Backend for ModelDB version 2
- General tools: git, maven, jdk 1.8
Note : This project requires JAVA version 8 to run. Please setup and configure JAVA version 8 prior to running and then run the following steps.
mvn clean
mvn package -Dmaven.test.skip=true
There are two ways a jar can be created:
- Stand-alone JAR generated by maven build and this jar doesn't contain other libraries. This jar is located after a successful build at
target/modeldb-1.0-SNAPSHOT.jar
. - Uber jar generated by spring-boot maven build, this JAR contains all the other maven libraries. This jar is located after a successful build at
target/modeldb-1.0-SNAPSHOT-client-build.jar
Run gRPC server using the following command
java -jar target/modeldb-1.0-SNAPSHOT-client-build.jar
The execution assumes the database schema matches the schema expected by hibernate entities. To run liquibase to
create the required tables or modify the schema set the environment variable LIQUIBASE_MIGRATION
:
LIQUIBASE_MIGRATION: true
If the execution should just update the database schema then set the variable LIQUIBASE_MIGRATION
:
RUN_LIQUIBASE_SEPARATE: false
The tests require a relational database to be running. Ensure the required JDBC connector is available in pom.xml. Set the details in config.yaml
test:
test-database:
DBType: relational
RdbConfiguration:
RdbDatabaseName: modeldb_test
RdbDriver: "org.postgresql.Driver"
RdbDialect: "org.hibernate.dialect.PostgreSQLDialect"
RdbUrl: "jdbc:postgresql://localhost:5432"
RdbUsername: postgres
RdbPassword: root
Set the VERTA_MODELDB_CONFIG environment variable to point to config.yaml
mvn package
mvn test
mvn -Dtest=TestName test
- Provide your existing database name in config.yaml file and DB user credentials.
database:
DBType: relational
timeout: 4
liquibaseLockThreshold: 60 #time in second
RdbConfiguration:
RdbDatabaseName: modeldb
RdbDriver: "org.postgresql.Driver"
RdbDialect: "org.hibernate.dialect.PostgreSQLDialect"
RdbUrl: "jdbc:postgresql://localhost:5432"
RdbUsername: postgres
RdbPassword: root
- If you are connecting to postgres, the database needs to be already created.
- Ensure the user mentioned in
RdbUsername
has create privileges on the database mentioned inRdbDatabaseName
.
If you have Docker Compose installed, you can bring up a ModelDB server with just a couple commands.
Note the following points when the setting up docker
- Populate config.yaml file with correct information of the database engine.
- As per docker hub suggestion for the wait-for-it.sh, the following links provide documentation of wait-for-it.sh script file.
Configure Host & Port number for ArtifactStore gRPC server like,
artifactStore_grpcServer:
host: localhost OR IP location
port: 8086
If the cloud store is s3 , then backend can use the local credentials of the machine
config.yaml --> artifactStoreConfig --> nfsRootPath : "root path"
- Configure port number for ArtifactStore gRPC server : artifactStore_grpcServer --> port : 8086
- Setup cloud storage name : artifactStoreConfig --> name : amazonS3 OR googleCloudStorage OR nfs
- Setup list of bucket name : artifactStoreConfig --> buckets_names : - bucket_name
Create code coverage reports by the following commands
mvn clean test
Show the created code coverage report for unit tests from following package directory, it has module wise code coverage report in 'html' format after executing unit tests
"target/site/jacoco-ut"
Download SonarQube from the following URL
https://www.sonarqube.org/downloads/
How to start SonarQube for view the analytics & coverage use following URL
https://docs.sonarqube.org/latest/setup/get-started-2-minutes/
Create code analytics & coverage reports by the following commands
mvn clean package sonar:sonar
Show the created code analytics & coverage report for the project from the following link
http://localhost:9000/dashboard?id=ai.verta.modeldb%3Amodeldb