def _analyze(self, dataset, task):
entity_ids = set(task.payload.get("entity_ids"))
analyzer = None
for entity in dataset.partials(entity_id=entity_ids):
if analyzer is None or analyzer.entity.id != entity.id:
if analyzer is not None:
entity_ids.update(analyzer.flush())
# log.debug("Analyze: %r", entity)
analyzer = Analyzer(dataset, entity, task.context)
analyzer.feed(entity)
if analyzer is not None:
entity_ids.update(analyzer.flush())
return list(entity_ids)
def analyze(dataset):
db = get_dataset(dataset, OP_ANALYZE)
analyzer = None
for entity in db.partials():
if analyzer is None or analyzer.entity.id != entity.id:
if analyzer is not None:
analyzer.flush()
# log.debug("Analyze: %r", entity)
analyzer = Analyzer(db, entity, {})
analyzer.feed(entity)
if analyzer is not None:
analyzer.flush()
def _analyze(self, dataset, task):
entity_ids = task.payload.get("entity_ids")
analyzer = None
for entity in dataset.partials(entity_id=entity_ids):
if analyzer is None or analyzer.entity.id != entity.id:
if analyzer is not None:
analyzer.flush()
# log.debug("Analyze: %r", entity)
analyzer = Analyzer(dataset, entity)
analyzer.feed(entity)
if analyzer is not None:
analyzer.flush()
return entity_ids
def _analyze(self, task):
entity_ids = task.payload.get('entity_ids')
dataset = Manager.get_dataset(task.stage, task.context)
analyzer = None
for entity in dataset.partials(entity_id=entity_ids):
if analyzer is None or analyzer.entity.id != entity.id:
if analyzer is not None:
analyzer.flush()
# log.debug("Analyze: %r", entity)
analyzer = Analyzer(dataset, entity)
analyzer.feed(entity)
if analyzer is not None:
analyzer.flush()
return entity_ids
def _tagged_entity(self, entity):
a = Analyzer(self.manager.dataset, entity, {})
a.feed(entity)
a.flush()
return self.get_emitted_by_id(entity.id)