KoreanSentenceAnalyser

Docs: https://hexdocs.pm/chinese_dictionary.

How to use

Add to mix.exs

def deps do
  [
    {:korean_sentence_analyser, "~> 0.1.0"}
  ]
end

Start elastic search via docker compose

docker-compose up -d

You can use the dockerfile and docker compose file in this project, and start it however you like. Just note that the version of elasticsearch has to be 6.1.1

For sentence analysis

KoreanSentenceAnalyser.analyse_sentence("한국어 배우기가 재미있어용")

Which will return a map, or nil if nothing is found

      %{
      "tokens" => [
        %{
          "end_offset" => 3,
          "position" => 0,
          "start_offset" => 0,
          "token" => "한국어",
          "type" => "Noun"
        },
        %{
          "end_offset" => 8,
          "position" => 1,
          "start_offset" => 4,
          "token" => "배우다",
          "type" => "Verb"
        },
        %{
          "end_offset" => 14,
          "position" => 2,
          "start_offset" => 9,
          "token" => "재미있다",
          "type" => "Adjective"
        }
      ]
      }

For a single word analysis

KoreanSentenceAnalyser.get_the_stem_of_a_word("완벽했지")

which will return a string, or nil if empty

"완벽하다"

Troubleshooting

In case Elasticsearch shuts down because of not enough virtual memory

sudo sysctl -w vm.max_map_count=262144