Skip to content

Track down the NodeNorm Biolink type changes on exp #108

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
gaurav wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Track down the NodeNorm Biolink type changes on exp #108

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
65cc441
Report all Biolink type differences at once in test_nodenorm_types
gaurav Jul 24, 2026
8cb809c
Update expected Biolink types to the NodeNorm `exp` baseline
gaurav Jul 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
189 changes: 105 additions & 84 deletions tests/nodenorm/test_nodenorm_types.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,16 @@
# is correct. These tests are intended to cover whether all the Biolink types are correct, for both normal entries
# and conflated entries.
#
# The expected values below are the baseline as of NodeNorm on `exp` (2026-07-24). They differ from
# dev/prod/test/ci, which still return the previous baseline; see the PR discussion for whether these
# changes should be accepted or fixed in NodeNorm. The three differences are:
# 1. `biolink:OntologyClass` is now included for cliques that previously lacked it (MESH:D014867,
# NCIT:C34373), inserted after the second-most-specific type.
# 2. `biolink:GeneOrGeneProductOrGeneFamily` is a new type on gene cliques, after
# `biolink:GeneOrGeneProduct`.
# 3. Because of (1), `biolink:OntologyClass` now appears in the main clique's types for
# MESH:D014867 rather than being appended by DrugChemical conflation, so its position moves.
#

import pytest
import requests
Expand All @@ -12,6 +22,55 @@
def nodenorm_url(target_info):
return target_info['NodeNormURL']


def check_biolink_types(nodenorm_url, expected, conflate, drug_chemical_conflate, allowed_individual_types):
"""
Query get_normalized_nodes for every CURIE in `expected` and compare the returned Biolink types.

Every mismatch across every CURIE is collected and reported in a single assertion, so that one
bad clique doesn't hide the others. `allowed_individual_types` is called with the types actually
returned for the clique and should return the set of types each equivalent identifier may have.
"""
response = requests.post(nodenorm_url + "get_normalized_nodes", json={
"curies": list(expected.keys()),
"conflate": conflate,
"drug_chemical_conflate": drug_chemical_conflate,
"individual_types": True,
})
response.raise_for_status()
results = response.json()

errors = []
for curie, expected_types in expected.items():
result = results.get(curie)
if result is None:
errors.append(f"{curie}: no result returned")
continue

actual_types = result['type']
if actual_types != expected_types:
missing = [t for t in expected_types if t not in actual_types]
extra = [t for t in actual_types if t not in expected_types]
reordered = not missing and not extra
errors.append(
f"{curie}: Biolink types differ ({'reordered only' if reordered else 'membership differs'})\n"
f" expected: {expected_types}\n"
f" actual: {actual_types}\n"
f" missing from actual: {missing}\n"
f" unexpected in actual: {extra}"
)

allowed = allowed_individual_types(actual_types)
for eq in result['equivalent_identifiers']:
if eq['type'] not in allowed:
errors.append(
f"{curie}: equivalent identifier {eq['identifier']} has type {eq['type']}, "
f"expected one of {sorted(allowed)}"
)

assert not errors, f"{len(errors)} difference(s) found:\n\n" + "\n\n".join(errors)


def test_unconflated_biolink_types(nodenorm_url):
"""
Check a few Biolink types without conflation turned on.
Expand All @@ -20,6 +79,7 @@ def test_unconflated_biolink_types(nodenorm_url):
"MESH:D014867": [
"biolink:SmallMolecule",
"biolink:MolecularEntity",
"biolink:OntologyClass",
"biolink:ChemicalEntity",
"biolink:PhysicalEssence",
"biolink:ChemicalOrDrugOrTreatment",
Expand All @@ -31,13 +91,15 @@ def test_unconflated_biolink_types(nodenorm_url):
"NCIT:C34373": [
"biolink:Disease",
"biolink:DiseaseOrPhenotypicFeature",
"biolink:OntologyClass",
"biolink:BiologicalEntity",
"biolink:ThingWithTaxon",
"biolink:NamedThing"
],
"NCBIGene:22059": [
'biolink:Gene',
'biolink:GeneOrGeneProduct',
'biolink:GeneOrGeneProductOrGeneFamily',
'biolink:GenomicEntity',
'biolink:ChemicalEntityOrGeneOrGeneProduct',
'biolink:PhysicalEssence',
Expand Down Expand Up @@ -81,26 +143,14 @@ def test_unconflated_biolink_types(nodenorm_url):
]
}

response = requests.post(nodenorm_url + "get_normalized_nodes", json={
"curies": list(curies_and_expected_results.keys()),
"conflate": False,
"drug_chemical_conflate": False,
"individual_types": True,
})
results = response.json()

# Pull out each Biolink types list and compare to expected results.
for query_curie in curies_and_expected_results.keys():
result = results[query_curie]

# Compare Biolink types with expected results.
biolink_types = result['type']
assert biolink_types == curies_and_expected_results[query_curie]

# Check the individual types, which -- without conflation -- should be identical to the first Biolink type.
first_biolink_type = biolink_types[0]
for equivalent_identifiers in result['equivalent_identifiers']:
assert equivalent_identifiers['type'] == first_biolink_type
# Without conflation, every individual type should be identical to the first Biolink type.
check_biolink_types(
nodenorm_url,
curies_and_expected_results,
conflate=False,
drug_chemical_conflate=False,
allowed_individual_types=lambda actual_types: {actual_types[0]},
)


def test_geneprotein_conflated_biolink_types(nodenorm_url):
Expand All @@ -111,6 +161,7 @@ def test_geneprotein_conflated_biolink_types(nodenorm_url):
"NCBIGene:22059": [
'biolink:Gene',
'biolink:GeneOrGeneProduct',
'biolink:GeneOrGeneProductOrGeneFamily',
'biolink:GenomicEntity',
'biolink:ChemicalEntityOrGeneOrGeneProduct',
'biolink:PhysicalEssence',
Expand All @@ -128,6 +179,7 @@ def test_geneprotein_conflated_biolink_types(nodenorm_url):
"PR:Q9Y6J0": [
'biolink:Gene',
'biolink:GeneOrGeneProduct',
'biolink:GeneOrGeneProductOrGeneFamily',
'biolink:GenomicEntity',
'biolink:ChemicalEntityOrGeneOrGeneProduct',
'biolink:PhysicalEssence',
Expand All @@ -144,25 +196,14 @@ def test_geneprotein_conflated_biolink_types(nodenorm_url):
]
}

response = requests.post(nodenorm_url + "get_normalized_nodes", json={
"curies": list(curies_and_expected_results.keys()),
"conflate": True,
"drug_chemical_conflate": False,
"individual_types": True,
})
results = response.json()

# Pull out each Biolink types list and compare to expected results.
for query_curie in curies_and_expected_results.keys():
result = results[query_curie]

# Compare Biolink types with expected results.
biolink_types = result['type']
assert biolink_types == curies_and_expected_results[query_curie]

# Check the individual types, which -- with GeneProtein conflation -- should be either 'biolink:Gene' or 'biolink:Protein'.
for equivalent_identifiers in result['equivalent_identifiers']:
assert equivalent_identifiers['type'] in {'biolink:Gene', 'biolink:Protein'}
# With GeneProtein conflation, every individual type should be either 'biolink:Gene' or 'biolink:Protein'.
check_biolink_types(
nodenorm_url,
curies_and_expected_results,
conflate=True,
drug_chemical_conflate=False,
allowed_individual_types=lambda actual_types: {'biolink:Gene', 'biolink:Protein'},
)


def test_drugchemical_conflated_biolink_types(nodenorm_url):
Expand All @@ -173,6 +214,7 @@ def test_drugchemical_conflated_biolink_types(nodenorm_url):
"MESH:D014867": [
"biolink:SmallMolecule",
"biolink:MolecularEntity",
"biolink:OntologyClass",
"biolink:ChemicalEntity",
"biolink:PhysicalEssence",
"biolink:ChemicalOrDrugOrTreatment",
Expand All @@ -181,7 +223,6 @@ def test_drugchemical_conflated_biolink_types(nodenorm_url):
"biolink:NamedThing",
"biolink:PhysicalEssenceOrOccurrent",
'biolink:Drug',
'biolink:OntologyClass',
'biolink:MolecularMixture',
'biolink:ChemicalMixture',
],
Expand All @@ -200,30 +241,18 @@ def test_drugchemical_conflated_biolink_types(nodenorm_url):
]
}

response = requests.post(nodenorm_url + "get_normalized_nodes", json={
"curies": list(curies_and_expected_results.keys()),
"conflate": False,
"drug_chemical_conflate": True,
"individual_types": True,
})
results = response.json()

# Pull out each Biolink types list and compare to expected results.
for query_curie in curies_and_expected_results.keys():
result = results[query_curie]

# Compare Biolink types with expected results.
biolink_types = result['type']
assert biolink_types == curies_and_expected_results[query_curie]

# Check the individual types, which -- without conflation -- should be one of the chemical types.
first_biolink_type = biolink_types[0]
for equivalent_identifiers in result['equivalent_identifiers']:
assert equivalent_identifiers['type'] in {
'biolink:Drug',
'biolink:SmallMolecule',
'biolink:ChemicalEntity',
}
# With DrugChemical conflation, every individual type should be one of the chemical types.
check_biolink_types(
nodenorm_url,
curies_and_expected_results,
conflate=False,
drug_chemical_conflate=True,
allowed_individual_types=lambda actual_types: {
'biolink:Drug',
'biolink:SmallMolecule',
'biolink:ChemicalEntity',
},
)



Expand All @@ -235,6 +264,7 @@ def test_fully_conflated_biolink_types(nodenorm_url):
"MESH:D014867": [
"biolink:SmallMolecule",
"biolink:MolecularEntity",
"biolink:OntologyClass",
"biolink:ChemicalEntity",
"biolink:PhysicalEssence",
"biolink:ChemicalOrDrugOrTreatment",
Expand All @@ -243,20 +273,21 @@ def test_fully_conflated_biolink_types(nodenorm_url):
"biolink:NamedThing",
"biolink:PhysicalEssenceOrOccurrent",
'biolink:Drug',
'biolink:OntologyClass',
'biolink:MolecularMixture',
'biolink:ChemicalMixture',
],
"NCIT:C34373": [
"biolink:Disease",
"biolink:DiseaseOrPhenotypicFeature",
"biolink:OntologyClass",
"biolink:BiologicalEntity",
"biolink:ThingWithTaxon",
"biolink:NamedThing"
],
"NCBIGene:22059": [
'biolink:Gene',
'biolink:GeneOrGeneProduct',
'biolink:GeneOrGeneProductOrGeneFamily',
'biolink:GenomicEntity',
'biolink:ChemicalEntityOrGeneOrGeneProduct',
'biolink:PhysicalEssence',
Expand Down Expand Up @@ -287,6 +318,7 @@ def test_fully_conflated_biolink_types(nodenorm_url):
"PR:Q9Y6J0": [
'biolink:Gene',
'biolink:GeneOrGeneProduct',
'biolink:GeneOrGeneProductOrGeneFamily',
'biolink:GenomicEntity',
'biolink:ChemicalEntityOrGeneOrGeneProduct',
'biolink:PhysicalEssence',
Expand All @@ -313,22 +345,11 @@ def test_fully_conflated_biolink_types(nodenorm_url):
]
}

response = requests.post(nodenorm_url + "get_normalized_nodes", json={
"curies": list(curies_and_expected_results.keys()),
"conflate": True,
"drug_chemical_conflate": True,
"individual_types": True,
})
results = response.json()

# Pull out each Biolink types list and compare to expected results.
for query_curie in curies_and_expected_results.keys():
result = results[query_curie]

# Compare Biolink types with expected results.
biolink_types = result['type']
assert biolink_types == curies_and_expected_results[query_curie]

# Check the individual types, which -- with full conflation -- should be one of the Biolink types for the overall clique.
for equivalent_identifiers in result['equivalent_identifiers']:
assert equivalent_identifiers['type'] in biolink_types
# With full conflation, every individual type should be one of the Biolink types for the overall clique.
check_biolink_types(
nodenorm_url,
curies_and_expected_results,
conflate=True,
drug_chemical_conflate=True,
allowed_individual_types=set,
)