Semantic Engine

The Semantic Engine transforms concise RDF data into a source-independent, semantically rich RDF model. It is designed to allow ontologists and programmers to work collaboratively on the creation and implementation of a target model that standardizes data from disparate sources. See the QuickStart Guide to get up and running. For more extensive description of the various components and a toy use case tutorial, see this Word document.

The system uses a syntax called the Semantic Engine Language to define RDF graph transformations. There are 2 types of configuration files that use this syntax: a Graph Specification file (.gs) defines the RDF target model, and a Transformation Instruction Set (.tis) maps another RDF schema to the target model. The intention is that this schema will have similar semantics to the original data source, whereas the target model will include rich semantics defined by an application ontology. Pre-existing public .gs and .tis files are available in drivetrain/config. For documentation about designing your own Semantic Engine Language configurations, see the Semantic Engine Language Feature Guide.

Installation

Prerequisites

You must have SBT and Java installed on your system. Unless running from a precompiled .jar file, interaction with the software will take place through the SBT console.

Setup

Clone the repository locally, and copy/remove the .template suffix from the following 4 files:

• turbo_properties.properties.template
• drivetrain/build.sbt.template
• drivetrain/project/plugins.sbt.template
• drivetrain/project/build.properties.template

Configure turbo_properties.properties to use the desired Semantic Engine Language files: Transformation Instruction Set and Graph Specification. Pre-existing options exist in the config folder.

SBT Commands

The following commands can be run via the SBT console. SBT should be started from the directory drivetrain/.

Data commands

run allUpdates Runs all update specifications from selected transformation instruction set in specified order. The graph will be modified.

run singleUpdate <some_process_URI> Runs only a specific update specification from the selected transformation instruction set. The graph will be modified.

run printQueryForUpdate <some_process_URI> Prints the generated update query to the console for a specific update specification. The graph will not be modified.

run buildTest Builds automated snapshot tests for all update specifications (even unqueued ones) in the selected transformation instruction set. The graph will not be modified.

run buildTest <some_process_URI> Builds an automated snapshot test for a specific update specification. The graph will not be modified.

run debug [--min] Generates a synthetic set of input triples for all queued update specifications and runs them in order, leaving the output for examination in the testing repository. If min flag is present, only the minimum required set of triples will be generated for each update specification, otherwise all possible triples will be generated. Note that if updates reference inputs that were the output of a previous update, instances of these input classes will still be created by the synthetic triple generation service, even if other instances of those classes already exist in the graph. The graph will not be modified.

run debug [--min] <some_process_URI> Generates a synthetic set of input triples for a specific update specification and then runs just that update specification, leaving the output for examination in the testing repository. If min flag is present, only the minimum required set of triples will be generated for each update specification, otherwise all possible triples will be generated. The graph will not be modified.

run validateModel Runs all Graph Model Validation steps without performing any other tasks. Errors will be apparent on the console as AssertionErrors or RuntimeExceptions for now. The graph will not be modified.

run updateModel Clears the model repository and uploads the Transformation Instruction Set, Graph Specification, Semantic Engine Language Ontology, and application ontology into the model repository. This functionality will also be executed automatically before any other data command is executed or test runs.

Test commands

test Runs all tests

testOnly edu.upenn.turbo.<test_class_name> Runs the tests inside a specific test class

testOnly edu.upenn.turbo.SnapshotTest Runs all Snapshot tests

testOnly edu.upenn.turbo.SnapshotTest -- -DfindTest={searchString} Runs Snapshot tests whose name includes the searchString

Click here for more detailed information on snapshot test generation.

Utility commands

run loadRepoFromFile <path_to_some_file> Loads an RDF file hosted locally into the production repository. Default format is RDF/XML. For now, format must be changed in the code in the DrivetrainDriver object. CTRL+F "loadRepoFromFile to find the appropriate section.

run loadRepoFromUrl <URL_to_some_file> Loads an RDF file hosted at the given URL into the production repository. Default format is RDF/XML. For now, format must be changed in the code in the DrivetrainDriver object. CTRL+F "loadRepoFromUrl to find the appropriate section.

run loadOntologyToProductionRepo Loads the ontology file provided as the value for key ontologyURL in the properties file into the production repository.

run loadOntologyToModelRepo Loads the ontology file provided as the value for key ontologyURL in the properties file into the model repository.