def test_ner_extract(self):
text = 'Das ist der Pudel von Angela Merkel. '
text = text * 5
entity = model.make_entity('PlainText')
entity.add('bodyText', text)
tag_entity(entity)
names = entity.get_type_values(registry.name)
assert 'Angela Merkel' in names, names
def test_pattern_extract(self):
text = "Mr. Flubby Flubber called the number tel:+919988111222 twice"
entity = model.make_entity('PlainText')
entity.add('bodyText', text)
tag_entity(entity)
phones = entity.get_type_values(registry.phone)
assert '+919988111222' in phones
countries = entity.get_type_values(registry.country)
assert 'in' in countries
def test_language_tagging(self):
text = "C'est le caniche d'Emmanuel Macron. " * 2
entity = model.make_entity('PlainText')
entity.add('bodyText', text)
tag_entity(entity)
names = entity.get_type_values(registry.name)
assert "d'Emmanuel Macron" in names, names
assert entity.get('detectedLanguage') == ['fra'], entity.get(
'detectedLanguage') # noqa
def _process_entity(entity, sync=False):
"""Perform pre-index processing on an entity, includes running the
NLP pipeline."""
if entity.id is None:
raise InvalidData("No ID for entity", errors=entity.to_dict())
tag_entity(entity)
if sync:
refresh_entity_id(entity.id)
# log.debug("Index: %r", entity)
return entity
def index_entities(stage, collection, iterable, sync=False):
entities = []
for entity in iterable:
if entity.id is None:
raise InvalidData("No ID for entity", errors=entity.to_dict())
tag_entity(entity)
entities.append(entity)
if len(entities) >= BULK_PAGE:
stage.report_finished(len(entities))
index_bulk(collection, entities, job_id=stage.job.id, sync=sync)
entities = []
if len(entities):
stage.report_finished(len(entities))
index_bulk(collection, entities, job_id=stage.job.id, sync=sync)
refresh_collection(collection)
def index_entities(collection, iterable, sync=False):
queue = get_queue(collection, OP_INDEX)
queue.progress.mark_pending(len(iterable))
entities = []
for entity in iterable:
if entity.id is None:
raise InvalidData("No ID for entity", errors=entity.to_dict())
tag_entity(entity)
entities.append(entity)
if len(entities) >= BULK_PAGE:
queue.progress.mark_finished(len(entities))
index_bulk(collection, entities, sync=sync)
entities = []
if len(entities):
queue.progress.mark_finished(len(entities))
index_bulk(collection, entities, sync=sync)
refresh_collection(collection)
