BaseIDProvider -- this class should not provide the connection of opinions and samples.

[nicolay-r]

Problem: the predefined template of the identifiers does not allow us to consider framework for other cases with different data structuring.

How to refactor: such meta-information should go alongside with the other data instead of being sealed into the specific format of the identifier; for example such columns as relation_index and doc_relation_id could be added.

• ❌ remove parse_label_in_sample_id
• ❌ remove parse_index_in_sample_id
• 🔧 the code below is depend on the id format, while it would be better do depend on the specific columns (mentioned above)

AREkit/arekit/contrib/utils/data/views/linkages/base.py

Lines 45 to 47 in a624ae1

	doc_df = self._storage.find_by_value(column_name=const.DOC_ID, value=doc_id)
	row_ids = [row_id for row_id in doc_df[const.ID]] # TODO. Adopt storage.
	doc_opin_ids = self.__iter_doc_opinion_ids(row_ids=row_ids)

• 🔧 the same is for sample view; in the snipped below we depend on the column ID

AREkit/arekit/common/data/views/samples.py

Lines 18 to 20 in a624ae1

	for row_index, sample_id in enumerate(storage.iter_column_values(const.ID)):
	sample_id = str(sample_id)
	opinion_id = self.__row_ids_provider.parse_opinion_in_sample_id(sample_id)

• ❌ _parse method
• ❌ convert_sample_id_to_opinion_id mehod
• BaseOpinionLinkagesView #492
• Remove TextOpinionLinkagesToOpinionConverterPipelineItem class #491
• RowIdProvider -- create_sample_id call with the related class are expected to be refactored #493