def void_dataset(graph_name):
"""Describe one RDF dataset"""
try:
logger.debug(
'[/void/] Loading VoID descriptions for dataset {}'.format(
graph_name))
graph = dataset.get_graph(graph_name)
if graph is None:
abort(404)
logger.debug('[/void/] Corresponding dataset found')
mimetype = request.accept_mimetypes.best_match([
"application/n-triples", "text/turtle", "application/xml",
"application/n-quads", "application/trig", "application/json",
"application/json+ld"
])
url = secure_url(request.url_root)
if url.endswith('/'):
url = url[0:len(url) - 1]
descriptor = VoidDescriptor(url, graph)
format, mimetype = choose_format(mimetype)
return Response(descriptor.describe(format), content_type=mimetype)
except Exception as e:
logger.error(e)
abort(500)
def index():
try:
url = secure_url(request.url)
dinfos = [dinfo for dinfo in dataset.describe(url)]
long_description = Markup(markdown(dataset.long_description))
return render_template("index_sage.html", dataset=dinfos, server_public_url=dataset.public_url, default_query=dataset.default_query, long_description=long_description)
except Exception as e:
gunicorn_logger.error(e)
abort(500)
def void_all():
"""Describe all RDF datasets hosted by the Sage endpoint"""
try:
mimetype = request.accept_mimetypes.best_match([
"application/n-triples", "text/turtle", "application/xml",
"application/n-quads", "application/trig", "application/json",
"application/json+ld"
])
url = secure_url(request.url_root)
if url.endswith('/'):
url = url[0:len(url) - 1]
format, mimetype = choose_format(mimetype)
description = many_void(url, datasets, format)
return Response(description, content_type=mimetype)
except Exception:
abort(500)
def index():
try:
datasets_infos = datasets._config["datasets"]
url = secure_url(request.url)
api_doc = {
"@context": "http://www.w3.org/ns/hydra/context.jsonld",
"@id": url,
"@type": "ApiDocumentation",
"title": "SaGe SPARQL API",
"description": "A SaGe interface which allow evaluation of SPARQL queries over RDF datasets",
"entrypoint": "{}sparql".format(url),
"supportedClass": []
}
for dinfo in datasets.describe(url):
api_doc["supportedClass"].append(dinfo)
long_description = Markup(markdown(datasets.long_description))
return render_template("index_sage.html", datasets=datasets_infos, api=api_doc, server_public_url=datasets.public_url, default_query=datasets.default_query, long_description=long_description)
except Exception as e:
print(e)
abort(500)
def sparql_index():
mimetype = request.accept_mimetypes.best_match([
"application/json", "application/xml",
"application/sparql-results+json",
"application/sparql-results+xml", "text/html"
])
try:
url = secure_url(request.base_url)
# parse arguments
if request.method == "GET":
query = request.args.get("query") or None
default_graph_uri = request.args.get(
"default-graph-uri") or None
next_link = request.args.get("next") or None
# ensure that both the query and default-graph-uri params are set
if (query is None or default_graph_uri is None) and (
next_link is None or default_graph_uri is None):
return sage_http_error(
"Invalid request sent to server: a GET request must contains both parameters 'query' and 'default-graph-uri'. See <a href='http://sage.univ-nantes.fr/documentation'>the API documentation</a> for reference."
)
elif request.method == "POST" and request.is_json:
# POST query
post_query, err = SageSparqlQuery().load(request.get_json())
if err is not None and len(err) > 0:
# TODO better formatting
return Response(format_marshmallow_errors(err), status=400)
query = post_query["query"]
default_graph_uri = post_query["defaultGraph"]
next_link = post_query["next"] if 'next' in post_query else None
else:
return sage_http_error(
"Invalid request sent to server: a GET request must contains both parameters 'query' and 'default-graph-uri'. See <a href='http://sage.univ-nantes.fr/documentation'>the API documentation</a> for reference."
)
# execute query
return execute_query(query, default_graph_uri, next_link, dataset,
mimetype, url)
except Exception as e:
logger.error(e)
abort(500)
def sparql_index():
try:
mimetype = request.accept_mimetypes.best_match(
["application/json", "text/html"])
url = secure_url(request.url)
api_doc = {
"@context": "http://www.w3.org/ns/hydra/context.jsonld",
"@id": url,
"@type": "ApiDocumentation",
"title": "SaGe SPARQL API",
"description":
"A SaGe interface which allow evaluation of SPARQL queries over RDF datasets",
"entrypoint": "{}sparql".format(url),
"supportedClass": []
}
for dinfo in datasets.describe(url):
api_doc["supportedClass"].append(dinfo)
if mimetype is "text/html":
return render_template("interfaces.html", api=api_doc)
return json.jsonify(api_doc)
except Exception:
abort(500)
def sparql_query(graph_name):
"""WARNING: old API, deprecated"""
graph = dataset.get_graph(graph_name)
if graph is None:
abort(404)
logger.debug('[/sparql/] Corresponding dataset found')
mimetype = request.accept_mimetypes.best_match([
"application/json", "application/xml",
"application/sparql-results+json", "application/sparql-results+xml"
])
url = secure_url(request.url)
try:
# A GET request always returns the homepage of the dataset
if request.method == "GET" or (not request.is_json):
dinfo = graph.describe(url)
dinfo['@id'] = url
void_desc = {
"nt": VoidDescriptor(url, graph).describe("ntriples"),
"ttl": VoidDescriptor(url, graph).describe("turtle"),
"xml": VoidDescriptor(url, graph).describe("xml")
}
return render_template("website/sage_dataset.html",
dataset_info=dinfo,
void_desc=void_desc)
engine = SageEngine()
post_query, err = QueryRequest().load(request.get_json())
if err is not None and len(err) > 0:
return Response(format_marshmallow_errors(err), status=400)
quota = graph.quota / 1000
max_results = graph.max_results
# Load next link
next_link = None
if 'next' in post_query:
next_link = decode_saved_plan(post_query["next"])
# build physical query plan, then execute it with the given quota
start = time()
plan, cardinalities = build_query_plan(post_query["query"],
dataset, graph_name,
next_link)
loading_time = (time() - start) * 1000 # convert in milliseconds
bindings, saved_plan, is_done = engine.execute(
plan, quota, max_results)
# compute controls for the next page
start = time()
next_page = None
if not is_done:
next_page = encode_saved_plan(saved_plan)
exportTime = (time() - start) * 1000 # convert in milliseconds
stats = {
"cardinalities": cardinalities,
"import": loading_time,
"export": exportTime
}
if mimetype == "application/sparql-results+json":
res = Response(responses.w3c_json_streaming(
bindings, next_page, stats, url),
content_type='application/json')
if mimetype == "application/xml" or mimetype == "application/sparql-results+xml":
res = Response(responses.w3c_xml(bindings, next_page, stats),
content_type="application/xml")
else:
res = Response(responses.raw_json_streaming(
bindings, next_page, stats, url),
content_type='application/json')
# set deprecation warning in headers
res.headers.add(
"Warning",
"199 SaGe/2.0 \"You are using a deprecated API. Consider uppgrading to the SaGe SPARQL query API. See http://sage.univ-nantes.fr/documentation fore more details.\""
)
return res
except Exception as e:
logger.error(e)
abort(500)
def lookup_entity(graph_name, entity):
"""Evaluates a DESCRIBE query over a RDF dataset"""
logger.debug('[IP: {}] [/lookup/] Querying {}'.format(
request.environ['REMOTE_ADDR'], graph_name))
graph = dataset.get_graph(graph_name)
if graph is None:
abort(404)
url = secure_url(request.url)
try:
engine = SageEngine()
# Get entity and possible next link
entity_uri = secure_url(request.base_url)
next_link = request.args.get("next", default=None)
post_query = build_describe_query(entity_uri)
logger.debug('[IP: {}] [/lookup/] Entity={}'.format(
request.environ['REMOTE_ADDR'], entity_uri))
quota = graph.quota / 1000
max_results = graph.max_results
# Load next link
if next_link is not None:
logger.debug(
'[/lookup/{}] Saved plan found, decoding "next" link'.
format(graph_name))
next_link = decode_saved_plan(next_link)
else:
logger.debug('[/lookup/{}] Query to evaluate: {}'.format(
graph_name, post_query))
# build physical query plan, then execute it with the given quota
logger.debug(
'[/lookup/{}] Starting query evaluation...'.format(graph_name))
# start = time()
plan, cardinalities = build_query_plan(post_query, dataset,
graph_name, next_link)
# loading_time = (time() - start) * 1000
bindings, saved_plan, is_done = engine.execute(
plan, quota, max_results)
logger.debug(
'[/lookup/{}] Query evaluation completed'.format(graph_name))
# compute controls for the next page
# start = time()
next_page = None
if is_done:
logger.debug(
'[/lookup/{}] Query completed under the time quota'.format(
graph_name))
else:
logger.debug(
'[/lookup/{}] The query was not completed under the time quota...'
.format(graph_name))
logger.debug(
'[/lookup/{}] Saving the execution to plan to generate a "next" link'
.format(graph_name))
next_page = encode_saved_plan(saved_plan)
logger.debug(
'[/lookup/{}] "next" link successfully generated'.format(
graph_name))
# exportTime = (time() - start) * 1000
# stats = {"import": loading_time, "export": exportTime}
triples = bindings_to_triple(entity_uri, bindings, url)
headers = dict()
if next_page is not None:
headers["X-Sage-Next"] = "{}?next={}".format(
entity_uri, next_page)
headers[
"Link"] = "<{}?next={}>; rel=\"next\"; title=\"Next page\"".format(
entity_uri, next_page)
return Response(responses.ntriples_streaming(triples),
content_type="application/ntriples",
headers=headers)
except Exception as e:
logger.error(e)
abort(500)
def sparql_query(dataset_name):
logger.info('[IP: {}] [/sparql/] Querying {}'.format(
request.environ['REMOTE_ADDR'], dataset_name))
dataset = datasets.get_dataset(dataset_name)
if dataset is None:
abort(404)
logger.debug('[/sparql/] Corresponding dataset found')
mimetype = request.accept_mimetypes.best_match([
"application/json", "application/xml",
"application/sparql-results+json", "application/sparql-results+xml"
])
url = secure_url(request.url)
try:
# A GET request always returns the homepage of the dataset
if request.method == "GET" or (not request.is_json):
dinfo = dataset.describe(url)
dinfo['@id'] = url
void_desc = {
"nt": VoidDescriptor(url, dataset).describe("ntriples"),
"ttl": VoidDescriptor(url, dataset).describe("turtle"),
"xml": VoidDescriptor(url, dataset).describe("xml")
}
return render_template("sage.html",
dataset_info=dinfo,
void_desc=void_desc)
engine = SageEngine()
post_query, err = QueryRequest().load(request.get_json())
if err is not None and len(err) > 0:
return Response(format_marshmallow_errors(err), status=400)
logger.info('[IP: {}] [/sparql/] Query={}'.format(
request.environ['REMOTE_ADDR'], post_query))
quota = dataset.quota / 1000
max_results = dataset.max_results
# Load next link
next_link = None
if 'next' in post_query:
logger.debug(
'[/sparql/{}] Saved plan found, decoding "next" link'.
format(dataset_name))
next_link = decode_saved_plan(post_query["next"])
else:
logger.debug('[/sparql/{}] Query to evaluate: {}'.format(
dataset_name, post_query))
# build physical query plan, then execute it with the given quota
logger.debug('[/sparql/{}] Starting query evaluation...'.format(
dataset_name))
start = time()
plan, cardinalities = build_query_plan(post_query["query"],
dataset, next_link)
loading_time = (time() - start) * 1000
bindings, saved_plan, is_done = engine.execute(
plan, quota, max_results)
logger.debug(
'[/sparql/{}] Query evaluation completed'.format(dataset_name))
# compute controls for the next page
start = time()
next_page = None
if is_done:
logger.debug(
'[/sparql/{}] Query completed under the time quota'.format(
dataset_name))
else:
logger.debug(
'[/sparql/{}] The query was not completed under the time quota...'
.format(dataset_name))
logger.debug(
'[/sparql/{}] Saving the execution to plan to generate a "next" link'
.format(dataset_name))
next_page = encode_saved_plan(saved_plan)
logger.debug(
'[/sparql/{}] "next" link successfully generated'.format(
dataset_name))
exportTime = (time() - start) * 1000
stats = {
"cardinalities": cardinalities,
"import": loading_time,
"export": exportTime
}
if mimetype == "application/sparql-results+json":
return json.jsonify(responses.json(bindings, next_page, stats))
if mimetype == "application/xml" or mimetype == "application/sparql-results+xml":
return Response(responses.xml(bindings, next_page, stats),
content_type="application/xml")
return json.jsonify(responses.raw_json(bindings, next_page, stats))
except Exception:
abort(500)