def sparql_query(
request: Request,
query: Optional[
str] = "SELECT * WHERE { <https://identifiers.org/OMIM:246300> <https://w3id.org/biolink/vocab/treated_by> ?drug . }"
):
# def sparql_query(query: Optional[str] = None):
"""
Send a SPARQL query to be executed.
- Example with a drug: https://identifiers.org/DRUGBANK:DB00394
- Example with a disease: https://identifiers.org/OMIM:246300
\f
:param query: SPARQL query input.
"""
if not query:
# TODO: return the SPARQL enndpoint service description
return {"SPARQL Service": "description"}
if request.headers['accept'] == 'text/csv':
# TODO: return in CSV format
return Response('a,b,c', media_type='text/csv')
else:
parsed_query = translateQuery(Query.parseString(query, parseAll=True))
query_operation = re.sub(r"(\w)([A-Z])", r"\1 \2",
parsed_query.algebra.name)
if query_operation != "Select Query":
return JSONResponse(
status_code=501,
content={"message": str(query_operation) + " not implemented"})
print(parsed_query)
print(query_operation)
predictions_list = query_classifier_from_sparql(parsed_query)
return predictions_list
def get_metadata(rq):
'''
Returns the metadata 'exp' parsed from the raw query file 'rq'
'exp' is one of: 'endpoint', 'tags', 'summary'
'''
yaml_string = "\n".join([row.lstrip('#+') for row in rq.split('\n') if row.startswith('#+')])
query_string = "\n".join([row for row in rq.split('\n') if not row.startswith('#+')])
query_metadata = yaml.load(yaml_string)
# If there is no YAML string
if query_metadata == None:
query_metadata = {}
query_metadata['query'] = query_string
try:
parsed_query = translateQuery(Query.parseString(rq, parseAll=True))
except ParseException:
app.logger.error("Could not parse query")
app.logger.error(query_string)
print traceback.print_exc()
query_metadata['type'] = parsed_query.algebra.name
if query_metadata['type'] == 'SelectQuery':
query_metadata['variables'] = parsed_query.algebra['PV']
return query_metadata
def get_metadata(rq, endpoint):
'''
Returns the metadata 'exp' parsed from the raw query file 'rq'
'exp' is one of: 'endpoint', 'tags', 'summary', 'request', 'pagination', 'enumerate'
'''
query_metadata = get_yaml_decorators(rq)
query_metadata['type'] = 'UNKNOWN'
try:
# THE PARSING
# select, describe, construct, ask
parsed_query = translateQuery(Query.parseString(rq, parseAll=True))
query_metadata['type'] = parsed_query.algebra.name
if query_metadata['type'] == 'SelectQuery':
# Projection variables
query_metadata['variables'] = parsed_query.algebra['PV']
# Parameters
query_metadata['parameters'] = get_parameters(rq, parsed_query.algebra['_vars'], endpoint, query_metadata)
elif query_metadata['type'] == 'ConstructQuery':
# Parameters
query_metadata['parameters'] = get_parameters(rq, parsed_query.algebra['_vars'], endpoint, query_metadata)
else:
glogger.warning("Query type {} is currently unsupported and no metadata was parsed!".format(query_metadata['type']))
except ParseException:
glogger.warning("Could not parse regular SELECT, CONSTRUCT, DESCRIBE or ASK query")
# glogger.warning(traceback.print_exc())
# insert queries won't parse, so we regex
# glogger.info("Trying to parse INSERT query")
# if static.INSERT_PATTERN in rq:
# query_metadata['type'] = 'InsertQuery'
# query_metadata['parameters'] = [u'_g_iri']
try:
# update query
glogger.info("Trying to parse UPDATE query")
parsed_query = UpdateUnit.parseString(rq, parseAll=True)
glogger.info(parsed_query)
query_metadata['type'] = parsed_query[0]['request'][0].name
if query_metadata['type'] == 'InsertData':
query_metadata['parameters'] = {'g': {'datatype': None, 'enum': [], 'lang': None, 'name': 'g', 'original': '?_g_iri', 'required': True, 'type': 'iri'},
'data': {'datatype': None, 'enum': [], 'lang': None, 'name': 'data', 'original': '?_data', 'required': True, 'type': 'literal'}}
glogger.info("Update query parsed with {}".format(query_metadata['type']))
# if query_metadata['type'] == 'InsertData':
# query_metadata['variables'] = parsed_query.algebra['PV']
except:
glogger.error("Could not parse query")
glogger.error(query_metadata['query'])
glogger.error(traceback.print_exc())
pass
glogger.debug("Finished parsing query of type {}".format(query_metadata['type']))
glogger.debug("All parsed query metadata (from decorators and content): ")
glogger.debug(pformat(query_metadata, indent=32))
return query_metadata
def get_parameters(rq):
"""
?_name The variable specifies the API mandatory parameter name. The value is incorporated in the query as plain literal.
?__name The parameter name is optional.
?_name_iri The variable is substituted with the parameter value as a IRI (also: number or literal).
?_name_en The parameter value is considered as literal with the language 'en' (e.g., en,it,es, etc.).
?_name_integer The parameter value is considered as literal and the XSD datatype 'integer' is added during substitution.
?_name_prefix_datatype The parameter value is considered as literal and the datatype 'prefix:datatype' is added during substitution. The prefix must be specified according to the SPARQL syntax.
"""
variables = translateQuery(Query.parseString(rq, parseAll=True)).algebra['_vars']
## Aggregates
internal_matcher = re.compile("__agg_\d+__")
## Basil-style variables
variable_matcher = re.compile("(?P<required>[_]{1,2})(?P<name>[^_]+)_?(?P<type>[a-zA-Z0-9]+)?_?(?P<userdefined>[a-zA-Z0-9]+)?.*$")
parameters = {}
for v in variables:
if internal_matcher.match(v):
continue
match = variable_matcher.match(v)
if match :
vname = match.group('name')
vrequired = True if match.group('required') == '_' else False
vtype = 'iri'
vlang = None
vdatatype = None
mtype = match.group('type')
muserdefined = match.group('userdefined')
if mtype in ['iri','number','literal']:
vtype = mtype
elif mtype:
vtype = 'literal'
if mtype:
if mtype in XSD_DATATYPES:
vdatatype = 'xsd:{}'.format(mtype)
elif len(mtype) == 2 :
vlang = mtype
elif muserdefined :
vdatatype = '{}:{}'.format(mtype, muserdefined)
parameters[vname] = {
'original': '?{}'.format(v),
'required': vrequired,
'name': vname,
'type': vtype,
'datatype': vdatatype,
'lang': vlang
}
return parameters
def get_metadata(rq):
'''
Returns the metadata 'exp' parsed from the raw query file 'rq'
'exp' is one of: 'endpoint', 'tags', 'summary', 'request', 'pagination', 'enumerate'
'''
query_metadata = get_yaml_decorators(rq)
try:
# select, describe, construct, ask
parsed_query = translateQuery(Query.parseString(rq, parseAll=True))
query_metadata['type'] = parsed_query.algebra.name
if query_metadata['type'] == 'SelectQuery':
query_metadata['variables'] = parsed_query.algebra['PV']
except ParseException:
glogger.warning("Could not parse SELECT, DESCRIBE, CONSTRUCT, ASK query")
# glogger.warning(traceback.print_exc())
pass
try:
# insert, update query
glogger.info("Trying to parse update query")
parsed_query = UpdateUnit.parseString(rq, parseAll=True)
glogger.info(parsed_query)
query_metadata['type'] = parsed_query[0]['request'][0].name
glogger.info("Update query parsed with {}".format(query_metadata['type']))
# if query_metadata['type'] == 'InsertData':
# query_metadata['variables'] = parsed_query.algebra['PV']
except:
glogger.error("Could not parse UPDATE query")
glogger.error(query_metadata['query'])
glogger.error(traceback.print_exc())
pass
glogger.info("Finished parsing query of type {}".format(query_metadata['type']))
return query_metadata
def process_shapes_file(shape_format, shapes_graph, rdf_file_path, repo_url,
branch, repo_description):
"""Process a file, check its content and add entry to the shapes graph
Large function, contain parsing for all formats: RDF, OBO, ShEx, OpenAPI, etc
"""
relative_filepath = str(rdf_file_path)[12:]
github_file_url = generate_github_file_url(repo_url, relative_filepath,
branch)
file_uri = URIRef(github_file_url)
shape_found = False
g = Graph()
if shape_format == 'obo':
# Get OBO ontologies
try:
graph = obonet.read_obo(github_file_url)
# for id_, data in graph.nodes(data=True):
for id_, data in graph.nodes(data=True):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add(
(file_uri, RDF.type, SIO['SIO_000623'])) # OBO ontology
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
shape_label = data.get('name')
if not shape_label:
shape_label = id_
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(shape_label)))
except Exception as e:
add_to_report('In repository: ' + repo_url + "\n> " + str(e),
github_file_url)
# Index OpenAPI files
elif shape_format == 'openapi':
try:
parser = ResolvingParser(github_file_url)
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SCHEMA['APIReference']))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
file_descriptions = []
if parser.specification['info']['title']:
file_descriptions.append(parser.specification['info']['title'])
if parser.specification['info']['description']:
file_descriptions.append(
parser.specification['info']['description'])
if len(file_descriptions) > 0:
shapes_graph.add((file_uri, RDFS.comment,
Literal(' - '.join(file_descriptions))))
# if not shape_label:
# shape_label = id_
# TODO: get operations hasPart?
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal('OpenAPI')))
except Exception as e:
pass
# TODO: YARRML? Search for prefixes and mappings at the root of YAML
# add_to_report('In repository: ' + repo_url + "\n> "
# + str(e), github_file_url)
# Search for shex files
elif shape_format == 'shex':
# No parsing possible for shex
shape_found = True
# TODO: use https://schema.org/SoftwareSourceCode ?
shapes_graph.add((file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SHEX.Schema))
shapes_graph.add((file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal('ShEx model')))
shapes_graph.add((file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
# Convert ShEx to RDF shex and parse it
# shex_rdf = ''
# if rdf_file_path.endswith('.shex'):
# with open(root / '../' + rdf_file_path, 'a') as f:
# shex_rdf = generate_shexj.parse(f.read())
# # if rdf_file_path.endswith('.shexj'):
# # with open(root / '../' + rdf_file_path, 'a') as f:
# # shex_rdf = f.read()
# logging.debug(shex_rdf)
# # for shape in g.subjects(RDF.type, SHEX.ShapeAnd):
# # add_shape_to_graph(shapes_graph, rdf_file_path, github_file_url, repo_url, shape, SHEX.schema)
# # for shape in g.subjects(RDF.type, SHEX.Shape):
# # add_shape_to_graph(shapes_graph, rdf_file_path, github_file_url, repo_url, shape, SHEX.schema)
# Parse SPARQL query files
elif shape_format == 'sparql':
# TODO: sparql+queries search failing might be due to a test SPARQL query hanging for long time
shape_found = True
shapes_graph.add((file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SH.SPARQLFunction))
shapes_graph.add((file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
try:
with open(rdf_file_path.absolute()) as file:
sparql_query = file.read()
# Parse SPARQL query (added fix for some malformed queries with =+ instead of #+)
sparql_query = "\n".join([
'#+' + row.lstrip('=+') for row in sparql_query.split('\n')
if row.startswith('=+')
])
yaml_string = "\n".join([
row.lstrip('#+') for row in sparql_query.split('\n')
if row.startswith('#+')
])
query_string = "\n".join([
row for row in sparql_query.split('\n')
if not row.startswith('#+')
])
shapes_graph.add(
(file_uri, SCHEMA['query'], Literal(sparql_query)))
grlc_metadata = {}
try: # Invalid YAMLs will produce empty metadata
grlc_metadata = yaml.load(yaml_string,
Loader=yaml.FullLoader)
except:
pass
# Get grlc query metadata
if grlc_metadata:
file_descriptions = []
if 'endpoint' in grlc_metadata:
sparql_endpoint = grlc_metadata['endpoint']
try:
shapes_graph.add((file_uri, VOID.sparqlEndpoint,
URIRef(sparql_endpoint)))
test_sparql_endpoint(sparql_endpoint)
except Exception as e:
logging.debug(
'Issue parsing SPARQL endpoint from .rq file')
logging.debug(e)
if 'summary' in grlc_metadata and grlc_metadata['summary']:
file_descriptions.append(grlc_metadata['summary'])
if 'description' in grlc_metadata and grlc_metadata[
'description']:
file_descriptions.append(grlc_metadata['description'])
# Add the query description to the graph
if len(file_descriptions) > 0:
shapes_graph.add(
(file_uri, RDFS.comment,
Literal(' - '.join(file_descriptions))))
# If default params described for grlc SPARQL query we add them as shapes
if 'defaults' in grlc_metadata:
for args in grlc_metadata['defaults']:
for arg, default_label in args.items():
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(arg)))
try:
# Parse the query to get its operation (select, construct..)
parsed_query = translateQuery(
Query.parseString(query_string, parseAll=True))
query_operation = re.sub(r"(\w)([A-Z])", r"\1 \2",
parsed_query.algebra.name)
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(query_operation)))
except:
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal('SPARQL Query')))
except:
logging.error('❌️ Issue opening file: ' + str(rdf_file_path))
# Parse RDF files
else:
try:
if shape_format == 'trig':
# Different graph required for trig to work
g = ConjunctiveGraph()
g.parse(str(rdf_file_path.absolute()), format=shape_format)
except Exception as e:
if shape_format == 'xml' and (str(rdf_file_path).endswith('.owl')
or
str(rdf_file_path).endswith('.rdf')):
# Try parsing with turtle for .owl and .rdf files
try:
g.parse(str(rdf_file_path.absolute()), format='ttl')
except:
add_to_report(
'RDF parsed as ' + shape_format + ', in repository: ' +
repo_url + "\n> " + str(e), github_file_url)
else:
add_to_report(
'RDF parsed as ' + shape_format + ', in repository: ' +
repo_url + "\n> " + str(e), github_file_url)
# Search for SHACL shapes
for shape in g.subjects(RDF.type, SH.NodeShape):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SH.Shape))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the shape
shape_label = label
# Fixing
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Search for CSV on the Web RDF (csvw)
# https://medium.swirrl.com/how-to-publish-csv-on-the-web-csvw-4ea6cbb603b4
# https://www.w3.org/ns/csvw
for shape_file in g.subjects(RDF.type, CSVW.Schema):
# for shape_file in g.objects(None, CSVW.tableSchema):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, CSVW.Schema))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
# Get file label
# for file_label in g.objects(shape_file, RDFS.label):
# shapes_graph.add((file_uri, RDFS.comment, Literal(str(file_label))))
# break
# Get columns label
for col_label in g.objects(shape_file, CSVW.column):
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(str(col_label))))
# Search for DCAT Datasets
for shape_file in g.subjects(RDF.type, DCAT.Dataset):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, DCAT.Dataset))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
# Get file label
for file_label in g.objects(shape_file, RDFS.label):
shapes_graph.add(
(file_uri, RDFS.comment, Literal(str(file_label))))
break
# shape_label = shape_file
# for label in g.objects(shape_file, RDFS.label):
# # Try to get the label of the shape
# shape_label = label
# # Fixing
# shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Search for nanopublication templates
for shape_file in g.subjects(RDF.type, NP_TEMPLATE.AssertionTemplate):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add(
(file_uri, RDF.type, NP_TEMPLATE.AssertionTemplate))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
# Get template label
for template_label in g.objects(shape_file, RDFS.label):
shapes_graph.add(
(file_uri, RDFS.comment, Literal(str(template_label))))
break
# TODO: get the shapes inside
nanopub_inputs = [
NP_TEMPLATE.GuidedChoicePlaceholder,
NP_TEMPLATE.LiteralPlaceholder,
NP_TEMPLATE.RestrictedChoicePlaceholder,
NP_TEMPLATE.UriPlaceholder
]
for np_input in nanopub_inputs:
for shape in g.subjects(RDF.type, np_input):
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the shape
shape_label = label
# Fixing
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Search for RML and R2RML mappings
for shape in g.subjects(RDF.type, R2RML.SubjectMap):
shape_found = True
is_rml_mappings = False
# Differenciate RML and R2RML mappings
if (None, RML.logicalSource, None) in g:
shapes_graph.add((file_uri, RDF.type, RML.LogicalSource))
else:
shapes_graph.add((file_uri, RDF.type, R2RML.TriplesMap))
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
shape_label = shape
# Try to get the label or URI of the subjectMap
for label in g.objects(shape, R2RML.template):
shape_label = label
for label in g.objects(shape, RDFS.label):
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Search for OWL classes
for shape in g.subjects(RDF.type, OWL.Class):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, OWL.Ontology))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the class
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Get rdfs:label of owl:Ontology and shaclTest:Validate for file description
file_descriptions = []
for shape in g.subjects(RDF.type, OWL.Ontology):
# Get one of the labels
for ontology_label in g.objects(shape, RDFS.label):
if len(file_descriptions) < 1:
file_descriptions.append(str(ontology_label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DC.title):
file_descriptions.append(str(label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DCTERMS.title):
file_descriptions.append(str(label))
# Now add the description
for comment in g.objects(shape, RDFS.comment):
file_descriptions.append(str(comment))
for label in g.objects(shape, RDFS.comment):
file_descriptions.append(str(label))
for description in g.objects(shape, DCTERMS.description):
file_descriptions.append(str(description))
for shape in g.subjects(
RDF.type, URIRef('http://www.w3.org/ns/shacl-test#Validate')):
for ontology_label in g.objects(shape, RDFS.label):
file_descriptions.append(str(ontology_label))
if len(file_descriptions) > 0:
shapes_graph.add((file_uri, RDFS.comment,
Literal(' - '.join(file_descriptions))))
# Get SKOS concepts and concept scheme
for shape in g.subjects(RDF.type, SKOS.Concept):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SKOS.ConceptScheme))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, SKOS.prefLabel):
# Try to get the label of the class
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
for shape in g.subjects(RDF.type, SKOS.ConceptScheme):
# Get one of the labels
for ontology_label in g.objects(shape, RDFS.label):
if len(file_descriptions) < 1:
file_descriptions.append(str(ontology_label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DC.title):
file_descriptions.append(str(label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DCTERMS.title):
file_descriptions.append(str(label))
# Now add the description
for comment in g.objects(shape, RDFS.comment):
file_descriptions.append(str(comment))
for label in g.objects(shape, RDFS.comment):
file_descriptions.append(str(label))
for description in g.objects(shape, DCTERMS.description):
file_descriptions.append(str(description))
# Search for ShEx Shapes and ShapeAnd
# TODO: Improve
for shape in g.subjects(RDF.type, SHEX.ShapeAnd):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SHEX.Schema))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the shape
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
for shape in g.subjects(RDF.type, SHEX.Shape):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SHEX.Schema))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add(
(file_uri, SCHEMA.codeRepository, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the shape
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Add the git repo to the graph
if shape_found:
logging.debug('[' + datetime.now().strftime("%m/%d/%Y, %H:%M:%S") +
'] ' + "✔️ Shape found in file " + github_file_url)
shapes_graph.add((URIRef(repo_url), RDF.type, SCHEMA['DataCatalog']))
shapes_graph.add(
(URIRef(repo_url), RDFS.label, Literal(repo_url.rsplit('/',
1)[1])))
if (repo_description):
shapes_graph.add(
(URIRef(repo_url), RDFS.comment, Literal(repo_description)))
return shapes_graph
def get_parameters(rq, endpoint):
"""
?_name The variable specifies the API mandatory parameter name. The value is incorporated in the query as plain literal.
?__name The parameter name is optional.
?_name_iri The variable is substituted with the parameter value as a IRI (also: number or literal).
?_name_en The parameter value is considered as literal with the language 'en' (e.g., en,it,es, etc.).
?_name_integer The parameter value is considered as literal and the XSD datatype 'integer' is added during substitution.
?_name_prefix_datatype The parameter value is considered as literal and the datatype 'prefix:datatype' is added during substitution. The prefix must be specified according to the SPARQL syntax.
"""
variables = translateQuery(Query.parseString(
rq, parseAll=True)).algebra['_vars']
## Aggregates
internal_matcher = re.compile("__agg_\d+__")
## Basil-style variables
variable_matcher = re.compile(
"(?P<required>[_]{1,2})(?P<name>[^_]+)_?(?P<type>[a-zA-Z0-9]+)?_?(?P<userdefined>[a-zA-Z0-9]+)?.*$"
)
parameters = {}
for v in variables:
if internal_matcher.match(v):
continue
match = variable_matcher.match(v)
# TODO: currently only one parameter per triple pattern is supported
tpattern_matcher = re.compile(
".*FROM\s+(?P<gnames>.*)\s+WHERE.*[\.\{][\n\t\s]*(?P<tpattern>.*\?"
+ re.escape(v) + ".*)\..*",
flags=re.DOTALL)
tp_match = tpattern_matcher.match(rq)
if match:
if tp_match:
vtpattern = tp_match.group('tpattern')
gnames = tp_match.group('gnames')
glogger.debug("Matched triple pattern with parameter")
# glogger.debug(vtpattern)
# glogger.debug(gnames)
codes_subquery = re.sub("SELECT.*\{.*\}",
"SELECT DISTINCT ?" + v + " FROM " +
gnames + " WHERE { " + vtpattern +
" . }",
rq,
flags=re.DOTALL)
headers = {'Accept': 'application/json'}
data = {'query': codes_subquery}
data_encoded = urllib.urlencode(data)
req = urllib2.Request(endpoint, data_encoded, headers)
glogger.debug("Sending code subquery request: " +
req.get_full_url() + "?" + req.get_data())
response = urllib2.urlopen(req)
codes_json = json.loads(response.read())
# glogger.debug(codes_json)
vcodes = []
for code in codes_json['results']['bindings']:
vcodes.append(code.values()[0]["value"])
# glogger.debug(vcodes)
vname = match.group('name')
vrequired = True if match.group('required') == '_' else False
vtype = 'literal'
vlang = None
vdatatype = None
mtype = match.group('type')
muserdefined = match.group('userdefined')
if mtype in ['iri', 'number', 'literal']:
vtype = mtype
elif mtype:
vtype = 'literal'
if mtype:
if mtype in static.XSD_DATATYPES:
vdatatype = 'xsd:{}'.format(mtype)
elif len(mtype) == 2:
vlang = mtype
elif muserdefined:
vdatatype = '{}:{}'.format(mtype, muserdefined)
parameters[vname] = {
'original': '?{}'.format(v),
'required': vrequired,
'name': vname,
'enum': sorted(vcodes),
'type': vtype,
'datatype': vdatatype,
'lang': vlang
}
return parameters
def process_shapes_file(shape_format, shapes_graph, rdf_file_path, repo_url,
branch, repo_description):
"""Process a Shapes file, check its content and add entry to the shapes graph
Large function, contain parsing for all formats: RDF, OBO, ShEx, OpenAPI...
"""
relative_filepath = str(rdf_file_path)[12:]
github_file_url = generate_github_file_url(repo_url, relative_filepath,
branch)
file_uri = URIRef(github_file_url)
shape_found = False
g = Graph()
if shape_format == 'obo':
# Get OBO ontologies
try:
graph = obonet.read_obo(github_file_url)
# for id_, data in graph.nodes(data=True):
for id_, data in graph.nodes(data=True):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add(
(file_uri, RDF.type, SIO['SIO_000623'])) # OBO ontology
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
shape_label = data.get('name')
if not shape_label:
shape_label = id_
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(shape_label)))
except Exception as e:
# print('[' + datetime.now().strftime("%m/%d/%Y, %H:%M:%S") + '] 🗑 Issue with OBO parser for file ' + github_file_url)
add_to_report('File: ' + github_file_url + "\n\n" +
'In repository: ' + repo_url + "\n> " + str(e) +
"\n\n---\n")
# Index OpenAPI files
elif shape_format == 'openapi':
try:
parser = ResolvingParser(github_file_url)
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SCHEMA['WebAPI']))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
file_descriptions = []
if parser.specification['info']['title']:
file_descriptions.append(parser.specification['info']['title'])
if parser.specification['info']['description']:
file_descriptions.append(
parser.specification['info']['description'])
if len(file_descriptions) > 0:
shapes_graph.add((file_uri, DC.description,
Literal(' - '.join(file_descriptions))))
# if not shape_label:
# shape_label = id_
# TODO: get operations hasPart?
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal('OpenAPI')))
except Exception as e:
pass
# print('[' + datetime.now().strftime("%m/%d/%Y, %H:%M:%S") + '] 🗑 Issue with OpenAPI parser for file ' + github_file_url)
# print(e)
# add_to_report('File: ' + github_file_url + "\n\n"
# + 'In repository: ' + repo_url + "\n> "
# + str(e) + "\n\n---\n")
# Search for shex files
elif shape_format == 'shex':
# no parsing possible for shex
shape_found = True
# TODO: use https://schema.org/SoftwareSourceCode ?
shapes_graph.add((file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SHEX.Schema))
shapes_graph.add((file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal('ShEx model')))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
# Convert ShEx to RDF shex and parse it
# shex_rdf = ''
# if rdf_file_path.endswith('.shex'):
# with open(root / '../' + rdf_file_path, 'a') as f:
# shex_rdf = generate_shexj.parse(f.read())
# # if rdf_file_path.endswith('.shexj'):
# # with open(root / '../' + rdf_file_path, 'a') as f:
# # shex_rdf = f.read()
# print(shex_rdf)
# # for shape in g.subjects(RDF.type, SHEX.ShapeAnd):
# # add_shape_to_graph(shapes_graph, rdf_file_path, github_file_url, repo_url, shape, SHEX.schema)
# # for shape in g.subjects(RDF.type, SHEX.Shape):
# # add_shape_to_graph(shapes_graph, rdf_file_path, github_file_url, repo_url, shape, SHEX.schema)
# Parse SPARQL query files
elif shape_format == 'sparql':
shape_found = True
shapes_graph.add((file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SH.SPARQLFunction))
shapes_graph.add((file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
with open(rdf_file_path.absolute()) as file:
sparql_query = file.read()
# Parse SPARQL query
yaml_string = "\n".join([
row.lstrip('#+') for row in sparql_query.split('\n')
if row.startswith('#+')
])
query_string = "\n".join([
row for row in sparql_query.split('\n')
if not row.startswith('#+')
])
shapes_graph.add(
(file_uri, SCHEMA['query'], Literal(query_string)))
grlc_metadata = {}
try: # Invalid YAMLs will produce empty metadata
grlc_metadata = yaml.load(yaml_string, Loader=yaml.FullLoader)
except:
pass
# Get metadata like grlc metadata
if grlc_metadata:
file_descriptions = []
if 'endpoint' in grlc_metadata:
sparql_endpoint = grlc_metadata['endpoint']
shapes_graph.add((file_uri, VOID.sparqlEndpoint,
Literal(sparql_endpoint)))
# TODO: check if in hashes of already tested endpoints valid and failing3
# Test endpoint with SPARQLWrapper, add it to hash of valid or failing endpoints
# Then, like repos, add them as schema:EntryPoint
if sparql_endpoint not in VALID_ENDPOINTS.keys(
) and sparql_endpoint not in FAILED_ENDPOINTS.keys():
sparql_test_query = 'SELECT * WHERE { ?s ?p ?o } LIMIT 10'
sparql = SPARQLWrapper(sparql_endpoint)
sparql.setReturnFormat(JSON)
sparql.setQuery(sparql_test_query)
try:
results = sparql.query().convert()
# Check SPARQL query sent back at least 5 triples
results_array = results["results"]["bindings"]
if len(results_array) > 4:
VALID_ENDPOINTS[sparql_endpoint] = {
'label': sparql_endpoint
}
else:
FAILED_ENDPOINTS[sparql_endpoint] = 'failed'
except Exception as e:
add_to_report('SPARQL endpoint failed: ' +
sparql_endpoint + "\n\n" + str(e) +
"\n\n---\n")
if 'summary' in grlc_metadata and grlc_metadata['summary']:
file_descriptions.append(grlc_metadata['summary'])
if 'description' in grlc_metadata and grlc_metadata[
'description']:
file_descriptions.append(grlc_metadata['description'])
if len(file_descriptions) > 0:
shapes_graph.add((file_uri, DC.description,
Literal(' - '.join(file_descriptions))))
# If default params described for grlc SPARQL query we add then as shapes
if 'defaults' in grlc_metadata:
for args in grlc_metadata['defaults']:
for arg, default_label in args.items():
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(arg)))
try:
# Parse query to get its operation (select, construct..)
parsed_query = translateQuery(
Query.parseString(query_string, parseAll=True))
query_operation = re.sub(r"(\w)([A-Z])", r"\1 \2",
parsed_query.algebra.name)
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal(query_operation)))
except:
shapes_graph.add(
(file_uri, DCTERMS.hasPart, Literal('SPARQL Query')))
# Parse RDF files
else:
try:
g.parse(str(rdf_file_path.absolute()), format=shape_format)
except Exception as e:
print('[' + datetime.now().strftime("%m/%d/%Y, %H:%M:%S") +
'] 🗑 RDF parser for ' + shape_format +
' did not worked for the file ' + github_file_url)
if not str(rdf_file_path).endswith('.xml') and not str(
rdf_file_path).endswith('.json'):
add_to_report('File: ' + github_file_url + " parsed as " +
shape_format + "\n\n" + 'In repository: ' +
repo_url + "\n> " + str(e) + "\n\n---\n")
# Search for SHACL shapes
for shape in g.subjects(RDF.type, SH.NodeShape):
# add_shape_to_graph(shapes_graph, rdf_file_path, github_file_url, repo_url, shape_uri, shape_type)
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SH.Shape))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the shape
shape_label = label
# Fixing
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Search for OWL classes, limit to max 300 classes/concepts retrieved
classes_limit = 300
classes_count = 0
for shape in g.subjects(RDF.type, OWL.Class):
# add_shape_to_graph(shapes_graph, rdf_file_path, github_file_url, repo_url, shape_uri, shape_type)
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, OWL.Ontology))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the class
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
classes_count += 1
if classes_count >= classes_limit:
break
# Get rdfs:label of owl:Ontology and shaclTest:Validate for file description
file_descriptions = []
for shape in g.subjects(RDF.type, OWL.ontology):
# Get one of the labels
for ontology_label in g.objects(shape, RDFS.label):
if len(file_descriptions) < 1:
file_descriptions.append(str(ontology_label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DC.title):
file_descriptions.append(str(label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DCTERMS.title):
file_descriptions.append(str(label))
# Now add the description
for comment in g.objects(shape, RDFS.comment):
file_descriptions.append(str(comment))
for label in g.objects(shape, DC.description):
file_descriptions.append(str(label))
for description in g.objects(shape, DCTERMS.description):
file_descriptions.append(str(description))
for shape in g.subjects(
RDF.type, URIRef('http://www.w3.org/ns/shacl-test#Validate')):
for ontology_label in g.objects(shape, RDFS.label):
file_descriptions.append(str(ontology_label))
if len(file_descriptions) > 0:
shapes_graph.add((file_uri, DC.description,
Literal(' - '.join(file_descriptions))))
# Get SKOS concepts and concept scheme
classes_count = 0
for shape in g.subjects(RDF.type, SKOS.Concept):
# add_shape_to_graph(shapes_graph, rdf_file_path, github_file_url, repo_url, shape_uri, shape_type)
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SKOS.ConceptScheme))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, SKOS.prefLabel):
# Try to get the label of the class
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
classes_count += 1
if classes_count >= classes_limit:
break
for shape in g.subjects(RDF.type, SKOS.ConceptScheme):
# Get one of the labels
for ontology_label in g.objects(shape, RDFS.label):
if len(file_descriptions) < 1:
file_descriptions.append(str(ontology_label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DC.title):
file_descriptions.append(str(label))
if len(file_descriptions) == 0:
for label in g.objects(shape, DCTERMS.title):
file_descriptions.append(str(label))
# Now add the description
for comment in g.objects(shape, RDFS.comment):
file_descriptions.append(str(comment))
for label in g.objects(shape, DC.description):
file_descriptions.append(str(label))
for description in g.objects(shape, DCTERMS.description):
file_descriptions.append(str(description))
# Search for ShEx Shapes and ShapeAnd
# TODO: Improve
for shape in g.subjects(RDF.type, SHEX.ShapeAnd):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SHEX.Schema))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the shape
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
for shape in g.subjects(RDF.type, SHEX.Shape):
shape_found = True
shapes_graph.add(
(file_uri, RDF.type, SCHEMA['SoftwareSourceCode']))
shapes_graph.add((file_uri, RDF.type, SHEX.Schema))
shapes_graph.add(
(file_uri, RDFS.label, Literal(rdf_file_path.name)))
shapes_graph.add((file_uri, DC.source, URIRef(repo_url)))
shape_label = shape
for label in g.objects(shape, RDFS.label):
# Try to get the label of the shape
shape_label = label
shapes_graph.add((file_uri, DCTERMS.hasPart, Literal(shape_label)))
# Add repository RDF
if shape_found:
shapes_graph.add(
(URIRef(repo_url), RDF.type, SCHEMA['codeRepository']))
# TODO: change, schema:codeRepository is a property, not a class, but not much available..
shapes_graph.add(
(URIRef(repo_url), RDFS.label, Literal(repo_url.rsplit('/',
1)[1])))
if (repo_description):
shapes_graph.add(
(URIRef(repo_url), RDFS.comment, Literal(repo_description)))
return shapes_graph