SPARQL is a key component of the Semantic Web, and its reusability of is crucial for maximizing productivity on the Semantic Web. While an increasing number of datasets have been published in the Resource Description Framework (RDF) and the SPARQL standard allows for an interoperable framework for querying the RDF datasets, the reuse of SPARQL queries remains limited. Herein, we present SPANG, a framework used to facilitate the reuse of SPARQL queries by using a template engine, one-liner functionalities, and application programming interfaces. We demonstrate its use for making SPARQL queries more reusable; assign a set of metadata for annotating a SPARQL query, which enables the identi cation of the SPARQL query and provides useful information for the query; parameterize SPARQL queries, which makes each query reusable under di erent settings; assign a URI for each query; and construct a server for providing the indices of SPARQL queries. SPANG makes SPARQL queries more reusable in the local system or across the Web and facilitate the construction of an eco-system wherein SPARQL queries are shared to maximize the productivity of Semantic Web developers and users. SPANG is available at https://spang.dbcls.jp/.
An increasing number of datasets have been published in the Resource Description Framework (RDF). SPARQL is the standardized interface to RDF and contributes to the e ective reuse of distributed RDF data. Although eco-systems have been developed for developing SPARQL queries, writing SPARQL codes is often a burden for Semantic Web developers and users. SPARQL queries have been written in many projects and are the result of developers' e orts. The value of the SPARQL queries accumulated thus far can be maximized by reusing SPARQL queries. For the e ective construction of Semantic Web applications, the reuse of such queries is invaluable. However, owing to the limitation Copyright c 2020 for this paper by its author. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Usage: spang2 [options] <SPARQL_TEMPLATE> [par1=val1 par2=val2 ...] Options: -e, --endpoint <ENDPOINT> target SPARQL endpoint -o, --outfmt <FORMAT> tsv, json, n-triples (nt), turtle (ttl), rdf/xml (rdfxml), n3, xml, html -a, --abbr abbreviate results using predefined prefixes -v, --vars variable names are included in output (in the case of tsv format) -S, --subject <SUBJECT> shortcut for specifying subject -P, --predicate <PREDICATE> shortcut for specifying predicate -O, --object <OBJECT> shortcut for specifying object -L, --limit <LIMIT> LIMIT output (use by itself or with -[SPOF]) -F, --from <FROM> shortcut to search FROM specific graph (use alone or with -[SPOLN]) -N, --number shortcut to COUNT results (use by itself or with -[SPO]) -G, --graph shortcut for searching for graph names (use by itself or with -[SPO]) -r, --prefix <PREFIX_FILES> read prefix declarations (default: SPANG_DIR/etc/prefix,~/.spang/prefix) -n, --ignore ignore user-specific file (~/.spang/prefix) for test purpose -m, --method <METHOD> GET or POST (default: "GET") -q, --show_query show query and quit -f, --fmt format the query -i, --indent <DEPTH> indent depth; use with --fmt (default: 2) -l, --list_nick_name list available nicknames of endpoints and quit -d, --debug debug (output query embedded in URL, or output AST with --fmt) -V, --version output version number -h, --help output usage information a default value of C0751955 (\Brain Infarction"). # @title Get genes involved in a specific disease # @endpoint http://rdf.disgenet.org/sparql/ # @prefix https://raw.githubusercontent.com/hchiba1/spang/master/prefix/bio # @param arg1=C0751955 SELECT DISTINCT ?gene ?score ?gene_label ?source ?gda ?pmid ?description WHERE { ?gda sio:SIO_000628 umls:{{arg1}} , ?gene ; a ?type ; sio:SIO_000253 ?source ; sio:SIO_000216/sio:SIO_000300 ?score . ?gene a ncit:C16612 ;
rdfs:label ?gene_label .
OPTIONAL { ?gda sio:SIO_000772 ?pmid ;
dct:description ?description .
} } ORDER BY DESC(?score) ?source ?pmid The Unix command-line environment is available as an interface for the SPANG template library. While the SPANG templates can be stored in the local system, users can also call templates that are publicly available across the Web if each query is assigned a dereferenceable URI. The following example command line gives a parameter to a local template query le for the DisGeNET endpoint.
spang2 disease gene.rq disease=C0751955 The template can also be a URL that returns a SPANG template (https://raw. githubusercontent.com/hchiba1/spang/master/library/disgenet/disease_ gene.rq). Thus, anyone can host the query libraries on GitHub. Other use cases include one-liner functionalities, such as
spang2 -e disgenet -N -O ncbigene:2247 where the query counts the number of triples that have ncbigene:2247 as the object in the DisGeNET endpoint. The target endpoints and URI pre xes can be prede ned via con guration les. The following command prints the internally generated SPARQL query according to the command-line options to the standard output.
spang2 -e disgenet -N -O ncbigene:2247 -q
Most of these one-liner functionalities are compatible with previous versions of
SPANG [
Use through API The SPARQL templates can also be called in the JavaScript code used to construct a website. The documentation is available at the project home page. Users can also execute templates that are prede ned and provided by the SPANG server through the REST API. Accessing the following URI returns the result of the query with the given parameter: https://spang-portal.dbcls.jp/api/ library/disgenet/disease_gene.rq?arg1=C0751955. The following URI returns the list of queries in the library: https://spang-portal.dbcls.jp/api/ library/disgenet.
Users can also obtain the results on the SPANG server by clicking on the web pages while browsing the templates on the server (https://spang-portal. dbcls.jp/library/disgenet/disease_gene.rq). 4
The SPARQL templating syntax used in the SPANG framework is an extension
of sparql-doc [
We implemented the SPANG template engine in JavaScript, where the
template is transformed into an abstract syntax tree, processed using the given
parameters, and then reconstructed into SPARQL to be submitted to an endpoint.
We used Parsing Expression Grammar (PEG) [
We also developed a SPANG server, wherein the SPARQL queries can be executed in a browser or through a REST API (https://spang.dbcls.jp/). The SPANG server is developed using Ruby on Rails, and it can run on Docker. Thus, users can also run a SPANG server on their own. 5
SPARQL templating mechanisms have been implemented in various projects.
SPARQL libraries, such as Bioqueries [