BigCQ

A big synthetic dataset of CQ templates to SPARQL-OWL templates mappings.

What is that?

BigCQ is a dataset of Competency Question templates paired with SPARQL-OWL query templates. These represent templates of ontology requirements formalizations which are then translated into SPARQL-OWL query language used to query T-Box level of ontologies. Thus, such a dataset can be used in various scenarios regarding ontology authoring:

• Provide a large scale dataset for automatization of CQ involving tasks (automatic extraction of Glossary of Terms from requirements, automatic translation of CQs into queries to check how mature given ontology is).
• Allow to understand better the relation between human-language and ontology constructs.
• Make Competency Question driven ontology authoring more popular, since, although CQs are suggested in many ontology design methodologies, there is very limited set of CQs made publicly available.
• Provide guidelines on how CQs can be constructed to target given modelling styles.

How to cite that work?

We generated a persistent URI for that dataset using Zenodo:

Where can I find the data?

In BigCQ_dataset there are the following files and folders:

• cq_templates_only.txt -- a list of all generated 77575 unique CQ-templates.
• sparqlowl_templates_only.txt -- a list of all generated 549 unique SPARQL-OWL query templates.
• query_templates_to_cq_template_mappings -- a folder with mapping from SPARQL-OWL templates to CQ templates. Each file is represents a different SPARQL-OWL query template is represented by a JSON document following the following schema:

  {
    'query': 'SPARQL_OWL_QUERY',
    'cqs': ['CQ1', 'CQ2', ...]
  }

Is it possible to modify/extend the dataset?

Sure! Along with the dataset we published the Python code creating the dataset from scratch. Just enter dataset_preparation_scripts and type: PYTHONPATH=. python3 make_dataset.py in your terminal to regenerate the dataset. If you want to add some new CQ templates or synonym sets, you can find them in dataset_preparation_scripts/statements_to_cqs_transformations/:

• File synonym_classes defines various synonymes sets.
• Files with filenames starting with cq_general_templates define CQ templates for various needs.

Just edit these files and run make_dataset.py to make your new (better? :) ) dataset!

How to use the templates to test my ontology?

We provide a sample (proof of concept) code (using owlready2), that can parse an ontology and fill the placeholders with labels / IRIs. You can use it to adapt to your own needs.

You can find the materialize.py script in template_materialization_poc folder

Is there an input verbalizations with axiom shapes file provided?

Sure! Please look at verbalization2turtle.csv in the main folder of the repository.

What is the license?

MIT License.

What are the weak points of the dataset?

We made an evaluation on a separate set of CQs and SPARQL-OWL queries. The list of cases that are not covered by the dataset is provided in evaluation folder. We hope that in the next version of BigCQ it will be possible to handle them properly.

We hope that the dataset will help you in your work.