Two transformation functions do not return the transformed dictionary

[eecavanna]

Summary

While running a transformation function named remove_doi_slots from within a Python notebook that imports the containing Python class from the nmdc-schema Python package on PyPI, Python raised an AttributeError exception because a value of None doesn't have an attribute named get.

Screenshot

Related snippets

This is where the exception was raised:

nmdc-schema/nmdc_schema/migration_recursion.py

Lines 382 to 384 in 918056b

	# Get dois from associated_dois
	associated_dois = [entry['doi_value']
	for entry in study.get('associated_dois', [])]

These two transformation functions don't return the transformed dictionary (i.e. they implicitly return None since they don't have return statements):

nmdc-schema/nmdc_schema/migration_recursion.py

Lines 316 to 327 in 918056b

	def process_doi(self, study: dict, doi_list: list, doi_category: str):
	id_value = study['id']
	for doi_updates in doi_list:
	if id_value == doi_updates['id']:
	new_doi = {
	'doi_value': doi_updates['doi'],
	'doi_category': doi_category,
	'doi_provider': doi_updates['doi_prov']
	}
	study.setdefault('associated_dois', []).append(new_doi)
	def fix_award_dois(self, study: dict):

nmdc-schema/nmdc_schema/migration_recursion.py

Lines 368 to 376 in 918056b

	def fix_ess_dive(self, study: dict):
	"""Move ess_dive_datasets values to associated_dois slot"""
	study.setdefault('associated_dois', []).extend(
	{'doi_value': dataset_doi, 'doi_category': 'dataset_doi',
	'doi_provider': 'ess_dive'}
	for dataset_doi in study.get('ess_dive_datasets', []))
	def remove_doi_slots(self, study: dict):

As a result, transformation functions "further down the pipeline" won't receive a dictionary as input.

Proposal

End each transformation function with:

return study

[eecavanna]

As a short-term workaround, in the notebook, I ignored the return values after confirming (by reviewing their source code) all the transformation functions modify the dictionary "in place."

I created the following PR which contains a long-term fix: #1312

[turbomam]

Thanks @eecavanna

@brynnz22 would be it be OK for me to remove myself as an assignee and just leave you?

[brynnz22]

@turbomam Yes! I just approved the PR. I think you already changed it? Or @eecavanna did. Thanks for the quick fix. Sorry about this.