Elasticsearch教程之五：AWS OpenSearch Node.js客户端的使用

Elasticsearch/OpenSearch可以说是企业级搜索应用的最佳选择。相较于传统的关系型数据库，Elasticsearch/OpenSearch更适合于存储海量数据并对其进行全文检索和数据分析（聚合）。下面介绍如何在Node.js中使用AWS OpenSearch。

Elasticsearch教程

Elasticsearch系列教程

• Elasticsearch教程之一：介绍
• Elasticsearch教程之二：索引，分词及映射
• Elasticsearch教程之三：API
• Elasticsearch教程之四：AWS OpenSearch
• Elasticsearch教程之五：AWS OpenSearch Node.js客户端的使用
• Elasticsearch教程之六：Python客户端的使用
• Elasticsearch教程之七：在OpenSearch中使用聚合实现Facet

安装

bash

npm install @opensearch-project/opensearch
npm install dotenv

关于如何使用Node.js调用OpenSearch API，请参考官网文档。下面介绍总结一个常用的功能。例子基本上都是在官网示例的基础上修改的。

创建索引

Javascript

const { Client } = require("@opensearch-project/opensearch");

let client = new Client({
  node: 'https://xxxxxx.eu-west-1.es.amazonaws.com'
});

const test_create_index = async () => {
  let settings = {
    settings: {
      index: {
        number_of_shards: 1,
        number_of_replicas: 1,
      },
    },
  };

  let response = await client.indices.create({
      index: "demoindex",
      body: settings,
  });

  console.log("Creating index:");
  console.log(response.body);
}

test_create_index();

获取索引列表

JavaScript

async getIndices() {
  return client.cat.indices({ format: 'json' });
},

只有当索引不存在时创建

JavaScript

async createIndexIfNotExist(indexName: string, settings: any) {
  const indices = (await this.getIndices()).body;
  if (indices) {
    for (let i = 0; i < indices.length; i += 1) {
      if (indices[i].index === indexName) {
        return indices[i];
      }
    }
  }
  return this.createIndex(indexName, settings);
},
async createIndex(indexName: string, settings: any) {
  const response = await client.indices.create({
    index: indexName,
    body: settings,
  });
  return response.body;
},

更新mapping

JavaScript

async updateMappings(indexName: string, mappings: any) {
  await client.indices.putMapping({
    index: indexName,
    body: mappings,
  });
},

传递过来的mapping类似于：

JavaScript

{
  "properties": {
    "author": {
      "type": "keyword"
    },
    "address": {
      "type": "text"
    },
  }
}

添加文档

JavaScript

async addDoc(indexName: string, id: string, doc: any) {
  const response = await client.index({
    id,
    index: indexName,
    body: doc,
    refresh: true,
  });
  return response.body;
},

删除索引/文档

删除索引

Javascript

let response = await client.indices.delete({
  index: index_name,
});

删除文档

Javascript

let response = await client.delete({
  index: index_name,
  id: id,
});

查询

基本查询

Javascript

const { Client } = require("@opensearch-project/opensearch");

let client = new Client({
  node: 'https://xxxxxx.eu-west-1.es.amazonaws.com'
});

const test_search = async () => {
  let query = {
    query: {
      match: {
        name: {
          query: "Harry",
        },
      },
    },
  };

  let response = await client.search({
    index: "demoindex",
    body: query,
  });

  console.log("Searching:");
  console.log(response.body);
}

test_search();

使用slop

Query可以使用OS中的各种关键字，比如：

JavaScript

let query = {
  query: {
    match: {
      name: {
        query: "food with cheese",
        slop: 10
      },
    },
  },
}

按照关键字查找

JavaScript

let query = {
  query: {
    match: {
      name: {
        query: "+(dessert | cake) -onion (cherry | stawberry)",
        default_field: "desc",
      },
    },
  },
}

进行term查询

JavaScript

query: {
  term: {
    age: 25
  }
}

JavaScript

query: {
  range: {
    age: {
      gte: 20,
      lte: 30
    }
  }
}

组合查询

JavaScript

query: {
  bool: {
    filter: [{range: {age: {ate: 18}}}],
    must: [{match_phrasae: {categories: 'Math'}}],
    should: [{match: {title: 'undergrade'}}],
    must_not: [{match: {content: 'gcse'}}],
  }
}

返回文档

返回文档的格式是这样的：

Javascript

[                                                                   
  {                                                                 
    _index: 'demoindex',                                            
    _type: '_doc',                                                  
    _id: '101',                                                     
    _score: 0.2876821,                                              
    _source: { name: 'Harry Potter', age: '12', school: 'Hogwarts' }
  }                                                                 
]                                                                   

Aggregation聚合

Javascript

const { Client } = require("@opensearch-project/opensearch");

let client = new Client({
  node: 'https://xxxxxx.eu-west-1.es.amazonaws.com'
});

const test_aggs = async () => {
  let query = {
    query: {
      aggs: {
        demo: {
          avg: { // avg, min, max, stats
            field: "age"
          }
        }
      }
    },
  };

  let response = await client.search({
    index: "demoindex",
    body: query,
  });

  console.log("Searching:");
  console.log(response.body);
}

test_search();

或者：

JavaScript

aggs: {
  demo: {
    histogram: { 
      field: "age",
      interval: 1
    }
  }
}