kvb rad

The application polls the position of the kvb bikes every minute and saves it in a managed cockroachdb database. It uses a selfhosted graphhopper server to calculate the shortest route between the two positions of a bike if it was moved between two polls. To view the data you can use the web application kvb-rad-ui.

commands

• Compile and run with h2 via: mvn spring-boot:run -Dspring.profiles.active=dryrun
• Compile and run via: mvn spring-boot:run
• Compile: mvn clean package -Dmaven.test.skip=true
• Compile and build docker container locally: mvn clean package -Dmaven.test.skip=true && docker build -t crowdsalat/kvb-rad .
• Start application with docker-compose: docker-compose up -d
• Start application with headless docker-compose: docker-compose --file ./docker-compose-headless.yml up -d
• Stop application: docker-compose down

Start with docker-compose

// with mysql and ui
docker-compose -f docker-compose.yml up

// with mysql and without ui
docker-compose -f docker-compose-headless.yml up

Start application with docker:

# create docker network
docker network create --driver bridge kvbbike

# create volume so db will be saved
docker volume create --driver local --name=kvbbike-db

# start mysql
docker run -d -p 3306:3306 --name kvb-mysql -network kvbbike -e MYSQL_ROOT_PASSWORD=myql -e MYSQL_DATABASE=kvbbike-db -v kvbbike-db:/var/lib/mysql mysql:latest

#start graphhopper
docker run -d -p 8989:8989 --name graphhopper --network kvbbike crowdsalat/graphhopper-cologne-bike

# start the application
docker run -d -p 8080:8080 --name kvb-rad --network=kvbbike crowdsalat/kvb-rad 

api consumption

graphhopper

This project uses a locally hosted Graphhopper server to calculate the distances and the waypoints of a bike movement. The route api is explained here. The waypoints are encoded to reduce bandwidth. To decode them there is a official [java] (https://github.com/graphhopper/graphhopper/blob/d70b63660ac5200b03c38ba3406b8f93976628a6/web/src/main/java/com/graphhopper/http/WebHelper.java#L43) and a official javascript library from graphhopper.

kvb

This project uses the nextbike api for cologne: [https://api.nextbike.net/maps/nextbike-live.json?city=14]. Alternatively there is a GBFS (General Bikeshare Feed Specification) conform api under: [https://api.nextbike.net/maps/gbfs/v1/nextbike_kg/gbfs.json]

used libraries

The Rest API is consumed via spring RestTemplate which uses Jackson (JSON Processor) annotations on java classes. The model classes for json were scaffolded with the vscode plugin Paste JSON as Code from Quicktype.

Alternatives for the use of RestTemplate would be:

• Jersey (JAX-RS)
• Spring WebClient from the WebFlux Stack (newer)

logging

Logging is done via slf4j with the default spring boot configuration.

data

The data is saved vie spring data repositories and jpa entities.

There is a MappedSuperclass AbstractEntity which provides the following generated fields:

• uuid
• creationDate (uses spring data annotation @CreatedDate, @EnableJpaAuditing and @EntityListeners(AuditingEntityListener.class))
• lastModified

The uuid works because of the hibernate annotation @GenericGenerator. The fields creationDate and lastModified use the spring data annotation @CreatedDate and @LastModifiedDate as well as the AuditingEntityListener. To activate the listener a configuration class is annotated with @EnableJpaAuditing.

TODO

• global exception handling: different methods
• save date with timezone and greater precision
• look at transactions
• handle bike points which could not been calculated
• handle bikes which are not in cologne: org.springframework.web.client.HttpClientErrorException$BadRequest: 400 Bad Request: [{"message":"Point 0 is out of bounds: 51.450448888889,7.0227255555556","hints":[{"message":"Point 0 is out of bounds: 51.450448888889,7.0227255555556","details":"com.graphhopper.util.exceptions.PointOutOfBoundsException","point_index":0}]}]