def _search_via_relational_roles(self, distance, provisional_roles,
relations):
for entity_key in sorted(self.entity_keys_to_visit):
node = self.nodes.get(entity_key)
if not node:
continue
entity = node.entity
relational_count = entity.roles_to_relation_count(
provisional_roles)
if 0 < distance and self.max_links < relational_count:
self.entity_keys_to_visit.remove(entity_key)
message = ' Pre-pruned {} [{}]'
message = message.format(entity.name, relational_count)
print(message)
if provisional_roles and distance < self.degree:
print(' Retrieving relational relations')
keys = sorted(self.entity_keys_to_visit)
step = 500
stop = len(keys)
for start in range(0, stop, step):
key_slice = keys[start:start + step]
print(' {}-{} of {}'.format(
start + 1,
min(start + step, stop),
stop,
))
relations.update(
PostgresRelation.search_multi(
key_slice,
roles=provisional_roles,
))
def _search_via_relational_roles(self, distance, provisional_roles, relations):
for entity_key in sorted(self.entity_keys_to_visit):
node = self.nodes.get(entity_key)
if not node:
continue
entity = node.entity
relational_count = entity.roles_to_relation_count(provisional_roles)
if 0 < distance and self.max_links < relational_count:
self.entity_keys_to_visit.remove(entity_key)
message = ' Pre-pruned {} [{}]'
message = message.format(entity.name, relational_count)
print(message)
if provisional_roles and distance < self.degree:
print(' Retrieving relational relations')
keys = sorted(self.entity_keys_to_visit)
step = 500
stop = len(keys)
for start in range(0, stop, step):
key_slice = keys[start:start + step]
print(' {}-{} of {}'.format(
start + 1,
min(start + step, stop),
stop,
))
relations.update(
PostgresRelation.search_multi(
key_slice,
roles=provisional_roles,
)
)
def _cross_reference(self, distance):
# TODO: We don't need to test all nodes, only those missing credit role
# relations. That may significantly reduce the computational
# load.
if not self.relational_roles:
print(' Skipping cross-referencing: no relational roles')
return
elif distance < 2:
print(
' Skipping cross-referencing: maximum distance less than 2')
return
else:
print(' Cross-referencing...')
relations = {}
entity_keys = sorted(self.nodes)
entity_keys.remove(self.center_entity.entity_key)
entity_key_slices = []
step = 250
for start in range(0, len(entity_keys), step):
entity_key_slices.append(entity_keys[start:start + step])
iterator = itertools.product(entity_key_slices, entity_key_slices)
for lh_entities, rh_entities in iterator:
print(' {} & {}'.format(len(lh_entities), len(rh_entities)))
found = PostgresRelation.search_bimulti(
lh_entities,
rh_entities,
roles=self.relational_roles,
)
relations.update(found)
self._process_relations(relations)
message = ' Cross-referenced: {} nodes / {} links'
message = message.format(len(self.nodes), len(self.links))
print(message)
def _cross_reference(self, distance):
# TODO: We don't need to test all nodes, only those missing credit role
# relations. That may significantly reduce the computational
# load.
if not self.relational_roles:
print(' Skipping cross-referencing: no relational roles')
return
elif distance < 2:
print(' Skipping cross-referencing: maximum distance less than 2')
return
else:
print(' Cross-referencing...')
relations = {}
entity_keys = sorted(self.nodes)
entity_keys.remove(self.center_entity.entity_key)
entity_key_slices = []
step = 250
for start in range(0, len(entity_keys), step):
entity_key_slices.append(entity_keys[start:start + step])
iterator = itertools.product(entity_key_slices, entity_key_slices)
for lh_entities, rh_entities in iterator:
print(' {} & {}'.format(len(lh_entities), len(rh_entities)))
found = PostgresRelation.search_bimulti(
lh_entities,
rh_entities,
roles=self.relational_roles,
)
relations.update(found)
self._process_relations(relations)
message = ' Cross-referenced: {} nodes / {} links'
message = message.format(len(self.nodes), len(self.links))
print(message)
