def demonstrate_source_lookup(source_uri):
print 'Demonstrating LookUp source API'
lookup = LookUp()
data = lookup.search_source(source_uri)
r = Result(data)
r.print_raw_result()
print
def get_all_tuples_of_relations(self):
'''
Returns list of tuples of relations which connects 'self.concepts' sequence in the path.
'''
relations_tuples = []
for index, (concept1, concept2) in enumerate(pairwise(self.concepts)):
search = Search(start=concept1, end=concept2)
data = search.search()
result = Result(data)
if result.get_num_found() > 0:
edges = result.parse_all_edges()
if index == 0:
for e in edges:
if not [e.rel] in relations_tuples:
relations_tuples.append([e.rel])
else:
for relation_tuple in relations_tuples:
if len(relation_tuple) == index:
for e in edges:
relation_tuple_copy = copy.deepcopy(
relation_tuple)
relation_tuple_copy.append(e.rel)
if not relation_tuple_copy in relations_tuples:
relations_tuples.append(
relation_tuple_copy)
for relation_tuple in relations_tuples:
if len(relation_tuple) != len(self.concepts) - 1:
relations_tuples.remove(relation_tuple)
return relations_tuples
def get_all_tuples_of_relations(self):
'''
Returns list of tuples of relations which connects 'self.concepts' sequence in the path.
'''
relations_tuples = []
for index, (concept1, concept2) in enumerate(pairwise(self.concepts)):
search = Search(start=concept1, end=concept2)
data = search.search()
result = Result(data)
if result.get_num_found() > 0:
edges = result.parse_all_edges()
if index == 0:
for e in edges:
if not [e.rel] in relations_tuples:
relations_tuples.append([e.rel])
else:
for relation_tuple in relations_tuples:
if len(relation_tuple) == index:
for e in edges:
relation_tuple_copy = copy.deepcopy(relation_tuple)
relation_tuple_copy.append(e.rel)
if not relation_tuple_copy in relations_tuples:
relations_tuples.append(relation_tuple_copy)
for relation_tuple in relations_tuples:
if len(relation_tuple) != len(self.concepts)-1:
relations_tuples.remove(relation_tuple)
return relations_tuples
def demonstrate_source_lookup(source_uri):
print 'Demonstrating LookUp source API'
lookup = LookUp()
data = lookup.search_source(source_uri)
r = Result(data)
r.print_raw_result()
print
def printRelations(self, term):
lookup = LookUp(limit=50)
response = lookup.search_concept(term)
response = Result(response)
edges = response.parse_all_edges(clean_self_ref = True)
for edge in edges:
print("%s --> %s --> %s" % (edge.start, edge.rel, edge.end))
def countRelations(self, term):
lookup = LookUp(limit=50)
response = lookup.search_concept(term)
response = Result(response)
edges = response.parse_all_edges(clean_self_ref = True)
i = 0
for edge in edges:
i+=1
return i
def makeConceptMap(self, term, modifier):
lookup = LookUp(limit=500)
response = lookup.search_concept(term)
response = Result(response)
edges = response.parse_all_edges(clean_self_ref = True)
conceptMap = {}
for edge in edges:
if edge.rel == self.getRelationUri(modifier):
conceptMap[edge.start] = edge.end
return conceptMap
def get_all_tuples_of_concepts(self):
'''
Returns list of tuples of concepts connected by 'self.relations' sequence in the path.
'''
graph = nx.DiGraph()
for relation in self.relations:
search = Search(rel=relation, limit=1000)
data = search.search()
result = Result(data)
edges = result.parse_all_edges()
for e in edges:
graph.add_node(e.start)
graph.add_node(e.end)
graph.add_edge(e.start, e.end)
graph[e.start][e.end]['relation'] = relation
concepts_tuples = []
for index, r in enumerate(self.relations):
for edge in graph.edges(data=True):
start = edge[0]
end = edge[1]
relation = edge[2]['relation']
if relation == r:
if index == 0:
concept_tuple = [start, end]
concepts_tuples.append(concept_tuple)
else:
for concept_tuple in concepts_tuples:
if len(concept_tuple) == index + 1:
concepts_tuples.remove(concept_tuple)
next_candidates = []
for next_edge in graph.edges(data=True):
next_edge_start = next_edge[0]
next_edge_end = next_edge[1]
next_edge_relation = next_edge[2][
'relation']
if next_edge_relation == r and concept_tuple[
-1] == next_edge_start:
next_candidates.append(next_edge_end)
for candidate in next_candidates:
concept_tuple_copy = copy.deepcopy(
concept_tuple)
concept_tuple_copy.append(candidate)
concepts_tuples.append(concept_tuple_copy)
return concepts_tuples
def does_exist(self, print_where_breaks=False):
'''
Checks whether this path exists in ConceptNet5.
:param print_where_breaks: prints the assertion, if the assertion does not exist in the path.
'''
for assertion in self.assertions:
search = Search(start=assertion.start, rel=assertion.relation, end=assertion.end, limit=1000)
data = search.search()
result = Result(data)
if result.get_num_found() == 0:
if print_where_breaks == True:
print('Assertion breaking the path: [ %s --> (%s) --> %s) ]' % (
assertion.start, assertion.relation, assertion.end))
return False
return True
def demonstrate_association():
print 'Demonstrating Association API'
a = Association(filter='/c/en/dog', limit=1)
data = a.get_similar_concepts('cat')
r = Result(data)
r.print_raw_result()
print
a = Association()
data = a.get_similar_concepts_by_term_list(
['toast', 'cereal', 'juice', 'egg'])
r = Result(data)
r.print_raw_result()
print
r.parse_all_edges()
print
def demonstrate_lookup(concept):
print 'Demonstrating LookUp concept API'
lookup = LookUp(offset=1, limit=1)
data = lookup.search_concept(concept)
r = Result(data)
edges = r.parse_all_edges()
for edge in edges:
edge.print_edge()
edge.print_all_attrs()
print
print 'Demonstrating LookUp concept API cleaning self referencing edges'
lookup = LookUp(offset=1, limit=1)
data = lookup.search_concept(concept)
r = Result(data)
r.print_raw_result()
print
def get_all_tuples_of_concepts(self):
'''
Returns list of tuples of concepts connected by 'self.relations' sequence in the path.
'''
graph = nx.DiGraph()
for relation in self.relations:
search = Search(rel=relation, limit=1000)
data = search.search()
result = Result(data)
edges = result.parse_all_edges()
for e in edges:
graph.add_node(e.start)
graph.add_node(e.end)
graph.add_edge(e.start, e.end)
graph[e.start][e.end]['relation'] = relation
concepts_tuples = []
for index, r in enumerate(self.relations):
for edge in graph.edges(data=True):
start = edge[0]
end = edge[1]
relation = edge[2]['relation']
if relation == r:
if index == 0:
concept_tuple = [start, end]
concepts_tuples.append(concept_tuple)
else:
for concept_tuple in concepts_tuples:
if len(concept_tuple) == index+1:
concepts_tuples.remove(concept_tuple)
next_candidates = []
for next_edge in graph.edges(data=True):
next_edge_start = next_edge[0]
next_edge_end = next_edge[1]
next_edge_relation = next_edge[2]['relation']
if next_edge_relation == r and concept_tuple[-1] == next_edge_start:
next_candidates.append(next_edge_end)
for candidate in next_candidates:
concept_tuple_copy = copy.deepcopy(concept_tuple)
concept_tuple_copy.append(candidate)
concepts_tuples.append(concept_tuple_copy)
return concepts_tuples
def does_exist(self, print_where_breaks=False):
'''
Checks whether this path exists in ConceptNet5.
:param print_where_breaks: prints the assertion, if the assertion does not exist in the path.
'''
for assertion in self.assertions:
search = Search(start=assertion.start,
rel=assertion.relation,
end=assertion.end,
limit=1000)
data = search.search()
result = Result(data)
if result.get_num_found() == 0:
if print_where_breaks == True:
print 'Assertion breaking the path: [ %s --> (%s) --> %s) ]' % (
assertion.start, assertion.relation, assertion.end)
return False
return True
def demonstrate_search():
print 'Demonstrating Search API'
s = Search(rel='/c/en/be_often_compare_to')
data = s.search()
r = Result(data)
r.print_raw_result()
print
s = Search(text='mariah carey', surfaceText='dion', something='anything')
data = s.search()
r = Result(data)
r.print_raw_result()
print
def demonstrate_association():
print 'Demonstrating Association API'
a = Association(filter='/c/en/dog', limit=1)
data = a.get_similar_concepts('cat')
r = Result(data)
r.print_raw_result()
print
a = Association()
data = a.get_similar_concepts_by_term_list(['toast', 'cereal', 'juice', 'egg'])
r = Result(data)
r.print_raw_result()
print
r.parse_all_edges()
print
def demonstrate_search():
print 'Demonstrating Search API'
s = Search(rel='/c/en/be_often_compare_to')
data = s.search()
r = Result(data)
r.print_raw_result()
print
s = Search(text='mariah carey', surfaceText='dion', something='anything')
data = s.search()
r = Result(data)
r.print_raw_result()
print
def demonstrate_lookup(concept):
print 'Demonstrating LookUp concept API'
lookup = LookUp(offset=1, limit=1)
data = lookup.search_concept(concept)
r = Result(data)
edges = r.parse_all_edges()
for edge in edges:
edge.print_edge()
edge.print_all_attrs()
print
print 'Demonstrating LookUp concept API cleaning self referencing edges'
lookup = LookUp(offset=1, limit=1)
data = lookup.search_concept(concept)
r = Result(data)
r.print_raw_result()
print
from conceptnet5_client.web.api import Search
from conceptnet5_client.web.api import Association
from conceptnet5_client.utils.result import Result
# get how similar cats and dogs
a = Association(filter='/c/en/dog', limit=1)
data = a.get_similar_concepts('cat')
r = Result(data)
# print results in key = value format
r.print_raw_result()
a = Association()
data = a.get_similar_concepts_by_term_list(['toast', 'cereal', 'juice', 'egg'])
r = Result(data)
# print results in key = value format
r.print_raw_result()
ques = prenlp.preprocess(q.question)
ans_a = prenlp.preprocess(q.a)
ans_b = prenlp.preprocess(q.b)
ans_c = prenlp.preprocess(q.c)
ans_d = prenlp.preprocess(q.d)
print "Question: " + str(q.question)
print "A: " + str(q.a)
print "B: " + str(q.b)
print "C: " + str(q.c)
print "D: " + str(q.d)
# Generate Semantic Graph from Question
a = Association(filter="/c/en", limit=30)
semnet = a.get_similar_concepts_by_term_list(ques)
r = Result(semnet)
# Parse Similarity
similar = r.get_similar()
if len(similar) > 0:
# Splice Leading API Directory
for word in similar:
word[0] = word[0][6:]
print "\n"
print similar
print "\n"
# Compute Score Probabilities
prob = [str(computeScore(similar, ans_a)),
ans_a = prenlp.preprocess(q.a) ans_b = prenlp.preprocess(q.b) ans_c = prenlp.preprocess(q.c) ans_d = prenlp.preprocess(q.d) print "[Question]: " + str(ques) + "\n" print "[Answer Choices]" print "A: " + str(ans_a) print "B: " + str(ans_b) print "C: " + str(ans_c) print "D: " + str(ans_d) print print "[Correct Answer]: " + str(q.ans) + "\n" # Perform Semantic Association Search on Question a = Association() data = a.get_similar_concepts_by_term_list(ques) r = Result(data) # Obtain Associtative Edges r.parse_all_edges() print # Perform an Exhaustive Entity Frequency Search # Check to see which one contains the most amount of relevant entities. #print "[Predicted Answer]: " #print "[Correct Answer]: " + str(q.ans) + "\n"
ques = prenlp.preprocess(q.question) ans_a = prenlp.preprocess(q.a) ans_b = prenlp.preprocess(q.b) ans_c = prenlp.preprocess(q.c) ans_d = prenlp.preprocess(q.d) print "[Question]: " + str(ques) + "\n" print "[Answer Choices]" print "A: " + str(ans_a) print "B: " + str(ans_b) print "C: " + str(ans_c) print "D: " + str(ans_d) print print "[Correct Answer]: " + str(q.ans) + "\n" # Perform Semantic Association Search on Question a = Association() data = a.get_similar_concepts_by_term_list(ques) r = Result(data) # Obtain Associtative Edges r.parse_all_edges() print # Perform an Exhaustive Entity Frequency Search # Check to see which one contains the most amount of relevant entities. #print "[Predicted Answer]: " #print "[Correct Answer]: " + str(q.ans) + "\n"
ques = prenlp.preprocess(q.question)
ans_a = prenlp.preprocess(q.a)
ans_b = prenlp.preprocess(q.b)
ans_c = prenlp.preprocess(q.c)
ans_d = prenlp.preprocess(q.d)
print "Question: " + str(q.question)
print "A: " + str(q.a)
print "B: " + str(q.b)
print "C: " + str(q.c)
print "D: " + str(q.d)
# Generate Semantic Graph from Question
a = Association(filter="/c/en", limit=30)
semnet = a.get_similar_concepts_by_term_list(ques)
r = Result(semnet)
# Parse Similarity
similar = r.get_similar()
if len(similar) > 0:
# Splice Leading API Directory
for word in similar:
word[0] = word[0][6:]
print "\n"
print similar
print "\n"
# Compute Score Probabilities
prob = [
