DocumentEntityID Assignation -- code parts are implicitly sinchronized with implementation

[nicolay-r]

How to fix: So in a snippet №2 we can use the entities mapping.

The reason is because the code below executed before:

AREkit/arekit/common/news/parsed/providers/base.py

Lines 22 to 39 in 1e1d354

	def init_parsed_news(self, parsed_news):
	assert(isinstance(parsed_news, ParsedNews))
	self._doc_entities = []
	self.__entity_map.clear()
	for index, entity in enumerate(parsed_news.iter_entities()):
	doc_entity = DocumentEntity(id_in_doc=index,
	value=entity.Value,
	e_type=entity.Type,
	display_value=entity.DisplayValue,
	group_index=entity.GroupIndex)
	self._doc_entities.append(doc_entity)
	if self.__entity_index_func is not None:
	self.__entity_map[self.__entity_index_func(entity)] = doc_entity

and then followed by this function in the inherited class:

AREkit/arekit/common/news/parsed/providers/entity_service.py

Lines 152 to 168 in 1e1d354

	def __calculate_entity_positions(self):
	""" Note: here we consider the same order as in self._entities.
	"""
	positions = []
	t_ind_in_doc = 0
	for s_ind, t_ind_in_sent, term in self.__iter_raw_terms_func():
	if isinstance(term, Entity):
	position = TermPosition(term_ind_in_doc=t_ind_in_doc,
	term_ind_in_sent=t_ind_in_sent,
	s_ind=s_ind)
	positions.append(position)
	t_ind_in_doc += 1
	return positions

Both are about assigning identifiers.
However, when it comes to nested entities, this is expected to be sinchronized in a single place

[nicolay-r]

Fixed in 8c1ee11