Elasticsearchで日本語の形態素解析を行おう。

56 min read読了の目安(約51200字

初めに

Elasticsearchのanalyzerを使用することで簡単に形態素解析を行うことが可能です。
今回は簡単な例を挙げながら、Elasticsearchっていいよね!!って布教して行こうと思います。

また、今回出てくるAPIリクエストは、Kibana の Dev Toolsで基本的に実行しています。

何も考えず形態素解析をしてみよう。

まず、indexを作成。( 日本語の形態素解析ができないのでダメダメという肩書を与えます。)

PUT /damedame_analyzer/

今回は全て以下の文章を形態素解析していきます。

私は、今年25歳になりました。昼は仕事をして夜はお酒を飲んでいます。趣味はバイクです。

以下を実行します。

GET /damedame_analyzer/_analyzer
{
    "analyzer": "standard",
    "explain" : true,
    "text": "私は、今年25歳になりました。昼は仕事をして夜はお酒を飲んでいます。趣味はバイクです。"
}
結果は、こちら

{
"detail" : {
"custom_analyzer" : false,
"analyzer" : {
"name" : "standard",
"tokens" : [
{
"token" : "私",
"start_offset" : 0,
"end_offset" : 1,
"type" : "<IDEOGRAPHIC>",
"position" : 0,
"bytes" : "[e7 a7 81]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 1,
"end_offset" : 2,
"type" : "<HIRAGANA>",
"position" : 1,
"bytes" : "[e3 81 af]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "今",
"start_offset" : 3,
"end_offset" : 4,
"type" : "<IDEOGRAPHIC>",
"position" : 2,
"bytes" : "[e4 bb 8a]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "年",
"start_offset" : 4,
"end_offset" : 5,
"type" : "<IDEOGRAPHIC>",
"position" : 3,
"bytes" : "[e5 b9 b4]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "25",
"start_offset" : 5,
"end_offset" : 7,
"type" : "<NUM>",
"position" : 4,
"bytes" : "[32 35]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "歳",
"start_offset" : 7,
"end_offset" : 8,
"type" : "<IDEOGRAPHIC>",
"position" : 5,
"bytes" : "[e6 ad b3]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "に",
"start_offset" : 8,
"end_offset" : 9,
"type" : "<HIRAGANA>",
"position" : 6,
"bytes" : "[e3 81 ab]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "な",
"start_offset" : 9,
"end_offset" : 10,
"type" : "<HIRAGANA>",
"position" : 7,
"bytes" : "[e3 81 aa]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "り",
"start_offset" : 10,
"end_offset" : 11,
"type" : "<HIRAGANA>",
"position" : 8,
"bytes" : "[e3 82 8a]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "ま",
"start_offset" : 11,
"end_offset" : 12,
"type" : "<HIRAGANA>",
"position" : 9,
"bytes" : "[e3 81 be]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "し",
"start_offset" : 12,
"end_offset" : 13,
"type" : "<HIRAGANA>",
"position" : 10,
"bytes" : "[e3 81 97]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "た",
"start_offset" : 13,
"end_offset" : 14,
"type" : "<HIRAGANA>",
"position" : 11,
"bytes" : "[e3 81 9f]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "昼",
"start_offset" : 15,
"end_offset" : 16,
"type" : "<IDEOGRAPHIC>",
"position" : 12,
"bytes" : "[e6 98 bc]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 16,
"end_offset" : 17,
"type" : "<HIRAGANA>",
"position" : 13,
"bytes" : "[e3 81 af]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "仕",
"start_offset" : 17,
"end_offset" : 18,
"type" : "<IDEOGRAPHIC>",
"position" : 14,
"bytes" : "[e4 bb 95]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "事",
"start_offset" : 18,
"end_offset" : 19,
"type" : "<IDEOGRAPHIC>",
"position" : 15,
"bytes" : "[e4 ba 8b]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "を",
"start_offset" : 19,
"end_offset" : 20,
"type" : "<HIRAGANA>",
"position" : 16,
"bytes" : "[e3 82 92]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "し",
"start_offset" : 20,
"end_offset" : 21,
"type" : "<HIRAGANA>",
"position" : 17,
"bytes" : "[e3 81 97]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "て",
"start_offset" : 21,
"end_offset" : 22,
"type" : "<HIRAGANA>",
"position" : 18,
"bytes" : "[e3 81 a6]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "夜",
"start_offset" : 22,
"end_offset" : 23,
"type" : "<IDEOGRAPHIC>",
"position" : 19,
"bytes" : "[e5 a4 9c]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 23,
"end_offset" : 24,
"type" : "<HIRAGANA>",
"position" : 20,
"bytes" : "[e3 81 af]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "お",
"start_offset" : 24,
"end_offset" : 25,
"type" : "<HIRAGANA>",
"position" : 21,
"bytes" : "[e3 81 8a]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "酒",
"start_offset" : 25,
"end_offset" : 26,
"type" : "<IDEOGRAPHIC>",
"position" : 22,
"bytes" : "[e9 85 92]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "を",
"start_offset" : 26,
"end_offset" : 27,
"type" : "<HIRAGANA>",
"position" : 23,
"bytes" : "[e3 82 92]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "飲",
"start_offset" : 27,
"end_offset" : 28,
"type" : "<IDEOGRAPHIC>",
"position" : 24,
"bytes" : "[e9 a3 b2]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "ん",
"start_offset" : 28,
"end_offset" : 29,
"type" : "<HIRAGANA>",
"position" : 25,
"bytes" : "[e3 82 93]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "で",
"start_offset" : 29,
"end_offset" : 30,
"type" : "<HIRAGANA>",
"position" : 26,
"bytes" : "[e3 81 a7]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "い",
"start_offset" : 30,
"end_offset" : 31,
"type" : "<HIRAGANA>",
"position" : 27,
"bytes" : "[e3 81 84]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "ま",
"start_offset" : 31,
"end_offset" : 32,
"type" : "<HIRAGANA>",
"position" : 28,
"bytes" : "[e3 81 be]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "す",
"start_offset" : 32,
"end_offset" : 33,
"type" : "<HIRAGANA>",
"position" : 29,
"bytes" : "[e3 81 99]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "趣",
"start_offset" : 34,
"end_offset" : 35,
"type" : "<IDEOGRAPHIC>",
"position" : 30,
"bytes" : "[e8 b6 a3]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "味",
"start_offset" : 35,
"end_offset" : 36,
"type" : "<IDEOGRAPHIC>",
"position" : 31,
"bytes" : "[e5 91 b3]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 36,
"end_offset" : 37,
"type" : "<HIRAGANA>",
"position" : 32,
"bytes" : "[e3 81 af]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "バイク",
"start_offset" : 37,
"end_offset" : 40,
"type" : "<KATAKANA>",
"position" : 33,
"bytes" : "[e3 83 90 e3 82 a4 e3 82 af]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "で",
"start_offset" : 40,
"end_offset" : 41,
"type" : "<HIRAGANA>",
"position" : 34,
"bytes" : "[e3 81 a7]",
"positionLength" : 1,
"termFrequency" : 1
},
{
"token" : "す",
"start_offset" : 41,
"end_offset" : 42,
"type" : "<HIRAGANA>",
"position" : 35,
"bytes" : "[e3 81 99]",
"positionLength" : 1,
"termFrequency" : 1
}
]
}
}
}

Elasticsearchは、標準では形態素解析を日本語では行えません。
( 全て一文字で解析されてしまいます。)
なので、 Elasticsearchのpluginである。kuromoji を利用します。
(その他、 sudachiなどがあります。)
kuromoji

kuromoji を使って形態素解析をしよう。

indexを新しく作成します。
ここでは、damedame_analyzerの時と違いsettingsでanalyzerをkuromojiに指定します。

PUT /iketeru_analyzer/
{
  "settings": {
    "index": {
      "analysis": {
        "analyzer": {
            "iketeru_analyzer" : {
              "tokenizer" : "kuromoji_tokenizer"
            }
          }
      }
    }
  }
}

では、同様に形態素解析しましょう。

GET /iketeru_analyzer/_analyze
{
    "analyzer": "iketeru_analyzer",
    "explain" : true,
    "text": "私は、今年25歳になりました。昼は仕事をして夜はお酒を飲んでいます。趣味はバイクです。"
}
結果はこちら

{
"detail" : {
"custom_analyzer" : true,
"charfilters" : [ ],
"tokenizer" : {
"name" : "kuromoji_tokenizer",
"tokens" : [
{
"token" : "私",
"start_offset" : 0,
"end_offset" : 1,
"type" : "word",
"position" : 0,
"baseForm" : null,
"bytes" : "[e7 a7 81]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-代名詞-一般",
"partOfSpeech (en)" : "noun-pronoun-misc",
"positionLength" : 1,
"pronunciation" : "ワタシ",
"pronunciation (en)" : "watashi",
"reading" : "ワタシ",
"reading (en)" : "watashi",
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 1,
"end_offset" : 2,
"type" : "word",
"position" : 1,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "今年",
"start_offset" : 3,
"end_offset" : 5,
"type" : "word",
"position" : 2,
"baseForm" : null,
"bytes" : "[e4 bb 8a e5 b9 b4]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-副詞可能",
"partOfSpeech (en)" : "noun-adverbial",
"positionLength" : 1,
"pronunciation" : "コトシ",
"pronunciation (en)" : "kotoshi",
"reading" : "コトシ",
"reading (en)" : "kotoshi",
"termFrequency" : 1
},
{
"token" : "25",
"start_offset" : 5,
"end_offset" : 7,
"type" : "word",
"position" : 3,
"baseForm" : null,
"bytes" : "[32 35]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-数",
"partOfSpeech (en)" : "noun-numeric",
"positionLength" : 1,
"pronunciation" : null,
"pronunciation (en)" : null,
"reading" : null,
"reading (en)" : null,
"termFrequency" : 1
},
{
"token" : "歳",
"start_offset" : 7,
"end_offset" : 8,
"type" : "word",
"position" : 4,
"baseForm" : null,
"bytes" : "[e6 ad b3]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-接尾-助数詞",
"partOfSpeech (en)" : "noun-suffix-classifier",
"positionLength" : 1,
"pronunciation" : "サイ",
"pronunciation (en)" : "sai",
"reading" : "サイ",
"reading (en)" : "sai",
"termFrequency" : 1
},
{
"token" : "に",
"start_offset" : 8,
"end_offset" : 9,
"type" : "word",
"position" : 5,
"baseForm" : null,
"bytes" : "[e3 81 ab]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-格助詞-一般",
"partOfSpeech (en)" : "particle-case-misc",
"positionLength" : 1,
"pronunciation" : "ニ",
"pronunciation (en)" : "ni",
"reading" : "ニ",
"reading (en)" : "ni",
"termFrequency" : 1
},
{
"token" : "なり",
"start_offset" : 9,
"end_offset" : 11,
"type" : "word",
"position" : 6,
"baseForm" : "なる",
"bytes" : "[e3 81 aa e3 82 8a]",
"inflectionForm" : "連用形",
"inflectionForm (en)" : "conjunctive",
"inflectionType" : "五段・ラ行",
"inflectionType (en)" : "5-row-cons-r",
"partOfSpeech" : "動詞-自立",
"partOfSpeech (en)" : "verb-main",
"positionLength" : 1,
"pronunciation" : "ナリ",
"pronunciation (en)" : "nari",
"reading" : "ナリ",
"reading (en)" : "nari",
"termFrequency" : 1
},
{
"token" : "まし",
"start_offset" : 11,
"end_offset" : 13,
"type" : "word",
"position" : 7,
"baseForm" : "ます",
"bytes" : "[e3 81 be e3 81 97]",
"inflectionForm" : "連用形",
"inflectionForm (en)" : "conjunctive",
"inflectionType" : "特殊・マス",
"inflectionType (en)" : "special-masu",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "マシ",
"pronunciation (en)" : "mashi",
"reading" : "マシ",
"reading (en)" : "mashi",
"termFrequency" : 1
},
{
"token" : "た",
"start_offset" : 13,
"end_offset" : 14,
"type" : "word",
"position" : 8,
"baseForm" : null,
"bytes" : "[e3 81 9f]",
"inflectionForm" : "基本形",
"inflectionForm (en)" : "base",
"inflectionType" : "特殊・タ",
"inflectionType (en)" : "special-da",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "タ",
"pronunciation (en)" : "ta",
"reading" : "タ",
"reading (en)" : "ta",
"termFrequency" : 1
},
{
"token" : "昼",
"start_offset" : 15,
"end_offset" : 16,
"type" : "word",
"position" : 9,
"baseForm" : null,
"bytes" : "[e6 98 bc]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-副詞可能",
"partOfSpeech (en)" : "noun-adverbial",
"positionLength" : 1,
"pronunciation" : "ヒル",
"pronunciation (en)" : "hiru",
"reading" : "ヒル",
"reading (en)" : "hiru",
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 16,
"end_offset" : 17,
"type" : "word",
"position" : 10,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "仕事",
"start_offset" : 17,
"end_offset" : 19,
"type" : "word",
"position" : 11,
"baseForm" : null,
"bytes" : "[e4 bb 95 e4 ba 8b]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-サ変接続",
"partOfSpeech (en)" : "noun-verbal",
"positionLength" : 1,
"pronunciation" : "シゴト",
"pronunciation (en)" : "shigoto",
"reading" : "シゴト",
"reading (en)" : "shigoto",
"termFrequency" : 1
},
{
"token" : "を",
"start_offset" : 19,
"end_offset" : 20,
"type" : "word",
"position" : 12,
"baseForm" : null,
"bytes" : "[e3 82 92]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-格助詞-一般",
"partOfSpeech (en)" : "particle-case-misc",
"positionLength" : 1,
"pronunciation" : "ヲ",
"pronunciation (en)" : "o",
"reading" : "ヲ",
"reading (en)" : "o",
"termFrequency" : 1
},
{
"token" : "し",
"start_offset" : 20,
"end_offset" : 21,
"type" : "word",
"position" : 13,
"baseForm" : "する",
"bytes" : "[e3 81 97]",
"inflectionForm" : "連用形",
"inflectionForm (en)" : "conjunctive",
"inflectionType" : "サ変・スル",
"inflectionType (en)" : "irregular-suru",
"partOfSpeech" : "動詞-自立",
"partOfSpeech (en)" : "verb-main",
"positionLength" : 1,
"pronunciation" : "シ",
"pronunciation (en)" : "shi",
"reading" : "シ",
"reading (en)" : "shi",
"termFrequency" : 1
},
{
"token" : "て",
"start_offset" : 21,
"end_offset" : 22,
"type" : "word",
"position" : 14,
"baseForm" : null,
"bytes" : "[e3 81 a6]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-接続助詞",
"partOfSpeech (en)" : "particle-conjunctive",
"positionLength" : 1,
"pronunciation" : "テ",
"pronunciation (en)" : "te",
"reading" : "テ",
"reading (en)" : "te",
"termFrequency" : 1
},
{
"token" : "夜",
"start_offset" : 22,
"end_offset" : 23,
"type" : "word",
"position" : 15,
"baseForm" : null,
"bytes" : "[e5 a4 9c]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-副詞可能",
"partOfSpeech (en)" : "noun-adverbial",
"positionLength" : 1,
"pronunciation" : "ヨル",
"pronunciation (en)" : "yoru",
"reading" : "ヨル",
"reading (en)" : "yoru",
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 23,
"end_offset" : 24,
"type" : "word",
"position" : 16,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "お",
"start_offset" : 24,
"end_offset" : 25,
"type" : "word",
"position" : 17,
"baseForm" : null,
"bytes" : "[e3 81 8a]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "接頭詞-名詞接続",
"partOfSpeech (en)" : "prefix-nominal",
"positionLength" : 1,
"pronunciation" : "オ",
"pronunciation (en)" : "o",
"reading" : "オ",
"reading (en)" : "o",
"termFrequency" : 1
},
{
"token" : "酒",
"start_offset" : 25,
"end_offset" : 26,
"type" : "word",
"position" : 18,
"baseForm" : null,
"bytes" : "[e9 85 92]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-一般",
"partOfSpeech (en)" : "noun-common",
"positionLength" : 1,
"pronunciation" : "サケ",
"pronunciation (en)" : "sake",
"reading" : "サケ",
"reading (en)" : "sake",
"termFrequency" : 1
},
{
"token" : "を",
"start_offset" : 26,
"end_offset" : 27,
"type" : "word",
"position" : 19,
"baseForm" : null,
"bytes" : "[e3 82 92]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-格助詞-一般",
"partOfSpeech (en)" : "particle-case-misc",
"positionLength" : 1,
"pronunciation" : "ヲ",
"pronunciation (en)" : "o",
"reading" : "ヲ",
"reading (en)" : "o",
"termFrequency" : 1
},
{
"token" : "飲ん",
"start_offset" : 27,
"end_offset" : 29,
"type" : "word",
"position" : 20,
"baseForm" : "飲む",
"bytes" : "[e9 a3 b2 e3 82 93]",
"inflectionForm" : "連用タ接続",
"inflectionForm (en)" : "conjunctive-ta-connection",
"inflectionType" : "五段・マ行",
"inflectionType (en)" : "5-row-cons-m",
"partOfSpeech" : "動詞-自立",
"partOfSpeech (en)" : "verb-main",
"positionLength" : 1,
"pronunciation" : "ノン",
"pronunciation (en)" : "non",
"reading" : "ノン",
"reading (en)" : "non",
"termFrequency" : 1
},
{
"token" : "で",
"start_offset" : 29,
"end_offset" : 30,
"type" : "word",
"position" : 21,
"baseForm" : null,
"bytes" : "[e3 81 a7]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-接続助詞",
"partOfSpeech (en)" : "particle-conjunctive",
"positionLength" : 1,
"pronunciation" : "デ",
"pronunciation (en)" : "de",
"reading" : "デ",
"reading (en)" : "de",
"termFrequency" : 1
},
{
"token" : "い",
"start_offset" : 30,
"end_offset" : 31,
"type" : "word",
"position" : 22,
"baseForm" : "いる",
"bytes" : "[e3 81 84]",
"inflectionForm" : "連用形",
"inflectionForm (en)" : "conjunctive",
"inflectionType" : "一段",
"inflectionType (en)" : "1-row",
"partOfSpeech" : "動詞-非自立",
"partOfSpeech (en)" : "verb-auxiliary",
"positionLength" : 1,
"pronunciation" : "イ",
"pronunciation (en)" : "i",
"reading" : "イ",
"reading (en)" : "i",
"termFrequency" : 1
},
{
"token" : "ます",
"start_offset" : 31,
"end_offset" : 33,
"type" : "word",
"position" : 23,
"baseForm" : null,
"bytes" : "[e3 81 be e3 81 99]",
"inflectionForm" : "基本形",
"inflectionForm (en)" : "base",
"inflectionType" : "特殊・マス",
"inflectionType (en)" : "special-masu",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "マス",
"pronunciation (en)" : "masu",
"reading" : "マス",
"reading (en)" : "masu",
"termFrequency" : 1
},
{
"token" : "趣味",
"start_offset" : 34,
"end_offset" : 36,
"type" : "word",
"position" : 24,
"baseForm" : null,
"bytes" : "[e8 b6 a3 e5 91 b3]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-一般",
"partOfSpeech (en)" : "noun-common",
"positionLength" : 1,
"pronunciation" : "シュミ",
"pronunciation (en)" : "shumi",
"reading" : "シュミ",
"reading (en)" : "shumi",
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 36,
"end_offset" : 37,
"type" : "word",
"position" : 25,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "バイク",
"start_offset" : 37,
"end_offset" : 40,
"type" : "word",
"position" : 26,
"baseForm" : null,
"bytes" : "[e3 83 90 e3 82 a4 e3 82 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-一般",
"partOfSpeech (en)" : "noun-common",
"positionLength" : 1,
"pronunciation" : "バイク",
"pronunciation (en)" : "baiku",
"reading" : "バイク",
"reading (en)" : "baiku",
"termFrequency" : 1
},
{
"token" : "です",
"start_offset" : 40,
"end_offset" : 42,
"type" : "word",
"position" : 27,
"baseForm" : null,
"bytes" : "[e3 81 a7 e3 81 99]",
"inflectionForm" : "基本形",
"inflectionForm (en)" : "base",
"inflectionType" : "特殊・デス",
"inflectionType (en)" : "special-desu",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "デス",
"pronunciation (en)" : "desu",
"reading" : "デス",
"reading (en)" : "desu",
"termFrequency" : 1
}
]
},
"tokenfilters" : [ ]
}
}

結果を見るとわかりますが、標準で行った時と違い。
きちんと形態素分析されている事がわかります。
また、読み仮名や品詞情報、動詞の原形などもわかるので文章内のそれぞれの単語を分類分けするのも
比較的簡単となります。(例 : 名詞だけ抽出、動詞の原形を抽出など )

もっと便利にしたい。

上記までで日本語の形態素分析が行える事がわかりました。
その上、品詞もわかって+動詞の原型もわかるのでだいぶ正規化された気がします。
もっと正規化したいです。
例えば、175㌢ => 175センチ (変な文字使うな)
    CoMpuTer => computer (小文字で統一したい) などです。

そういったこともElasticsearchで可能です。

例文を変えます。
例文 : 私は老眼なのでCoMpuTerから50㌢は離れないと見えない。

先ほどので形態素解析をしてみましょう。

GET /iketeru_analyzer/_analyze
{
    "analyzer": "iketeru_analyzer",
    "explain" : true,
    "text": "私は老眼なのでCoMpuTerから50㌢は離れないと見えない。"
}
結果はこちら

{
"detail" : {
"custom_analyzer" : true,
"charfilters" : [ ],
"tokenizer" : {
"name" : "kuromoji_tokenizer",
"tokens" : [
{
"token" : "私",
"start_offset" : 0,
"end_offset" : 1,
"type" : "word",
"position" : 0,
"baseForm" : null,
"bytes" : "[e7 a7 81]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-代名詞-一般",
"partOfSpeech (en)" : "noun-pronoun-misc",
"positionLength" : 1,
"pronunciation" : "ワタシ",
"pronunciation (en)" : "watashi",
"reading" : "ワタシ",
"reading (en)" : "watashi",
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 1,
"end_offset" : 2,
"type" : "word",
"position" : 1,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "老眼",
"start_offset" : 2,
"end_offset" : 4,
"type" : "word",
"position" : 2,
"baseForm" : null,
"bytes" : "[e8 80 81 e7 9c bc]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-一般",
"partOfSpeech (en)" : "noun-common",
"positionLength" : 1,
"pronunciation" : "ローガン",
"pronunciation (en)" : "rogan",
"reading" : "ロウガン",
"reading (en)" : "rōgan",
"termFrequency" : 1
},
{
"token" : "な",
"start_offset" : 4,
"end_offset" : 5,
"type" : "word",
"position" : 3,
"baseForm" : "だ",
"bytes" : "[e3 81 aa]",
"inflectionForm" : "体言接続",
"inflectionForm (en)" : "uninflected-connection",
"inflectionType" : "特殊・ダ",
"inflectionType (en)" : "special-ta",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "ナ",
"pronunciation (en)" : "na",
"reading" : "ナ",
"reading (en)" : "na",
"termFrequency" : 1
},
{
"token" : "ので",
"start_offset" : 5,
"end_offset" : 7,
"type" : "word",
"position" : 4,
"baseForm" : null,
"bytes" : "[e3 81 ae e3 81 a7]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-接続助詞",
"partOfSpeech (en)" : "particle-conjunctive",
"positionLength" : 1,
"pronunciation" : "ノデ",
"pronunciation (en)" : "node",
"reading" : "ノデ",
"reading (en)" : "node",
"termFrequency" : 1
},
{
"token" : "CoMpuTer",
"start_offset" : 7,
"end_offset" : 15,
"type" : "word",
"position" : 5,
"baseForm" : null,
"bytes" : "[43 6f 4d 70 75 54 65 72]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-一般",
"partOfSpeech (en)" : "noun-common",
"positionLength" : 1,
"pronunciation" : null,
"pronunciation (en)" : null,
"reading" : null,
"reading (en)" : null,
"termFrequency" : 1
},
{
"token" : "から",
"start_offset" : 15,
"end_offset" : 17,
"type" : "word",
"position" : 6,
"baseForm" : null,
"bytes" : "[e3 81 8b e3 82 89]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-格助詞-一般",
"partOfSpeech (en)" : "particle-case-misc",
"positionLength" : 1,
"pronunciation" : "カラ",
"pronunciation (en)" : "kara",
"reading" : "カラ",
"reading (en)" : "kara",
"termFrequency" : 1
},
{
"token" : "50",
"start_offset" : 17,
"end_offset" : 19,
"type" : "word",
"position" : 7,
"baseForm" : null,
"bytes" : "[35 30]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-数",
"partOfSpeech (en)" : "noun-numeric",
"positionLength" : 1,
"pronunciation" : null,
"pronunciation (en)" : null,
"reading" : null,
"reading (en)" : null,
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 20,
"end_offset" : 21,
"type" : "word",
"position" : 8,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "離れ",
"start_offset" : 21,
"end_offset" : 23,
"type" : "word",
"position" : 9,
"baseForm" : "離れる",
"bytes" : "[e9 9b a2 e3 82 8c]",
"inflectionForm" : "未然形",
"inflectionForm (en)" : "imperfective",
"inflectionType" : "一段",
"inflectionType (en)" : "1-row",
"partOfSpeech" : "動詞-自立",
"partOfSpeech (en)" : "verb-main",
"positionLength" : 1,
"pronunciation" : "ハナレ",
"pronunciation (en)" : "hanare",
"reading" : "ハナレ",
"reading (en)" : "hanare",
"termFrequency" : 1
},
{
"token" : "ない",
"start_offset" : 23,
"end_offset" : 25,
"type" : "word",
"position" : 10,
"baseForm" : null,
"bytes" : "[e3 81 aa e3 81 84]",
"inflectionForm" : "基本形",
"inflectionForm (en)" : "base",
"inflectionType" : "特殊・ナイ",
"inflectionType (en)" : "special-nai",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "ナイ",
"pronunciation (en)" : "nai",
"reading" : "ナイ",
"reading (en)" : "nai",
"termFrequency" : 1
},
{
"token" : "と",
"start_offset" : 25,
"end_offset" : 26,
"type" : "word",
"position" : 11,
"baseForm" : null,
"bytes" : "[e3 81 a8]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-格助詞-引用",
"partOfSpeech (en)" : "particle-case-quote",
"positionLength" : 1,
"pronunciation" : "ト",
"pronunciation (en)" : "to",
"reading" : "ト",
"reading (en)" : "to",
"termFrequency" : 1
},
{
"token" : "見え",
"start_offset" : 26,
"end_offset" : 28,
"type" : "word",
"position" : 12,
"baseForm" : "見える",
"bytes" : "[e8 a6 8b e3 81 88]",
"inflectionForm" : "未然形",
"inflectionForm (en)" : "imperfective",
"inflectionType" : "一段",
"inflectionType (en)" : "1-row",
"partOfSpeech" : "動詞-自立",
"partOfSpeech (en)" : "verb-main",
"positionLength" : 1,
"pronunciation" : "ミエ",
"pronunciation (en)" : "mie",
"reading" : "ミエ",
"reading (en)" : "mie",
"termFrequency" : 1
},
{
"token" : "ない",
"start_offset" : 28,
"end_offset" : 30,
"type" : "word",
"position" : 13,
"baseForm" : null,
"bytes" : "[e3 81 aa e3 81 84]",
"inflectionForm" : "基本形",
"inflectionForm (en)" : "base",
"inflectionType" : "特殊・ナイ",
"inflectionType (en)" : "special-nai",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "ナイ",
"pronunciation (en)" : "nai",
"reading" : "ナイ",
"reading (en)" : "nai",
"termFrequency" : 1
}
]
},
"tokenfilters" : [ ]
}
}

結果を見ると、㌢(環境依存文字)は無視され、CoMpuTerは、そのままlowercaseなどされることなく出力されています。
このままだと、データとしては使いづらいかもです。

そこで、Elasticsearchのchar_filterを利用します。

char_filterとは、今まで行っていたtokenizerの前に文字列を正規化( 前処理 )するものです。
イメージは、以下です。
( また、tokenizerで分割後の要素を正規化する token_filterというのも存在しますが割愛します。 )

では、実際にindex作成して実行します。


PUT /motto_iketeru_analyzer/
{
  "settings": {
    "index": {
      "analysis": {
        "analyzer": {
            "motto_iketeru_analyzer" : {
              "tokenizer" : "kuromoji_tokenizer",
              "char_filter" :  [ "icu_normalizer" ] 
            }
          }
      }
    }
  }
}
GET /motto_iketeru_analyzer/_analyze
{
    "analyzer": "motto_iketeru_analyzer", 
    "explain" : true,
    "text": "私は老眼なのでCoMpuTerから50㌢は離れないと見えない。"
}

結果はこちら

{
"detail" : {
"custom_analyzer" : true,
"charfilters" : [
{
"name" : "icu_normalizer",
"filtered_text" : [
"私は老眼なので"
]
}
],
"tokenizer" : {
"name" : "kuromoji_tokenizer",
"tokens" : [
{
"token" : "私",
"start_offset" : 0,
"end_offset" : 1,
"type" : "word",
"position" : 0,
"baseForm" : null,
"bytes" : "[e7 a7 81]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-代名詞-一般",
"partOfSpeech (en)" : "noun-pronoun-misc",
"positionLength" : 1,
"pronunciation" : "ワタシ",
"pronunciation (en)" : "watashi",
"reading" : "ワタシ",
"reading (en)" : "watashi",
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 1,
"end_offset" : 2,
"type" : "word",
"position" : 1,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "老眼",
"start_offset" : 2,
"end_offset" : 4,
"type" : "word",
"position" : 2,
"baseForm" : null,
"bytes" : "[e8 80 81 e7 9c bc]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-一般",
"partOfSpeech (en)" : "noun-common",
"positionLength" : 1,
"pronunciation" : "ローガン",
"pronunciation (en)" : "rogan",
"reading" : "ロウガン",
"reading (en)" : "rōgan",
"termFrequency" : 1
},
{
"token" : "な",
"start_offset" : 4,
"end_offset" : 5,
"type" : "word",
"position" : 3,
"baseForm" : "だ",
"bytes" : "[e3 81 aa]",
"inflectionForm" : "体言接続",
"inflectionForm (en)" : "uninflected-connection",
"inflectionType" : "特殊・ダ",
"inflectionType (en)" : "special-ta",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "ナ",
"pronunciation (en)" : "na",
"reading" : "ナ",
"reading (en)" : "na",
"termFrequency" : 1
},
{
"token" : "ので",
"start_offset" : 5,
"end_offset" : 7,
"type" : "word",
"position" : 4,
"baseForm" : null,
"bytes" : "[e3 81 ae e3 81 a7]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-接続助詞",
"partOfSpeech (en)" : "particle-conjunctive",
"positionLength" : 1,
"pronunciation" : "ノデ",
"pronunciation (en)" : "node",
"reading" : "ノデ",
"reading (en)" : "node",
"termFrequency" : 1
},
{
"token" : "computer",
"start_offset" : 7,
"end_offset" : 15,
"type" : "word",
"position" : 5,
"baseForm" : null,
"bytes" : "[63 6f 6d 70 75 74 65 72]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-一般",
"partOfSpeech (en)" : "noun-common",
"positionLength" : 1,
"pronunciation" : null,
"pronunciation (en)" : null,
"reading" : null,
"reading (en)" : null,
"termFrequency" : 1
},
{
"token" : "から",
"start_offset" : 15,
"end_offset" : 17,
"type" : "word",
"position" : 6,
"baseForm" : null,
"bytes" : "[e3 81 8b e3 82 89]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-格助詞-一般",
"partOfSpeech (en)" : "particle-case-misc",
"positionLength" : 1,
"pronunciation" : "カラ",
"pronunciation (en)" : "kara",
"reading" : "カラ",
"reading (en)" : "kara",
"termFrequency" : 1
},
{
"token" : "50",
"start_offset" : 17,
"end_offset" : 19,
"type" : "word",
"position" : 7,
"baseForm" : null,
"bytes" : "[35 30]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-数",
"partOfSpeech (en)" : "noun-numeric",
"positionLength" : 1,
"pronunciation" : null,
"pronunciation (en)" : null,
"reading" : null,
"reading (en)" : null,
"termFrequency" : 1
},
{
"token" : "センチ",
"start_offset" : 19,
"end_offset" : 20,
"type" : "word",
"position" : 8,
"baseForm" : null,
"bytes" : "[e3 82 bb e3 83 b3 e3 83 81]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "名詞-接尾-助数詞",
"partOfSpeech (en)" : "noun-suffix-classifier",
"positionLength" : 1,
"pronunciation" : "センチ",
"pronunciation (en)" : "senchi",
"reading" : "センチ",
"reading (en)" : "senchi",
"termFrequency" : 1
},
{
"token" : "は",
"start_offset" : 20,
"end_offset" : 21,
"type" : "word",
"position" : 9,
"baseForm" : null,
"bytes" : "[e3 81 af]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-係助詞",
"partOfSpeech (en)" : "particle-dependency",
"positionLength" : 1,
"pronunciation" : "ワ",
"pronunciation (en)" : "wa",
"reading" : "ハ",
"reading (en)" : "ha",
"termFrequency" : 1
},
{
"token" : "離れ",
"start_offset" : 21,
"end_offset" : 23,
"type" : "word",
"position" : 10,
"baseForm" : "離れる",
"bytes" : "[e9 9b a2 e3 82 8c]",
"inflectionForm" : "未然形",
"inflectionForm (en)" : "imperfective",
"inflectionType" : "一段",
"inflectionType (en)" : "1-row",
"partOfSpeech" : "動詞-自立",
"partOfSpeech (en)" : "verb-main",
"positionLength" : 1,
"pronunciation" : "ハナレ",
"pronunciation (en)" : "hanare",
"reading" : "ハナレ",
"reading (en)" : "hanare",
"termFrequency" : 1
},
{
"token" : "ない",
"start_offset" : 23,
"end_offset" : 25,
"type" : "word",
"position" : 11,
"baseForm" : null,
"bytes" : "[e3 81 aa e3 81 84]",
"inflectionForm" : "基本形",
"inflectionForm (en)" : "base",
"inflectionType" : "特殊・ナイ",
"inflectionType (en)" : "special-nai",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "ナイ",
"pronunciation (en)" : "nai",
"reading" : "ナイ",
"reading (en)" : "nai",
"termFrequency" : 1
},
{
"token" : "と",
"start_offset" : 25,
"end_offset" : 26,
"type" : "word",
"position" : 12,
"baseForm" : null,
"bytes" : "[e3 81 a8]",
"inflectionForm" : null,
"inflectionForm (en)" : null,
"inflectionType" : null,
"inflectionType (en)" : null,
"partOfSpeech" : "助詞-格助詞-引用",
"partOfSpeech (en)" : "particle-case-quote",
"positionLength" : 1,
"pronunciation" : "ト",
"pronunciation (en)" : "to",
"reading" : "ト",
"reading (en)" : "to",
"termFrequency" : 1
},
{
"token" : "見え",
"start_offset" : 26,
"end_offset" : 28,
"type" : "word",
"position" : 13,
"baseForm" : "見える",
"bytes" : "[e8 a6 8b e3 81 88]",
"inflectionForm" : "未然形",
"inflectionForm (en)" : "imperfective",
"inflectionType" : "一段",
"inflectionType (en)" : "1-row",
"partOfSpeech" : "動詞-自立",
"partOfSpeech (en)" : "verb-main",
"positionLength" : 1,
"pronunciation" : "ミエ",
"pronunciation (en)" : "mie",
"reading" : "ミエ",
"reading (en)" : "mie",
"termFrequency" : 1
},
{
"token" : "ない",
"start_offset" : 28,
"end_offset" : 30,
"type" : "word",
"position" : 14,
"baseForm" : null,
"bytes" : "[e3 81 aa e3 81 84]",
"inflectionForm" : "基本形",
"inflectionForm (en)" : "base",
"inflectionType" : "特殊・ナイ",
"inflectionType (en)" : "special-nai",
"partOfSpeech" : "助動詞",
"partOfSpeech (en)" : "auxiliary-verb",
"positionLength" : 1,
"pronunciation" : "ナイ",
"pronunciation (en)" : "nai",
"reading" : "ナイ",
"reading (en)" : "nai",
"termFrequency" : 1
}
]
},
"tokenfilters" : [ ]
}
}

結果を見ると、
CoMpuTerは、computerに変換され、㌢は、センチへ変換され
より使いやすいデータが取り出せるようになりました。

最後に

Elasticsearchのanalyzerは簡易に形態素解析をする際にとても便利ですが奥が深く結構大変です。
もっと勉強しないとダメですね😢