remixman/vedas

VEDAS

VEDAS is a RDF store engine that be able to query with SPARQL and run on single GPU.

Dependencies

• ModernGPU
• Thrust
• Raptor RDF Syntax Library
• Rasqal RDF Query Library

Build

make

Build the VEDAS database

First, you should prepare the RDF data in N-triple format or .nt extension. vdBuild is used for load the triple data into VEDAS internal format

./vdBuild <database_name> <path_to_nt_file>

For example

./vdBuild watdiv500M /home/username/data/watdiv/watdiv.500M.nt

The internal database file <database_name>.vdd and <database_name>.vds will be generated.

Query RDF data

VEDAS support query only from file. The vdQuery is the query engine that load the RDF data and wait for the input file.

./vdQuery <database_name>

The prompt will shown after finish loaded data. To submit the query, use command sparql <path_to_sparql_query_file> and exit to terminate the program.

You can use -sparql-path option to speccify the sparql file path.

./vdQuery <database_name> -sparql-path=<path_to_sparql_query_file>

Visualize the RDF Graph

After load the database with vdBuild, it will construct the graph vertex and edge files, named tools/nodes.txt and edges/nodes.txt. You can generate the GraphML file with the following command

cd tools
pip install -r requirements.txt
python graphml.py

The output file triple-data.graphml can opened with any supported software e.g. Graphia, Gephi etc.

BibTeX

@Article{vedas2021,
  author={Makpaisit, Pisit and Chantrapornchai, Chantana},
  title={VEDAS: an efficient GPU alternative for store and query of large RDF data sets},
  journal={Journal of Big Data},
  year={2021},
  month={Sep},
  day={16},
  volume={8},
  number={1},
  pages={125},
  issn={2196-1115},
  doi={10.1186/s40537-021-00513-y},
  url={https://doi.org/10.1186/s40537-021-00513-y}
}

@article{makpisit2023sparql,
  title={SPARQL Query Optimizations for GPU RDF Stores},
  author={Makpisit, Pisit and others},
  journal={ECTI Transactions on Computer and Information Technology (ECTI-CIT)},
  volume={17},
  number={2},
  pages={235--244},
  year={2023}
}