Elasticsearch Spring Boot Geosearch

RESTful API that allows a user to find Starbucks locations based on one of the following criteria:

• Location (latitude, longitude & distance)
• City
• State
• Zip code

Geosearch Options

This project demonstrates two ways in which you can utilize Elasticsearch's geospatial search capabilities via the Spring Data Elasticsearch library.

1. ElasticsearchRepository method - implementation automatically taken care of by Spring Data Elasticsearch via naming convention findByLocationNear()

@RepositoryRestResource(path = "/starbucks-locations", collectionResourceRel = "/starbucks-locations")
public interface StarbucksSearchRepository extends ElasticsearchRepository<Starbucks, Long>, StarbucksSearchRepositoryCustom {

  Page<Starbucks> findByLocationNear(@Param("location") Point point, @Param("distance") Distance distance, Pageable pageable);

}

2. Custom Repository method - implementation defined in StarbucksSearchRepositoryImpl

@Repository
public class StarbucksSearchRepositoryImpl implements StarbucksSearchRepositoryCustom {

	private final JestClient jestClient;
	private final JestElasticsearchTemplate elasticsearchTemplate;

	public StarbucksSearchRepositoryImpl(JestClient jestClient) {
		this.jestClient = jestClient;
		this.elasticsearchTemplate = new JestElasticsearchTemplate(this.jestClient);
	}

	@Override
	public Page<Starbucks> findByLocationWithin(Point point, Distance distance, Pageable pageable) {
		return elasticsearchTemplate.queryForPage(getGeoQuery(point, distance, pageable), Starbucks.class);
	}

	private CriteriaQuery getGeoQuery(Point point, Distance distance, Pageable pageable) {
		return new CriteriaQuery(
				new Criteria("location").within(point, distance),
				pageable
		);
	}

}

Technologies

• Spring Boot
• Spring Data Elasticsearch
• Spring Data JPA
• Spring Data REST
• Spring Data Jest
• Docker
• AWS Elasticsearch Service
• AWS Relational Database Service

Configuration Options

Spring Profile	Database	Elasticsearch
default	in-memory H2	in-memory
docker	latest MySQL image	Elasticsearch 2.4.5-alpine image
aws	AWS RDS	AWS Elasticsearch Service

Running Locally

Prerequisites

• Java 8
• Maven

Steps

1. Set the Spring profile to default in application.yml (only if previously changed)

spring.profiles.active: default

2. Run the Java project using mvn spring-boot:run on the command line or using your favorite IDE

Running on Docker

Prerequisites

• Docker Engine

Steps

1. Set the Spring profile to docker in application.yml

spring.profiles.active: docker

2. From the root directory of this project, run docker-compose up on the command line

3. Run the Java project using mvn spring-boot:run on the command line or using your favorite IDE

Running on AWS Elasticsearch Service & RDS

Prerequisites

• AWS Elasticsearch Service cluster
• RDS instance up and running (or any other MySQL database)

Steps

1. Set the Spring profile to aws in application.yml

spring.profiles.active: aws

2. Modify the application.yml file with your Elasticsearch & RDS settings.

spring:
  profiles: aws
  data.jest:
    uri: https://[AWS_ELASTICSEARCH_URI]  # URI for AWS Elasticsearch index
    aws-region: [AWS_REGION_NAME]
  datasource:  # RDS settings
    driverClassName: com.mysql.jdbc.Driver
    url: jdbc:mysql://[AWS_DB_HOST]:3306/starbucks
    username: [DB_USERNAME]
    password: [DB_PASSWORD]

3. Run the Java project using mvn spring-boot:run on the command line or using your favorite IDE

Running the Tests

There are a complete set of unit tests covering core search capabilities as well as data population/synchronization from a MySQL database. To run the tests from the command line:

mvn test

Acknowledgements

Starbucks location data provided by Socrata