Data Preparation

1. Primary datasets The four main datasets—FB15K, NELL-995, WN18, and LUBM-1—are listed in the datasets_csv file.

2. Additional datasets Datasets such as LUBM-? are generated by UBA-1.7 (see https://swat.cse.lehigh.edu/projects/lubm/) and then processed with:

   python ./py_script/owl2rdf.py

Environment Setup

1. Python dependencies

   pip install -r ./py_script/requirements.txt

2. Java SE Install Java SE 23.

3. PostgreSQL (optional) If you wish to run on PostgreSQL, install and configure a server—see the official guide: https://www.postgresql.org/docs/17/index.html

Compile the System

bash ./lib/import_local_jars.sh
mvn package

Running the System

java -jar -Xmx{memory_limit} ./target/QSC-1.0.jar -I {absolute_path_of_program} {dataset_name} \
     -d {pg|duckdb} \
     -i {pg_db_path|duckdb_db_path} \
     -s {experiment_name} \
     -a {acquisition_function}

Example:

java -jar -Xmx32G ./target/QSC-1.0.jar -I ${abs path of this project} WN18 \
     -d pg \
     -i localhost:5432/wn18 \
     -s test#WN18 \
     -a qehvi