Ad-hoc light weight SPARQL endpoint from a file, using Python Flask and RDFLib
Sometimes one needs to set up a quick SPARQL endpoint for a local Turtle file without installing a triplestore. This is a quick demo how to use the Python 3 microframework Flask and RDFLib to provide an Ad-Hoc SPARQL endpoint based on a Turtle or RDF/XML file.
adhs requires python > 3. You can install dependencies with pip
:
pip install -r requirements.txt
Example:
./adhs.py file.ttl -i turtle -p 5000
The -i
parameter is optional but should be used to specify the format of the file. If it's missing, rdflib.util.guess_format()
will be used.
The parameter --host
is optional too, and defines the interfaces where the hosts is listening (127.0.0.1
for local only, 0.0.0.0
for all interfaces).
The -p
parameter is optional as well. If not provided, flask will be accessible on default port 5000. Navigate to
http://localhost:5000/sparql
to get a very basic form to enter and submit your queries (the form is just some icing on the cake).
Additionally, the SPARQL endpoint accepts GET and POST requests.
Example of a GET request via URL:
http://localhost:5000/sparql?query=select distinct ?s where { ?s a ?c }&format=<CONTENT-TYPE>
The format
has to be specified via the accept type (see below) but in future there might be more fine-grained options.
Note: Instead of format
, the parameter output
can be used as well.
Example with cURL:
curl -d "query=select distinct ?c where {?s a ?c}" \
-H "Accept: text/html" \
localhost:5000/sparql
The content type has to be application/x-www-form-urlencoded
, whereas the accept type can be one of the following:
text/html
application/sparql-results+json
application/sparql-results+xml
Additionally, the accept type can also be set with the format
(or alternatively) output
parameter.
If no accept type is specified, the default type is text/html
unless it's overridden with format
or output
. When an accept type is set via content negotiation and as well with the format
or output
parameter, the parameter takes precedence over the content negotiation.
adhs is also available as an automatic image build at hub.docker.com, so you can serve a local file with the following docker command:
docker run -i -t --rm -p 80:80 -v $PWD:/data -e ADHS_FILE=/data/ontology.ttl eccenca/adhs:latest
adhs docker understands the following environment variables:
ADHS_FILE
corresponds to the file parameter of adhsADHS_INPUT
corresponds to the optional input parameter -i
.In dockerized adhs, the port is set to 80
and the host to 0.0.0.0
(all interfaces).
Volume is set to /data
so you can mount your data directory into it.
In addition to that /opt/adhs/templates
is also a volume if you want to overwrite the templates.
One use case behind the adhs was to provide a quick possibility to set up different ontologies to test them with the Visual SPARQL Builder (VSB) without having to load them into a triplestore like Virtuoso. Additionally, since the VSB needs SPARQL 1.1, it is a convenient alternative on systems don't have high enough version of Virtuoso installed.
ADHS is actually the German abbreviation for ADHD.
moved to TODO.md