elastic-muto

Easy expressive search queries for Elasticsearch with customisation! Build complicated elasticsearch queries without having to use the DSL. Expressions get compiled into native Elasticsearch queries, offering the same performance as if it had been hand coded.

elastic-muto is built using PEG.js. If you are curious about how the parsing works, check this out. The parser was originally developed for parsing filter conditions for the GET score API of Boolean.

Check out the API reference documentation.

Note: The library includes TypeScript definitions for a superior development experience.

Elasticsearch compatibility

elastic-muto can be used with elasticsearch v2.x and above.

Install

npm install elastic-muto --save

Usage

// Import the library
const muto = require('elastic-muto');

// muto.parse returns an elastic-builder BoolQuery object
const qry = muto.parse('["elasticsearch"] == "awesome" and ["unicorn"] exists');
qry.toJSON();
{
  "bool": {
    "must": [
      {
        "term": { "elasticsearch.keyword": "awesome" }
      },
      {
        "exists": { "field": "unicorn" }
      }
    ]
  }
}

Classes have also been provided for building the where expressions. Use whatever floats your boat 😉.

const qry = muto.parse(
    muto.where(muto.cn('elasticsearch').eq('awesome'))
        .and(muto.cn('unicorn').exists())
);

elastic-muto uses debug with the namespace elastic-muto. To enable debug logs, refer this.

Where Conditions

Where conditions can either be single(ex: '["key"] == value') or multiple. Multiple conditions can be combined with and/or.

Supported data types:

Condition types:

Both and, or cannot be used in the same level, because if you do, the desired query is not clear.

it('throws error if both and, or are called', () => {
    expect(
        () => muto.where()
            .and(muto.cn('anime').notContains('fillers'))
            .or(muto.cn('elasticsearch').eq('awesome'))
    ).toThrowError('Illegal operation! Join types cannot be mixed!');
});

Expressions can be nested using paranthesis. This allows to use both and, or:

const qry = muto.parse(
    '["elasticsearch"] == "awesome" and ["language"] == "node.js"' +
        'and (["library"] == "elastic-muto" or ["library"] == "elastic-builder")'
)

Elasticsearch Mapping

elastic-muto makes some assumptions for the mapping of data types. Following are the recommended mappings:

• String mapping:

  {
  "type": "text",
    "fields": {
        "keyword": {
        "type": "keyword",
        "ignore_above": 256
      }
    }
  }
  

  This is the default since elasticsearch v5.x

• Date mapping

  {
    "type": "date",
    "format": "strict_date_time_no_millis||strict_date_optional_time||epoch_millis"
  }
  

• Number mapping

  { "type" : "double" }
  

• Boolean mapping

  { "type": "boolean" }
  

If your mapping doesn't match, you might need to tweak the elasticsearch query generated with customisation.

Customisation

Elasticsearch queries generated by elastic-muto can be customised. Read more here. Check out a contrived example here.

REPL

Try it out on the command line using the node REPL:

# Start the repl
node ./node_modules/elastic-muto/repl.js
# Use the library loaded in context as `muto`
elastic-muto > muto.prettyPrint('["elasticsearch"] == "awesome" and ["unicorn"] exists')

API Reference

API reference can be accessed here - http://muto.js.org/docs.

API documentation was generated using documentation.js. It is being hosted with help from this awesome project - https://github.com/js-org/dns.js.org

Tests

Run unit tests:

npm test

The parser is tested extensively with upto 5 levels of nested queries!

Related

• elastic-builder - An elasticsearch query body builder for node.js
• FiltrES.js - A simple, safe, ElasticSearch Query compiler

License

MIT