def ready(self):
if 'UWKGM_STATE' in os.environ and os.environ[
'UWKGM_STATE'] == 'running':
verbose.info('Initializing NLTK wordnet...', KnowledgeConfig)
try:
_create_unverified_https_context = ssl._create_unverified_context
except AttributeError:
pass
else:
ssl._create_default_https_context = _create_unverified_https_context
nltk.download('wordnet')
verbose.info('Testing MongoDB database connection...')
pymongo.MongoClient('mongodb://%s:%d/' %
(env.resolve('database.mongo.address'),
env.resolve('database.mongo.port')))
verbose.info('MongoDB connection test pass')
def openie(text: str) -> List[Tuple[str, str, str]]:
"""Extracts triples using Stanford CoreNLP OpenIE library via CoreNLPConnector in Java REST API
:param text: A string to be extracted
:ex: Barack Obama born in Hawaii
:return: A list of triples
"""
client = env.resolve('servers.java')
verbose.info('Extracting triples using OpenIE at: ' + client['address'], caller=openie)
return requests.get('%s/openie/triples' % client['address'], params={'text': text}).json()
def dbpedia(triple: Tuple[str, str, str], predicate: str) -> bool:
"""Verifies domain-range agreement using DBpedia's ontology
:param triple: A triple (subject's entity, predicate, object's entity)
:ex: ["http://dbpedia.org/resource/Barack_Obama", "bear in", "http://dbpedia.org/resource/Hawaii"]
:param predicate: Entity of the predicate to be verified
:ex: http://dbpedia.org/ontology/birthPlace
:return: True if the predicate does not constitute domain-range violation
"""
# NOTE: Since DBpedia is expected to be the primary service for ontology look-up in this project,
# removing part of the URI helps reduce redundant information showed on a screen.
# The following lines of code should be modified when the service endpoint is changed.
config = env.resolve('database.virtuoso')
e = '%s:%d/sparql/' % (
config['address'], config['port']
) if 'port' in config else '%s/sparql/' % config['address']
verbose.info(
"Verifying domain-range of a predicate '%s' with SPARQL server: %s" %
(predicate.replace('http://dbpedia.org/ontology/', ''), e), dbpedia)
sparql = SPARQLWrapper(e)
sparql.setQuery('SELECT DISTINCT ?vr WHERE {{ '
'<{0}> <http://www.w3.org/2000/01/rdf-schema#range> ?rp. '
'<{1}> ?vr ?rp. }}'.format(predicate, triple[2]))
sparql.setReturnFormat(JSON)
try:
results = sparql.query().convert()
for result in results['results']['bindings']:
if len(result['vr']['value']) > 0:
return True
except QueryBadFormed as error:
verbose.error(str(error), dbpedia)
return False
# Domain-range violation concluded
return False
"""Virtuoso driver for graph database
The UWKGM project
:copyright: (c) 2020 Ichise Laboratory at NII & AIST
:author: Rungsiman Nararatwong
"""
from dorest import env
from SPARQLWrapper import SPARQLWrapper, JSON
from database.database.graph import default_graph_uri
config = env.resolve('database.virtuoso')
if 'port' in config:
client = SPARQLWrapper('%s:%s/sparql/' %
(config['address'], config['port']))
else:
client = SPARQLWrapper('%s/sparql/' % config['address'])
client.addDefaultGraph(default_graph_uri)
client.setReturnFormat(JSON)
"""MongoDB driver for triple-modifier database
The UWKGM project
:copyright: (c) 2020 Ichise Laboratory at NII & AIST
:author: Rungsiman Nararatwong
"""
import os
import pymongo
from dorest import env
mongo_config = env.resolve('database.mongo')
mongo_address = 'mongodb://%s:%d/' % (mongo_config['address'],
mongo_config['port'])
if 'UWKGM_MONGO_USERNAME' not in os.environ or 'UWKGM_MONGO_PASSWORD' not in os.environ \
or len(os.environ['UWKGM_MONGO_USERNAME']) == 0 or len(os.environ['UWKGM_MONGO_PASSWORD']) == 0:
mongo_client = pymongo.MongoClient(mongo_address)
else:
mongo_client = pymongo.MongoClient(
mongo_address,
username=os.environ['UWKGM_MONGO_USERNAME'],
password=os.environ['UWKGM_MONGO_PASSWORD'])
mongo = mongo_client['uwkgm']