Project to develop Neo4j as the backing store for InterMine.
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.
Download the latest source code by cloning the git repository.
git clone https://github.com/intermine/neo4j.git
Add Neo4j Database URL, InterMine Service URL, your credentials and other information in neo4jloader.properties.
Place the apoc-3.1.*.jar file to the plugins directory inside the neo4j home directory. For example, /home/yourusername/neo4j-community-3.2.0/plugins.
Install the remaining prerequisites as per the instructions given in their official websites.
Download the following prerequisites from their official websites.
- JDK 8
- Apache Tomcat 7
- Neo4j Graph Database 3.2
- Neo4j APOC 3.1+
Change the current directory to project home by running cd path/to/project/.
Build the project JAR using Apache Ant by running ant jar.
The metadata package deals with handling metadata of the Neo4j InterMine graph. The description of the metagraph can be found in metadata.md.
The cypher package deals with the conversion of PathQuery into Cypher.
For detailed documentation, you can checkout the Java Docs.
Loads a mine into Neo4j using parameters in neo4jloader.properties.
./run org.intermine.neo4j.Neo4jLoader
A variation on Neo4jLoader which stores up a bunch of queries and submits them to Neo4j as a batch rather than one at a time. Written to compare timing.
./run org.intermine.neo4j.Neo4jBatchLoader
Completes the references and collections for nodes that have only had attributes stored as a result of being loaded as references and collections by Neo4jLoader.
This is very handy to fill out the graph without adding new nodes. neo4jloader.properties is used, especially loaded.classes to indicate which types of nodes you'd like completed.
./run org.intermine.neo4j.Neo4jCompleter
Load a single node into Neo4j identified by its class and InterMine id. Other parameters are read from neo4jloader.properties.
./run org.intermine.neo4j.Neo4jNodeLoader Gene 4295368
Load edges into Neo4j given by source class, edge reference or collection, and target as referenced in the edge class. This converts IM objects
which store relations, like Location, into Neo4j edges with properties. Since some target names, e.g. chromosomeLocation, refer to simpler-named classes, like Location, there are two
parameters in neo4jloader.properties, intermine.edge.classes and neo4j.edge.types, which map IM classes to Neo4j edge types so that, for example,
chromosomeLocation edges are given the type "location" in Neo4j.
./run org.intermine.neo4j.Neo4jEdgeLoader Gene chromosomeLocation locatedOn
Simple utility dump an InterMine model in XML form to standard output.
./run org.intermine.neo4j.InterMineModelXMLDumper http://www.synbiomine.org/synbiomine/service
Parses a Neo4j-annotated model XML file, which contains instructions like neo4j-include="true" in the class, attribute, reference and collection definitions to indicate that those items should be ignored in Neo4j loading. The main method takes an XML file as a parameter and spits out the data model with "X" denoting ignored items. Methods like isIgnored(ReferenceDescriptor) provide an easy way to see if a given reference (for example) should be ignored during loading.
Generates the schema of the Neo4j database using org.intermine.neo4j.metadata package and outputs the model to the console. Also, tests each validation method in the Model class.
./run TestSchemaGenerator
Generates the schema, extracts the data model and tests the model methods.
./run TestModel <node label> <property | relationship label>
Tests the conversion of PathQuery XML to a Cypher query.
./run TestQueryGenerator
Tomcat webapps for doing interactive demos.
The webapp page, currently holds a textarea to submit PathQuery XML to...
A servlet which receives an XML PathQuery then submits it to: (1) a mine's web service endpoint; (2) the mine using the Java PathQuery API; (3) a Neo4j using a reverse-engineered Cypher query,
which may or may not actually work, and returns all the results along with query execution + mine-to-webapp timing. Parameters are in web.xml.
A servlet which mimics the IM query endpoint: it receives an XML PathQuery with format=tab and then queries Neo4j using a reverse-engineered Cypher query, returning the tabular results.