Tiamat is the Stop Place Register. It is used nationally in Norway, and other places. Tiamat is created with technologies like Spring Boot, Hibernate, Postgis, Jersey and Jackson.
- Supports different pre steps and merging options for stop places, handling bad data quality.
- Assigns unique IDs to stop places (if desired).
- Validates incoming data against the XML schema.
Supports exporting stop places and other entities to the http://netex-cen.eu/ format. There are many options for exports:
- Asynchronous exports to google cloud storage. Asynchronous exports handles large amount of data, even if exporting thousands of stop places.
- Synchronous exports directly returned
- Several export parameters and filtering (ex: query or administrative polygons filtering)
- Exports can be validated against the NeTEx schema, ensuring quality.
Tiamat provides a rich GraphQL API for stop places, topographic places, path links, tariff zones and so on, support the same parameters as the NeTEx export API. It also supports mutations. So you can update or create entities. There are also graphql processes (named functions) which allows functionality like merging quays or stop places.
A frontend for Tiamat is available. It's name is Abzu. See https://github.com/entur/abzu
Tiamat uses Hazelcast memory grid to communicate with other instances in kubernetes. This means that you can run multiple instances.
After import stop places and assigning new IDs to stop places, tiamat keeps olds IDs in a mapping table. The mapping table between old and new IDs is available through the GraphQL API and a REST endpoint.
Tiamat supports looking up and populating references to tariff zones and topographic places from polygon matches when saving a stop place.
Stop places and other entities are versioned. This means that you have full version history of stop places and what person that made those changes. Tiamat also includes a diff tool. This is used to compare and show the difference between two versions of a stop place (or other entity).
mvn clean install
You need the directory /deployments/data
with rights for the user who
performs the build.
Tiamat uses testcontainers to run integration tests against a real database. To run Testcontainers-based tests, you need a Docker-API compatible container runtime for more detail see https://www.testcontainers.org/supported_docker_environment/
(default profiles are set in application.properties)
To run Tiamat with Docker compose, you need to have a docker-compose.yml file. In docker-compose folder you will find a compose.yml file.:
docker compose up
This will start Tiamat with PostgreSQL and Hazelcast. and you can access Tiamat on http://localhost:1888 and the database on http://localhost:5433 and graphiql on http://localhost:8777/services/stop_places/graphql , At start up tiamat copy empty schema to the database. Spring properties are set in application.properties. Security is disabled in this setup.
To run with PostgreSQL you ned an external application.properties. Below is an example of application.properties:
spring.jpa.database=POSTGRESQL
spring.datasource.platform=postgres
spring.jpa.properties.hibernate.hbm2ddl.import_files_sql_extractor=org.hibernate.tool.hbm2ddl.MultipleLinesSqlCommandExtractor
spring.jpa.hibernate.hbm2ddl.import_files_sql_extractor=org.hibernate.tool.hbm2ddl.MultipleLinesSqlCommandExtractor
spring.jpa.hibernate.ddl-auto=none
spring.http.gzip.enabled=true
#spring.jpa.properties.hibernate.format_sql=true
spring.jpa.properties.hibernate.order_updates=true
spring.jpa.properties.hibernate.batch_versioned_data=true
spring.flyway.enabled=true
spring.flyway.table =schema_version
server.compression.mime-types=application/json,application/xml,text/html,text/xml,text/plain
spring.jpa.hibernate.id.new_generator_mappings=true
spring.jpa.hibernate.use-new-id-generator-mappings=true
spring.jpa.properties.hibernate.cache.use_second_level_cache=false
spring.jpa.properties.hibernate.cache.use_query_cache=false
spring.jpa.properties.hibernate.cache.use_minimal_puts=false
spring.jpa.properties.hibernate.cache.region.factory_class=org.rutebanken.tiamat.hazelcast.TiamatHazelcastCacheRegionFactory
netex.import.enabled.types=MERGE,INITIAL,ID_MATCH,MATCH
hazelcast.performance.monitoring.enabled=true
hazelcast.performance.monitoring.delay.seconds=2
management.endpoints.web.exposure.include=info,env,metrics
management.endpoints.prometheus.enabled=true
management.metrics.endpoint.export.prometheus.enabled=true
spring.jpa.properties.hibernate.jdbc.batch_size=20
spring.jpa.properties.hibernate.default_batch_fetch_size=16
spring.jpa.properties.hibernate.generate_statistics=false
changelog.publish.enabled=false
jettyMaxThreads=10
jettyMinThreads=1
spring.datasource.hikari.maximumPoolSize=40
spring.datasource.hikari.leakDetectionThreshold=30000
tiamat.locals.language.default=eng
tariffZoneLookupService.resetReferences=true
debug=true
# Disable feature detection by this undocumented parameter. Check the org.hibernate.engine.jdbc.internal.JdbcServiceImpl.configure method for more details.
spring.jpa.properties.hibernate.temp.use_jdbc_metadata_defaults = false
# Because detection is disabled you have to set correct dialect by hand.
spring.jpa.database-platform=org.hibernate.dialect.PostgreSQL9Dialect
tariffzoneLookupService.resetReferences=true
spring.jpa.properties.hibernate.dialect=org.hibernate.spatial.dialect.postgis.PostgisDialect
spring.database.driverClassName=org.postgresql.Driver
spring.datasource.url=jdbc:postgresql://localhost:5436/tiamat
spring.datasource.username=tiamat
spring.datasource.password=tiamat
#OAuth2 Resource Server
spring.security.oauth2.resourceserver.jwt.issuer-uri=https:http://localhost:8082/realms/entur
tiamat.oauth2.resourceserver.auth0.ror.jwt.audience=abzu
tiamat.oauth2.resourceserver.auth0.ror.claim.namespace=role_assignments
spring.cloud.gcp.pubsub.enabled=false
aspect.enabled=true
netex.id.valid.prefix.list={TopographicPlace:{'KVE','WOF','OSM','ENT','LAN'},TariffZone:{'*'},FareZone:{'*'},GroupOfTariffZones:{'*'}}
server.port=1888
blobstore.gcs.blob.path=exports
blobstore.gcs.bucket.name=tiamat-test
blobstore.gcs.credential.path=gcloud-storage.json
blobstore.gcs.project.id=carbon-1287
security.basic.enabled=false
management.security.enabled=false
authorization.enabled = true
rutebanken.kubernetes.enabled=false
async.export.path=/tmp
publicationDeliveryUnmarshaller.validateAgainstSchema=false
publicationDeliveryStreamingOutput.validateAgainstSchema=false
netex.validPrefix=NSR
netex.profile.version=1.12:NO-NeTEx-stops:1.4
blobstore.local.folder=/tmp/local-gcs-storage/tiamat/export
spring.profiles.active=local-blobstore,activemq
To start Tiamat with this configuration, specify spring.config.location:
java -jar -Dspring.config.location=/path/to/tiamat.properties --add-opens java.base/java.lang=ALL-UNNAMED -Denv=dev tiamat-0.0.2-SNAPSHOT.jar
Tiamat is using HikariCP. Most properties should be be possible to be specified in in application.properties, like spring.datasource.initializationFailFast=false
. More information here. https://github.com/brettwooldridge/HikariCP/wiki/Configuration
See also http://stackoverflow.com/a/26514779
docker run -it -d -p 5435:5432 --name postgress-13 -e POSTGRES_USER=tiamat -e POSTGRES_PASSWORD="tiamat" -e POSTGRES_INITDB_ARGS="-d" postgis/postgis:13-master
During the implementation of Tiamat was desirable to produce NeTEx IDs for stop places more or less gap less. The reason for this implementation was legacy systems with restrictions of maximum number of digits.
It is possible to control whether IDs should be generated outside Tiamat or not. See the class ValidPrefixList.
Setting the property netex.validPrefix
tells Tiamat to generate IDs for new entities.
Please note that it is not possible to do an initial import (see ImportType) multiple times with the same IDs.
It's all initiated by an entity listener annotated with PrePersist
on the class IdentifiedEntity
called IdentifiedEntityListener
.
NetexIdAssigner
determines if the entity already has an ID or not. NetexIdProvider
either return a new ID or handles explicity claimed IDs if the configured prefix matches. See ValidPrefixList
for the configuration of valid prefixes, and prefixes for IDs generated elsewhere. The GaplessIdGeneratorService
uses Hazelcast to sync state between instances and avoid conflicts.
Both Tiamat and Abzu are set up to be used with Keycloak or Auth0. A detailed guide on how to setup Keycloak can be found here.
It is possible to configure if tiamat should validate incoming and outgoing NeTEx xml when unmarshalling or marshalling publication deliveries. Default values are true. Can be deactivated with setting properties to false.
publicationDeliveryStreamingOutput.validateAgainstSchema=false
publicationDeliveryUnmarshaller.validateAgainstSchema=true
It is possible to export stop places and topographic places directly to NeTEx format. This is the endpoint: https://api.dev.entur.io/stop-places/v1/netex
https://api.dev.entur.io/stop-places/v1/netex?q=Arne%20Garborgs%20vei
https://api.dev.entur.io/stop-places/v1/netex?q=3115
https://api.dev.entur.io/stop-places/v1/netex?stopPlaceType=RAIL_STATION
It is also possible with multiple types.
https://api.dev.entur.io/stop-places/v1/netex?municipalityReference=KVE:TopographicPlace:1003
https://api.dev.entur.io/stop-places/v1/netex?countyReference=KVE:TopographicPlace:11
https://api.dev.entur.io/stop-places/v1/netex?size=1000
https://api.dev.entur.io/stop-places/v1/netex?page=1
You can specify a list of NSR stop place IDs to return
https://api.dev.entur.io/stop-places/v1/netex?idList=NSR:StopPlace:3378&idList=NSR:StopPlace:123
allVersions
. Acceptable values are true or false. If set to true, all versions of matching stop places will be returned.
If set to false, the highest version by number will be returned for matching stop places. This parameter is not enabled when using the version valitity parmaeter.
Match only stop places without location
Use the parameter: withoutLocationOnly=true
The parameter topographicPlaceExportMode
can be set to NONE, RELEVANT or ALL
Relevant topographic places will be found from the exported list of stop places.
The parameter tariffZoneExportMode
can be set to NONE, RELEVANT or ALL
Relevant tariff zones with be found from the exported list of stop places. Because stop places can have a list of tariff zone refs.
The parameter groupOfStopPlacesExportMode
can be set to NONE, RELEVANT or ALL
Relevant group of stop places can be found from the exported list of stop places.
The versionValidity
parameter controls what stop places to return.
- ALL: returns all stops in any version (See allVersions attribute), regardless of version validity
- CURRENT: returns only stop place versions valid at the current time
- FUTURE_CURRENT: returns only stop place versions valid at the current time, as well as versions valid in the future.
https://api.dev.entur.io/stop-places/v1/netex?tariffZoneExportMode=RELEVANT&topographicPlaceExportMode=RELEVANT&groupOfStopPlacesExportMode=NONE&q=Nesbru&versionValidity=CURRENT&municipalityReference=KVE:TopographicPlace:0220
Returns stop places with current version validity now, matching the query 'Nesbru' and exists in municipality 0220. Fetches relevant tariff zones and topographic places.
Asynchronous export uploads exported data to google cloud storage. When initiated, you will get a job ID back. When the job is finished, you can download the exported data.
Most of the parameters from synchronous export works with asynchronous export as well!
curl https://api.dev.entur.io/stop-places/v1/netex/export/initiate
Pro tip: Pipe the output from curl to xmllint to format the output:
curl https://api.dev.entur.io/stop-places/v1/netex/export/initiate | xmllint --format -
curl https://api.dev.entur.io/stop-places/v1/netex/export
curl https://api.dev.entur.io/stop-places/v1/netex/export/130116/content | zcat | xmllint --format - > export.xml
See also https://rutebanken.atlassian.net/browse/NRP-924
Clean existing data in postgresql (streamline if frequently used):
TRUNCATE stop_place CASCADE;
TRUNCATE quay CASCADE;
TRUNCATE topographic_place CASCADE;
If you are running this from spring:run
, then you need to make sure that you have enough memory available for the java process (in case of large data sets).
Another issue is thread stack size, which might need to be increased when coping with really large NeTEx imports.
Example:
export MAVEN_OPTS='-Xms256m -Xmx1712m -Xss256m -XX:NewSize=64m -XX:MaxNewSize=128m -Dfile.encoding=UTF-8'
This NeTEx file should not contain NSR ID. (The NSR prefix is configurable in the class ValidPrefixList)
- Tiamat will match existing stops based on name and coordinates.
- Tiamat will merge Quays inside stops that are close, have the same original ID and does not have too different compass bearing.
Tiamat will return the modified NeTEx structure with it's own NSR IDs. Original IDs will be present in key value list on each object.
curl -XPOST -H"Content-Type: application/xml" -d@my-nice-netex-file.xml http://localhost:1997/services/stop_places/netex
When importing with importType=INITIAL, a parallel stream will be created, spawning the original process. During import, user authorizations is checked, thus accessing SecurityContextHolder. By default, SecurityContextHolder use DEFAULT_LOCAL_STRATEGY. When using INITIAL importType, you should tell Spring to use MODE_INHERITABLETHREADLOCAL for SecurityContextHolder, allowing Spring to duplicate Security Context in spawned threads. This can be done setting env variable :
-Dspring.security.strategy=MODE_INHERITABLETHREADLOCAL
If not, the application may complain about user not being authenticated if Spring tries to check authorization in a spawned process
GraphQL endpoint is available on
https://api.dev.entur.io/stop-places/v1/graphql
Tip: GraphiQL UI available on https://api.dev.entur.io/graphql-explorer/stop-places using GraphiQL:
https://github.com/graphql/graphiql
(Use e.g. Modify Headers
for Chrome to add bearer-token for mutations)
To create the database for tiamat, download and use the flyway command line tool: https://flywaydb.org/documentation/commandline/
Migrations are executed when tiamat is started.
Create a new file according to the flyway documentation in the folder resources/db/migrations
.
Commit the migration together with code changes that requires this schema change. Follow the naming convention.
Various queries and scripts related to tiamat, has been collected here: https://github.com/entur/tiamat-scripts
Tiamat is built using CircleCI. See the .circleci folder.