"w") as f:
for e in projection.entityToPreferredLabels:
for v in projection.entityToPreferredLabels[e]:
f.write("%s preferred_label %s\n" % (e, v))
for a in annotations:
f.write("%s\n" % " ".join(a))
# read URI document
# two parts: walks, axioms (if the axiom file exists)
walk_sentences, axiom_sentences, URI_Doc = list(), list(), list()
if "URI_Doc" in config["DOCUMENT"] and config["DOCUMENT"]["URI_Doc"] == "yes":
print("\nGenerate URI document ...")
# walker_type=config['DOCUMENT']['walker']
walks_ = get_rdf2vec_walks(
onto_file=ontology_file,
walker_type=config["DOCUMENT"]["walker"],
walk_depth=int(config["DOCUMENT"]["walk_depth"]),
classes=entities,
)
print("Extracted %d walks for %d seed entities" %
(len(walks_), len(entities)))
walk_sentences += [list(map(str, x)) for x in walks_]
axiom_file = os.path.join(config["DOCUMENT"]["cache_dir"], "axioms.txt")
if os.path.exists(axiom_file):
for line in open(axiom_file).readlines():
axiom_sentence = [item for item in line.strip().split()]
axiom_sentences.append(axiom_sentence)
print("Extracted %d axiom sentences" % len(axiom_sentences))
URI_Doc = walk_sentences + axiom_sentences
print("\n 1.Extract corpus and learning embedding ... \n")
classes = [line.strip() for line in open(FLAGS.class_file).readlines()]
candidate_num = len(classes)
uri_label = dict()
annotations = list()
for line in open(FLAGS.annotation_file).readlines():
tmp = line.strip().split()
if tmp[1] == 'http://www.w3.org/2000/01/rdf-schema#label':
uri_label[tmp[0]] = pre_process_words(tmp[2:])
elif tmp[0] in classes:
annotations.append(tmp)
walk_sentences, axiom_sentences = list(), list()
if FLAGS.URI_Doc.lower() == 'yes':
walks_ = get_rdf2vec_walks(onto_file=FLAGS.onto_file, walker_type=FLAGS.walker,
walk_depth=FLAGS.walk_depth, classes=classes)
print('Extracted {} walks for {} classes!'.format(len(walks_), len(classes)))
walk_sentences += [list(map(str, x)) for x in walks_]
for line in open(FLAGS.axiom_file).readlines():
axiom_sentence = [item for item in line.strip().split()]
axiom_sentences.append(axiom_sentence)
print('Extracted %d axiom sentences' % len(axiom_sentences))
URI_Doc = walk_sentences + axiom_sentences
Lit_Doc = list()
if FLAGS.Lit_Doc.lower() == 'yes':
for annotation in annotations:
processed_words = pre_process_words(annotation[2:])
if len(processed_words) > 0:
Lit_Doc.append(uri_label[annotation[0]] + processed_words)
print('Extracted %d literal annotations' % len(Lit_Doc))
encoding='utf-8') as f:
for e in projection.entityToPreferredLabels:
for v in projection.entityToPreferredLabels[e]:
f.write('%s preferred_label %s\n' % (e, v))
for a in annotations:
f.write('%s\n' % ' '.join(a))
# read URI document
# two parts: walks, axioms (if the axiom file exists)
walk_sentences, axiom_sentences, URI_Doc = list(), list(), list()
if 'URI_Doc' in config['DOCUMENT'] and config['DOCUMENT']['URI_Doc'] == 'yes':
print('\nGenerate URI document ...')
#walker_type=config['DOCUMENT']['walker']
walks_ = get_rdf2vec_walks(onto_file=ontology_file,
walker_type=config['DOCUMENT']['walker'],
walk_depth=int(
config['DOCUMENT']['walk_depth']),
classes=entities)
print('Extracted %d walks for %d seed entities' %
(len(walks_), len(entities)))
walk_sentences += [list(map(str, x)) for x in walks_]
axiom_file = os.path.join(config['DOCUMENT']['cache_dir'], 'axioms.txt')
if os.path.exists(axiom_file):
for line in open(axiom_file).readlines():
axiom_sentence = [item for item in line.strip().split()]
axiom_sentences.append(axiom_sentence)
print('Extracted %d axiom sentences' % len(axiom_sentences))
URI_Doc = walk_sentences + axiom_sentences