es-base.ts

// https://qiita.com/shin_hayata/items/41c07923dbf58f13eec4
// http://319ring.net/blog/archives/2906/
// https://medium.com/hello-elasticsearch/elasticsearch-833a0704e44b
// http://blog.shibayu36.org/entry/2016/08/31/110000
// https://gist.github.com/karmi/4526141

export const settings = {
  index: {
    analysis: {
      tokenizer: {
        ja_text_tokenizer: {
          type: "kuromoji_tokenizer",
          mode: "search",
        },
        roman_text_tokenizer: {
          type: "ngram",
          min_gram: 2,
          max_gram: 3,
        },
      },
      analyzer: {
        original_form_analyzer: {
          type: "custom",
          tokenizer: "ja_text_tokenizer",
          filter: ["kuromoji_stemmer"],
        },
        normalize_form_analyzer: {
          type: "custom",
          tokenizer: "ja_text_tokenizer",
          char_filter: ["icu_normalizer", "kuromoji_iteration_mark"],
          filter: ["kuromoji_baseform", "kuromoji_part_of_speech", "ja_stop", "kuromoji_number", "kuromoji_stemmer"],
        },
        reading_form_analyzer: {
          type: "custom",
          tokenizer: "ja_text_tokenizer",
          char_filter: ["icu_normalizer", "kuromoji_iteration_mark"],
          filter: ["kuromoji_part_of_speech", "ja_stop", "kuromoji_readingform", "kuromoji_stemmer"],
        },
        roman_analyzer: {
          tokenizer: "roman_text_tokenizer",
        },
      },
    },
  },
};

export const integer = { type: "integer" };
export const float = { type: "float" };
export const boolean = { type: "boolean" };
export const date = { type: "date" };
export const keyword = { type: "keyword" };
export const nested = { type: "nested" };
export const text = {
  type: "text",
  analyzer: "original_form_analyzer",
  fields: {
    strict: {
      type: "keyword",
    },
    normalized: {
      type: "text",
      analyzer: "normalize_form_analyzer",
    },
    kana: {
      type: "text",
      analyzer: "reading_form_analyzer",
    },
  },
};
export const idtext = {
  type: "keyword",
  fields: {
    fuzzy: {
      type: "text",
      analyzer: "roman_analyzer",
    },
  },
};

export const extractCharRefs = (str: string) =>
  str
  .replace(/"/g, '"')
  .replace(/&lt;/g, '<')
  .replace(/&gt;/g, '>')
  .replace(/&amp;/g, '&')
  ;