9:11 PM Hyeonguk Lee wrote: >...

@ShambhuT In openkoreantext, everything is static, which means nothing gets instantiated except for data classes. I wonder if your problem is related to Scala overhead, but it should be minimal....

Ah, that's totally legit. For now, I would not recommend running this in an Android environment. Can you build an API backend server for this? We have a source code...

Thanks for the test. OKT uses dynamic programming to parse. Since all the objects are immutable in scala, memory consumption is chosen to be inefficient. But in this extreme case,...

That's a good point. But unfortunately open-korean-text does not take context into account. So, it is not possible for now. There's an ml-based korean text processor https://github.com/kakao/khaiii. It looks promising...

ah, if you're considering all possible stems, it is possible with open-korean-text. I'm not sure if https://github.com/open-korean-text/open-korean-text/blob/master/src/main/scala/org/openkoreantext/processor/tokenizer/KoreanTokenizer.scala#L108 would return multiple verb/adjective candidates. I will need to test it a few...

Thank you! I just solved the issue by copying all the files except models from liblinear-1.51/ruby to the directory that holds my ruby scripts.

This seems to be a scala version conflict. I can change the scala version openkoreantext is using. Please let me know if it is needed. On Mon, Jul 30, 2018...