Getting started

Please note that Java 8 is necessary in order to build the project.

Source code

git clone https://bitbucket.org/jummp/jummpindexer/

Maven is used to manage the build process.

Build procedure

All dependencies are listed in the project's pom.xml.

Both Jummp and the indexer rely on a common set of domain classes. To avoid duplication, these have been packaged into a separate project available at https://bitbucket.org/jummp/annotationstore/

This dependency is retrieved automatically from the EBI Maven repository during the build process. The instructions for building a custom version of the domain classes are available at https://bitbucket.org/jummp/annotationstore

Use mvn package to build the binaries or mvn verify if, in addition, you also wish to run the integration tests.

Execution process

The project can be ran from command line

java -jar JummpSolrIndexer-1.0-SNAPSHOT.jar modelInfo.json

or by calling the RequestParser facade directly from your application:

#!java
import net.biomodels.jummp.indexing.solrindexer.RequestParser;

String pathToModelInfo = "./modelInfo.json";
RequestParser.handleRequest(pathToModelInfo);

In either case, the indexer expects to receive the path to a JSON file giving generic information about the model that is to be indexed. In addition, the file should also specify any additional resources that are relevant for the indexing process. These currently include

• the externalised configuration for Jummp -- this helps the indexer use the same database connection as the main application;
• the URI for the Solr instance which should be populated with the information that is extracted from the model;
• the path to an export of the Identifiers.org Registry, used to resolve annotations that link components of the model with external resources such as ontologies.

This file is typically produced by Jummp, although it need not be the case.