Updates targeting Collection and Upload

[yuanzhou]

Corresponding to hubmapconsortium/entity-api#632, first switch to use the /documents/<id> for retrieving the Collections:

search-api/src/hubmap_translator.py

Lines 1094 to 1101 in a3a1013

	def get_collection(self, entity_id):
	logger.info(f"Start executing get_collection() on uuid: {entity_id}")
	# The entity-api returns public collection with a list of connected public/published datasets, for either
	# - a valid token but not in HuBMAP-Read group or
	# - no token at all
	# Here we do NOT send over the token
	url = self.entity_api_url + "/entities/" + entity_id

Then switch to the new fields when iterating over the datasets associated with a Collection or Upload during the index procedure.

search-api/src/hubmap_translator.py

Lines 704 to 709 in a3a1013

	# Used for Upload and Collection index
	def add_datasets_to_entity(self, entity):
	logger.info("Start executing add_datasets_to_entity()")
	datasets = []
	if 'datasets' in entity:

• For Collection, change to use member_uuids
• For Upload, change to use dataset_uuids

Keep using datasets as the final field that gets indexed otherwise it'll break the Harvard portal-ui:

search-api/src/hubmap_translator.py

Line 739 in a3a1013

entity['datasets'] = datasets

Later, ask the Harvard team and IU (not sure if they are affected) to switch from Collection.datasets parsing to Collection.member_uuids, the same applies to Upload.datasets -> Upload.dataset_uuids.

[yuanzhou]

Move to backlog