softalias-kg

Authors: Hector López and Daniel Garijo

This repository contains the mappings, cleaning code and construct queries used for creating Softalias-KG. At the moment, the KG contains two main sources of software aliases:

The CZI software mention dataset (Dryad): A large dataset of software mentions in the biomedical literature containing thousands of aliases. This dataset consists on two sets of files. On the one hand, a set of pickle files describe the software mention groups matched to each external platform (Pypi, CRAN, etc.), together with a file assigning an id to each software mention in a publication. On the other hand, a set of CSVs contain the metadata for each software entry on each external platform. The input data used for this project are the .pkl files from the 'disambiguated' folder, the .csv files from 'linked' and the .pkl files from 'intermediate_files'. The folder czi_dataset/transformation_scripts contains all the Python scripts used to reading, cleaning, transforming and verifying the data (see the readme file for their execution).
Wikidata: we have enriched Softalias-KG with over 8K software applications (free software, wd:Q341) and more than 3k alternative labels (skos:altLabel) from Wikidata. The folder wikidata contains the python script with the queries used to retrieve software aliases from Wikidata. The category Software is not used directly because it incorporates many undesired results (e.g., videogames)

The folder vocab contains a description of the vocabulary used to describe aliases. Namely, we have class "SoftwareAlias" and the "alias" property (designed to indicate that a SoftwareApplication has a certain alias)

Requirements

The pipeline has been tested with python 3.9. The following libraries are required:

pandas
json
pickle
shutil
numpy
jaro (for Jaro-Winkler metric calculation)
Sparqlwrapper (for Wikidata construct queries)
morph_kgc

For CZI transformation details, please see the associated README file.

Knowledge Graph

The CZI dataset files resultant from ou pipeline are available in Zenodo (DOI: https://doi.org/10.5281/zenodo.7988427).
The Aliases, labels and metadata obtained from Wikidata are available in the wikidata folder.

A public SPARQL Fuseki endpoint has been made available here at https://softalias.linkeddata.es/softalias/sparql. To perform a query, just URL encode the query as shown below. For example, returning all the URLs associated with an alias:

PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
SELECT distinct ?url WHERE {
  ?soft <https://w3id.org/softalias/alias>/<https://schema.org/name> "spss" .
  ?soft <https://schema.org/url> ?url .
}

Results can be obtained with the following curl command:

curl https://softalias.linkeddata.es/softalias/sparql --data query=PREFIX%20rdf%3A%20%3Chttp%3A%2F%2Fwww.w3.org%2F1999%2F02%2F22-rdf-syntax-ns%23%3E%0APREFIX%20rdfs%3A%20%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23%3E%0ASELECT%20distinct%20%3Furl%20WHERE%20%7B%0A%20%20%3Fsoft%20%3Chttps%3A%2F%2Fw3id.org%2Fsoftalias%2Falias%3E%2F%3Chttps%3A%2F%2Fschema.org%2Fname%3E%20%22spss%22%20.%0A%20%20%3Fsoft%20%3Chttps%3A%2F%2Fschema.org%2Furl%3E%20%3Furl%20.%0A%7D%20 -X POST

The results should be something like:

{ "head": {
    "vars": [ "url" ]
  } ,
  "results": {
    "bindings": [
      { 
        "url": { "type": "literal" , "value": "https://scicrunch.org/browse/resources/SCR_002865" }
      } ,
      { 
        "url": { "type": "literal" , "value": "http://www-01.ibm.com/software/uk/analytics/spss/" }
      }
    ]
  }
}

SoftwareUnderstanding/softalias-kg

softalias-kg

Requirements

Knowledge Graph