Hi everbody,
I try to send select and paths-queries via pystardog.
Example python-code:
*with stardog.Connection(self.stardog_database, *self.conn_details) as conn:
The triples are stored in the named graph "urn:named_graph/me/tco/plant_0780_year_2023".
To clarify the data structure in the db see the result from stardog studio:
query:
SELECT * WHERE
{
GRAPH ?g {
{?s ?p ?o}
}
}
Result:
should work like the FROM ... in the query - but it does not.
The above is also true for a paths-query issued by stardog (see above) and stardog studio.
As a workaround I can insert the FROM statement into the select query.
But for a path query this workaround is not possible - there i need the parameter
default_graph_uri=["urn:named_graph/me/tco/plant_0780_year_2023"]
Where is the error in using the default_graph_uri - parameter?
Thanks,
Thomas
I am not sure if I follow your code example since it is a bit difficult to read due to the formatting. You should be able to retrieve the from the named graph using the default_graph_uri . For example, if you insert data in the DB as follows
Hi Lars,
thanks for the quick reply.
Actually, what you write is exactly what i do - with one difference: I do write the triples to the db using the following code:
def write_file_to_graph(self, graph_uri: str, file_name: str):
with stardog.Connection(self.stardog_database, **self.conn_details) as conn:
conn.begin()
file_path = path.join(loadEnvVar('DAYDIR'), f'{file_name}_{date.today()}.ttl')
conn.add(content=stardog.content.File(file_path), graph_uri=graph_uri)
conn.commit()
i tried your solution but have to say, that it does not work.
I added the data to a db where there are already some data in the default graph.
So using stardog studio, i can differentiate between querying the default graph and the named graph
"urn:myNamedGraph".
Trying the same via the pystardog api always returns the data from the default graph. It seems that the parameter "default_graph_uri" is not effective.
Can you confirm that your solution does actually work (try-out done)?
Or might there be an implementation error in the pystardog module?
Thanks,
Thomas
I can confirm that the default_graph_uri parameter works as expected for pystardog. For example, when I execute the following example script:
import stardog
import stardog.content_types
conn_details = {
'endpoint': 'http://localhost:5820',
'username': 'user',
'password': 'pw'
}
data_dg = """
[] rdfs:label "Default Graph" .
"""
data_ng = """
[] rdfs:label "Named Graph" .
"""
with stardog.Connection('testing', **conn_details) as conn:
# Clear the data in the DB and add data to the default and named graph
conn.begin()
conn.clear()
conn.add(content=stardog.content.Raw(data_dg, 'text/turtle', name='dg.ttl'))
conn.add(content=stardog.content.Raw(data_ng, 'text/turtle', name='ng.ttl'), graph_uri="urn:ng")
conn.commit()
# Query the data
results = conn.select('select ?o { ?a ?p ?o }', content_type=stardog.content_types.CSV)
print(f"Default Graph: {results}")
results = conn.select('select ?o { ?a ?p ?o }', default_graph_uri=["urn:ng"], content_type=stardog.content_types.CSV)
print(f"Named Graph: {results}")
results = conn.select('select ?o from <urn:ng> { ?a ?p ?o }', content_type=stardog.content_types.CSV)
print(f"Named Graph (FROM): {results}")
results = conn.select('select ?o from named <urn:ng> { graph ?g { ?a ?p ?o } }', content_type=stardog.content_types.CSV)
print(f"Named Graph (FROM NAMED): {results}")
results = conn.select('select ?o from stardog:context:local { ?a ?p ?o }', content_type=stardog.content_types.CSV)
print(f"All local graphs: {results}")
I get the expected results:
Default Graph: b'o\nDefault Graph\n'
Named Graph: b'o\nNamed Graph\n'
Named Graph (FROM): b'o\nNamed Graph\n'
Named Graph (FROM NAMED): b'o\nNamed Graph\n'
All local graphs: b'o\nDefault Graph\nNamed Graph\n'
You should be able to reproduce these results on a simple DB for testing.
Maybe one thing to check in your code is that the provided graph URIs are not encoded in <...>, that is in SPARQL you would use FROM <urn:graph> but in the default_graph_uri parameter takes urn:graph instead.
again thanks a lot for this precise help.
I run your script with one change - i used an already existing database.
And my output corresponds to my former results (see line 2):
Default Graph: b'o\nDefault Graph\n'
Named Graph: b'o\nDefault Graph\n'
Named Graph (FROM): b'o\nNamed Graph\n'
Named Graph (FROM NAMED): b'o\nNamed Graph\n'
All local graphs: b'o\nDefault Graph\nNamed Graph\n'
So the problem seems not to be in the code but maybe in the database configuration?
Regards,
Thomas
you saved my week
I had the version 0.12.0. running.
After upgrading to 0.17.0 I finally get results from my named graphs!
Thanks a lot!
Best regards,
Thomas
