DEER Instance Matching Operator

It is a sub-task of Knowledge Graph Fusion.

Guide: How to the example

In the example folder you can find the example of DEER instance matching operator. The folder contains source data, target data and plugin jar file. In this example we are comparing "foaf:Person" in "data_nobelprize_org.nt" with "foaf:Person" in "lov_linkeddata_es_dataset_lov.nt".

Step 1: Clone the repo and change branch to instanceMatch

git clone https://github.com/binhudakhalid/DEER-InstanceMachingOperator.git --branch instanceMatch

Step 2:

cd DEER-InstanceMachingOperator

Step 3: Change directory to example, Here you can find the configuration.ttl file.

cd example/

Step 4: Run the instance matching operator through docker and feed the data as well as configuration file

For windows:

docker run -it --rm -v %cd%/data:/data -v %cd%/plugins:/plugins -v %cd%/:/config dicegroup/deer:latest /config/configuration.ttl

For linux base system:

docker run -it --rm -v $(pwd)/data:/data -v $(pwd)/plugins:/plugins -v $(pwd)/:/config dicegroup/deer:latest /config/configuration.ttl

Resut

Output: In the end you will get the Reificated output as a jena model. Here we are only showing simplified output. The entity http://www.inf.kcl.ac.uk/staff/simonm/ is matching with http://data.nobelprize.org/resource/laureate/680 and the confidence is 0.707

[f9069a7b-5f4e-4081-a118-26d2dc1258ab, http://www.w3.org/1999/02/22-rdf-syntax-ns#type, http://www.w3.org/1999/02/22-rdf-syntax-ns#Statement] 

[f9069a7b-5f4e-4081-a118-26d2dc1258ab, http://www.w3.org/1999/02/22-rdf-syntax-ns#subject, http://data.nobelprize.org/resource/laureate/680]

[f9069a7b-5f4e-4081-a118-26d2dc1258ab, http://www.w3.org/1999/02/22-rdf-syntax-ns#predicate, http://www.w3.org/2002/07/owl#sameAs] 

[f9069a7b-5f4e-4081-a118-26d2dc1258ab, http://www.w3.org/1999/02/22-rdf-syntax-ns#object, http://www.inf.kcl.ac.uk/staff/simonm/] 

[f9069a7b-5f4e-4081-a118-26d2dc1258ab, https://w3id.org/deer/confidence, "0.7071067811865475"] 

[https://w3id.org/deer/datasetTarget, https://w3id.org/deer/path, data/lov_linkeddata_es_dataset_lov.nt]
[https://w3id.org/deer/datasetSource, https://w3id.org/deer/path, data/data_nobelprize_org.nt]
[https://w3id.org/deer/objectType#1, https://w3id.org/deer/is, http://www.w3.org/1999/02/22-rdf-syntax-ns#type] 
[https://w3id.org/deer/objectType#2, https://w3id.org/deer/is, http://xmlns.com/foaf/0.1/Person] 
[https://w3id.org/deer/subjectType#1, https://w3id.org/deer/is, http://www.w3.org/1999/02/22-rdf-syntax-ns#type]
[https://w3id.org/deer/subjectType#2, https://w3id.org/deer/is, http://xmlns.com/foaf/0.1/Person] 

Note: if you are on windows please use docker for running DEER because sometimes it has issue on windows.

Configuration file

The DEER operator requires multiple parametes such as type, source data, target data, sourceRestriction and targetRestriction. Additional parameters are tabuSourceProperty, tabuTargetProperty, coverage and maxLimit.

coverage: You can set the coverage of each property. Calulated by "number of instances with this proptery/Total number of Instances".

maxLimit: Set the maximum number of properties to used for comparision,

type: The type of input data. It can be file or url endpoint.

source: Path to data file or url of endpoint.

target: Path to data file or url of endpoint.

sourceRestriction: Set the restrictions for source data

targetRestriction: Set the restrictions for target data

tabuSourceProperty: Source tabu properties are those properties that we don't want to use for instance matching

tabuTargetProperty: Source tabu properties are those properties that we don't want to use for instance matching

Example:

Example configuration files can found at DEER-InstanceMachingOperator\example\configurationFiles.

The data set can be downloaded from:

https://hobbitdata.informatik.uni-leipzig.de/teaching/theses/skuhlmann/query_models/

https://hobbitdata.informatik.uni-leipzig.de/teaching/theses/skuhlmann/link_models/

How to build Instance Matching Operator

Using mvn clean package in this folder will generate the plugin under ./target/plugin-starter-${version}-plugin.jar. Copy the plugin into a folder named plugins/ in the working directory from which you want to invoke DEER and it will automatically be loaded. Copy the source and target data files to data folder.

In order to invoke DEER, either download it from GitHub or use our Docker image:

DEER requires two parametes one is path to configuration file and other is path to jar file of Instance Matching operator.

For Linux

docker run -it --rm \
   -v $(pwd)/plugins:/plugins -v $(pwd)/src/test/resources:/config dicegroup/deer:latest \
   /config/configuration.ttl

If you want to give data files to DEER operator. you need to add $(pwd)/data:/data

docker run -it --rm -v $(pwd)/data:/data -v $(pwd)/plugins:/plugins -v $(pwd)/:/config dicegroup/deer:latest /config/configuration.ttl

For Windows

1. You need to create a folder plugins inside deer-plugin-starter directory.

2. Copy the newly generated plugin under ./target/plugin-starter-${version}-plugin.jar to plugins folder.

3. Run the docker command from the deer-plugin-starter directory.

docker run -it --rm  -v %cd%/plugins:/plugins -v %cd%/src/test/resources:/config dicegroup/deer:latest /config/configuration.ttl