• Python 3.4 or above (For Knowldge Graph Summarization)
• numpy
• scipy
• pandas
• JAVA 11 or above (For TeBaQA)
• ElasticSearch 6.6.1 (For TeBaQA)

Summarization is done over DBPedia Knowldge Graph. Data sets we used are

• Mapping Based Objects
• Mapping Based Literals
• Labels
• Specific Mapping Based Properties
• Person Data

• Clone or download the project in desired directory
• In lines 16-18 of base.py, change the path of DBPedia dataset according to your local data directories.
• In addition to this, you may need to change according to your directory structure and file naming conventions:
  • The subdirectory where queries are saved for the user(i.e. user0, user1, and user2 for QALD8, QALD9, and LcQuAD respectively), The queries should be in WebQSP Format. If the queries are not in WebQSP format, you may change to script folder in order to convert the queries in desired format.
• To generate the summary, run this command in the base directory
  python3 main.py
• Incase you want to change code behaviour, you can edit arguments of parse_args method in main.py class.
• Summaries are generated in the directory out0,out1,and out2 of the Project, according to user.

• These generated summaries along with DBPedia DataSets are then used by QA system(i.e. TeBaQA)
• clone or download the TeBaQA.
• DataSets are indexed on the ElasticSearch via TeBaQA class GenerateIndexes.
• update the entityLinking.properties file in the base folder, according to directory where datasets are stored.
• After updating the properties file, run the following command in the base directory
  mvn exec:java -Dexec.mainClass="de.uni.leipzig.tebaqa.indexGeneration.GenerateIndexes"
• Once indexes are generated, Configure each microservice to run the Question Aswering System.
• Finally, we build the project and then run via scripts named build-script.sh and run-script.sh in the base folder.
• Project is started on the following url
  http://localhost:8080

1. Safavi, Tara et al. “Personalized Knowledge Graph Summarization: From the Cloud to YourPocket.” 2019 IEEE International Conference on Data Mining (ICDM) (2019): 528-537.
2. Vollmers, Daniel et al. “Knowledge Graph Question Answering using Graph-Pattern Isomor-phism.” ArXiv abs/2103.06752 (2021): n. pag