elastic-orm
v0.7.0
Published
Elasticsearch object relational mapper.
Downloads
93
Maintainers
Readme
ElasticORM
ElasticORM is an ORM tool that can run in javascript and helps developers to define a mapping of entities in one place by using decorators.
Installation
Install package from npm.
npm i elastic-orm @elastic/elasticsearch
Usage
1. Entity Definition
import { EsEntity, EsProperty, EsId } from 'elastic-orm';
@EsEntity('elastic_index') // specify elastic index
export class MyEntity {
@EsId()
public id: string;
@EsProperty('integer')
public foo: number;
@EsProperty({ type: 'keyword' })
public bar: string;
@EsProperty('boolean')
public fingersCrossedForUkraine: boolean = true;
@EsProperty('date')
public createdAt: Date;
}
2. Client and Repository Configuration
import { EsRepository } from 'elastic-orm';
import { Client } from '@elastic/elasticsearch';
const repository = new EsRepository(
MyEntity,
new Client(/* client configuration */),
);
Client configuration is provided here https://www.elastic.co/guide/en/elasticsearch/client/javascript-api/7.17/client-connecting.html
3. Create Mapping
You can easily create index from you entity definition by running:
import { FactoryProvider } from 'elastic-orm';
const schema =
FactoryProvider.makeSchemaManager().generateIndexSchema(MyEntity);
await repository.createIndex(schema);
Storing and Loading
const myEntity = new MyEntity()
myEntity.foo = 99
myEntity.bar = 'hello elastic search'
myEntity.isTutorial = true
myEntity.createdAt = new Date()
const createdEntity = await repository.create(myEntity)
const {entities, raw} = await repository.find({
query: {
term: {
foo: 99,
},
},
})
- entities contain deserialized entities.
- raw contains raw response from elastic.
Repository methods
Entity methods
- create: creates entity
- createMultiple: creates multiple entities
- update: updates entity
- updateMultiple: updates multiple entities
- index: index entity
- indexMultiple: index multiple entities
- delete: deletes entity
- deleteMultiple: deletes multiple entities
- findOne: finds single entity by specified query
- findOneOrFail: finds single entity or throw exception
- find: finds entities by query
- findById finds single entity by id
Indices/Mapping methods
- createIndex: creates index
- deleteIndex: delete index
- updateMapping: update mapping
Additional methods
- on: registers function on specific repository actions
Advanced Usage
Query
Native query interface
Repository requires specific elastic structure object by default. Thanks to typescript query can be written really easily with autosuggestion. It will hint also properties from defined entity.
Query Builder
If you are more familiar with builders. Bodybuilder can be used easily.
- Install bodybuilder package.
npm i bodybuilder
- Builder usage
import * as bodybuilder from 'bodybuilder';
const body = bodybuilder().query('match', 'foo', 1).build();
const res = await repository.findOne(body);
Nested Entities
import { EsEntity, EsProperty, EsId } from 'elastic-orm';
@EsEntity('elastic_index')
export class MyEntity {
@EsId()
public id: string;
@EsProperty('integer')
public foo: number;
@EsProperty({ type: 'nested', entity: MyNestedEntity })
public nestedItem: MyNestedEntity;
@EsProperty({ type: 'nested', entity: MyNestedEntity })
public nestedItems: MyNestedEntity[];
}
export class MyNestedEntity {
@EsProperty('keyword')
public name: string;
}
// query for nested entities
const res = await repository.findOne({
query: {
nested: {
path: 'nestedItem',
query: {
bool: {
must: [{ match: { 'nestedItem.name': 'find this string' } }],
},
},
},
},
});
Generating Entity Identifier
Identifier is automatically generated. But if you want to control generation of ids, you can use this approach:
@EsEntity('elastic_index')
export class MyEntity {
@EsId({
generator: () => {
// apply custom logic here
return 'customGeneratedId';
},
})
public id: string;
}
Additional Field Options
For additional property options use parameter additionalFieldOptions
.
import { EsEntity, EsProperty, EsId } from 'elastic-orm';
@EsEntity('elastic_index')
export class MyEntity {
@EsId()
public id: string;
@EsProperty('text', {
additionalFieldOptions: {
enabled: false,
fields: {
raw: {
type: 'keyword',
},
},
},
})
public foo: string;
}
Aliases and Settings
@EsEntity('elastic_index', {
aliases: ['elastic_index_alias_read', 'elastic_index_alias_write'],
settings: {
number_of_replicas: 1,
number_of_shards: 5,
// and other settings definitions
},
mapping: {
dynamic: 'strict',
},
})
export class MyEntity{}
Update Mapping
import { FactoryProvider } from 'elastic-orm';
const mapping = FactoryProvider.makeSchemaManager().buildMapping(
FactoryProvider.makeMetaLoader().getReflectMetaData(TestingNestedClass),
);
await repository.updateMapping(mapping);
Delete Index
await repository.deleteIndex();
Global Request Manipulation
Enhancing elastic search requests is sometimes useful in one place. To do so you can register custom function which will be executed before every request on elastic.
For example:
- Enable explain for non production environments
repository.on('beforeRequest', (action, esParams, args) => {
if (process.env.NODE_ENV !== 'production') {
esParams.explain = true;
}
});
- In case of find method replace index for alias
repository.on('beforeRequest', (action, esParams, args) => {
if (action === 'find') {
esParams.index = 'elastic_index_alias_read';
}
});