This repository contains reference serializations and a generation script for the Swift Vocabulary, an SKOS-based vocabulary on the Swift programming language. This is a work to appear at the "Resources Track" of ISWC 2020.
To run the generate.py script, you first need to install a couple of Python packages:
pip install beautifulsoup4 spacy spacy-lookups-data rdflib
python3 -m spacy download en_core_web_smNow you can run the script:
python3 generate.pyIt was tested on an Ubuntu 20.04 running Python 3.8, as well as on macOS Catalina (10.15.6) running Python 3.6. The execution takes less than a minute.
The generation script provides the following features:
- Extraction of concepts & resources from the "Swift book"
- Cleansing of concept names (removing non-alphabetical characters, applying camel case
- Creation of the RDF graph (metadata, concept scheme, concepts & labels, resources)
- Serialization to Turtle & XML format
Some manual curation is needed for the generated files. This can be done, e.g., using Protégé.
- Selection of final concepts
- Selection of best resources per concept
- Creation of associative links (
skos:related) and hierarchical links (skos:broader) - Determination of the scheme's top concepts (
skos:hasTopConcept) - Alignment with DBpedia (cf.
dbpedia.txtfor common programming-related DBpedia concepts)
Reference serializations in Turtle and XML format after manual curation can be found in this repository.
- Documentation
- SPARQL endpoint
- Swift Semantic: A Swift utility to retrieve the main concept from the Swift Vocabulary for a selection in a Swift source file. Based on SwiftSyntax.
- ALMA 4 Code: An extension for Visual Studio Code allowing to retrieve learning material related to a selected piece of code from the ALMA repository. Currently supporting the Swift programming language by using Swift Semantic.
- Swift Vocabulary as a dataset on the Linked Open Data Cloud
As a canonical citation, please use:
Grévisse, C. and Rothkugel, S. (2020). Swift Vocabulary. http://purl.org/lu/uni/alma/swift
The citation to the paper will be provided after its publication.