Skip to content

Latest commit

 

History

History
59 lines (42 loc) · 2.7 KB

README.md

File metadata and controls

59 lines (42 loc) · 2.7 KB
Raw

LOVStats

DOI

A LODStats extension to compute restriction type statistics for RDF data models, currently with modules for axioms (OWL, RDFS) and constraints (SHACL). Whereas the actual extension consists of own statistical modules (src/restriction-types-stats, src/shape-constraints-stats) and a different commandline script (src/lodstats.py) compared to the original LODStats.

Example, calling the lodstats.sh script to execute lodstats within a docker container

# calling lodstats
bash lodstats.sh

# execute our statistics (if folder input mounted to docker container)
bash lodstats.sh -o restriction-types-stats /input/foaf.nt 

# execute our statistics on an HTTP uri
bash lodstats.sh -o restriction-types-stats http://xmlns.com/foaf/spec/index.rdf

please note that lodstats is called in the container, thus needed directories needs to be mounted in the docker-compose file or docker run command. In the shown example, the input is taken from the mounted input directory which is mounted in the docker-compose file.

How does LODStats work?

  • LODStats is a tool implemented by AKSW Leipzig, in python 2.7.
  • The tool parses input RDF and handles it as a stream.
  • Multiple stats modules exist (provided in the stats folder), and the commandline script LODStats/build/scripts-2.7/lodstats calls the LODStats library with provided stats modules.
  • For each triple the provided stats modules are executed.
  • A stats module follows a specific interface (LODStats/lodstats/stats/RDFStatInterface.py), thus new stats modules can inherit from that interface.
  • The count function is called for each triple, and information can be stored in own data structures.
  • The postproc function is called when the RDF stream ends, own aggregations can then be performed on self-defined data structures.

How does our extension work?

  • We adapted the lodstats script, to take also another stats directory into account (./src/lodstats.py with -o option).
  • Our stats modules for axioms are in the restriction-types-stats directory (including __init__.py, which makes the stats modules available
  • Our stats modules for constraints are in the shape-constraints-stats directory (including __init__.py, which makes the stats modules available

How to install LODStats?

# Create a docker image to execute LODStats with our extension
docker-compose build lodstats-execute

# Run the script
docker-compose run lodstats-execute /LODStats/scripts/lodstats

How to run tests?

# Create a docker image to run tests
docker-compose build lodstats-test

# Execute the tests (Does currently not work)
docker-compose up lodstats-test