Elasticsearch のユーザ辞書の設定

Elasticsearch でユーザ辞書を利用する方法のメモ。
■辞書（config/dic/user_dic.txt）

リバーシブル,リバーシブル,リバーシブル,カスタム名詞
ヨガマット,ヨガ マット,ヨガ マット,カスタム名詞

■settings（settings_dic.json）

{
  "settings": {
    "analysis": {
      "tokenizer": {
        "kuromoji": {
          "type": "kuromoji_tokenizer",
          "user_dictionary": "dic/user_dic.txt"
        }
      },
      "analyzer": {
        "ja_analyzer": {
          "tokenizer": "kuromoji",
          "type": "custom",
          "mode":"search"
        }
      }
    }
  }
}

■settings_dic.json を反映

curl http://localhost:9200/settings_dic/?pretty \
     -XPUT \
     -H "Content-Type: application/json" \
     -T settings_dic.json

■解析実行

TEXT='リバーシブルヨガマットを買った'
curl -XGET \
     http://localhost:9200/settings_dic/_analyze?pretty \
     -H "Content-Type: application/json" \
     -d "
{
  \"analyzer\": \"ja_analyzer\",
  \"text\": \"${TEXT}\"
}"

■解析結果

{
  "tokens" : [
    {
      "token" : "リバーシブル",
      "start_offset" : 0,
      "end_offset" : 6,
      "type" : "word",
      "position" : 0
    },
    {
      "token" : "ヨガ",
      "start_offset" : 6,
      "end_offset" : 8,
      "type" : "word",
      "position" : 1
    },
    {
      "token" : "マット",
      "start_offset" : 8,
      "end_offset" : 11,
      "type" : "word",
      "position" : 2
    },
    {
      "token" : "を",
      "start_offset" : 11,
      "end_offset" : 12,
      "type" : "word",
      "position" : 3
    },
    {
      "token" : "買っ",
      "start_offset" : 12,
      "end_offset" : 14,
      "type" : "word",
      "position" : 4
    },
    {
      "token" : "た",
      "start_offset" : 14,
      "end_offset" : 15,
      "type" : "word",
      "position" : 5
    }
  ]
}

■解析結果（辞書なしの場合）

{
  "tokens" : [
    {
      "token" : "リバー",
      "start_offset" : 0,
      "end_offset" : 3,
      "type" : "word",
      "position" : 0
    },
    {
      "token" : "リバーシブルヨガマット",
      "start_offset" : 0,
      "end_offset" : 11,
      "type" : "word",
      "position" : 0,
      "positionLength" : 2
    },
    {
      "token" : "シブルヨガマット",
      "start_offset" : 3,
      "end_offset" : 11,
      "type" : "word",
      "position" : 1
    },
    {
      "token" : "を",
      "start_offset" : 11,
      "end_offset" : 12,
      "type" : "word",
      "position" : 2
    },
    {
      "token" : "買っ",
      "start_offset" : 12,
      "end_offset" : 14,
      "type" : "word",
      "position" : 3
    },
    {
      "token" : "た",
      "start_offset" : 14,
      "end_offset" : 15,
      "type" : "word",
      "position" : 4
    }
  ]
}