Merge pull request #620 from iQuxLE/bgee_ingest_split_odd_intersectio…

…n_symbol

splitting apart intersection-symbol IDs in bgee ingest

src/monarch_ingest/ingests/bgee/gene_to_expression_utils.py

@@ -36,18 +36,20 @@ def write_group(rows: List, koza_app: KozaApp):
         koza_app (KozaApp): The KozaApp to use for output of rows.
     """
     for row in rows:
-        association = GeneToExpressionSiteAssociation(
-            id="uuid:" + str(uuid.uuid1()),
-            subject="ENSEMBL:" + row['Gene ID'],
-            predicate='biolink:expressed_in',
-            object=row['Anatomical entity ID'],
-            primary_knowledge_source="infores:bgee",
-            aggregator_knowledge_source=["infores:monarchinitiative"],
-            knowledge_level=KnowledgeLevelEnum.knowledge_assertion,
-            agent_type=AgentTypeEnum.not_provided,
-        )
-
-        koza_app.write(association)
+        anatomical_entities = row['Anatomical entity ID'].split(' ∩ ')
+        for anatomical_entity in anatomical_entities:
+            association = GeneToExpressionSiteAssociation(
+                id="uuid:" + str(uuid.uuid1()),
+                subject="ENSEMBL:" + row['Gene ID'],
+                predicate='biolink:expressed_in',
+                object=anatomical_entity.strip(),
+                primary_knowledge_source="infores:bgee",
+                aggregator_knowledge_source=["infores:monarchinitiative"],
+                knowledge_level=KnowledgeLevelEnum.knowledge_assertion,
+                agent_type=AgentTypeEnum.not_provided,
+            )
+
+            koza_app.write(association)
 
 
 def get_row_group(koza_app: KozaApp, col: str = 'Gene ID') -> Union[List, None]:
@@ -68,12 +70,25 @@ def get_row_group(koza_app: KozaApp, col: str = 'Gene ID') -> Union[List, None]:
     elif koza_app.previous_row is None:
         return None
 
-    rows = [koza_app.previous_row]
-    current_row = koza_app.get_row()
-
-    while rows[0][col] == current_row[col]:
-        rows.append(current_row)
-        current_row = koza_app.get_row()
+    rows = []
+    current_row = koza_app.previous_row
+
+    while current_row[col] == koza_app.previous_row[col]:
+        if " ∩ " in current_row['Anatomical entity ID']:
+            multiple_entities = [
+                entity.strip().replace('"','') for entity in current_row['Anatomical entity ID'].split(' ∩ ')
+            ]
+            for entity in multiple_entities:
+                split_row = current_row.copy()
+                split_row['Anatomical entity ID'] = entity.strip()
+                rows.append(split_row)
+        else:
+            rows.append(current_row)
+
+        try:
+            current_row = koza_app.get_row()
+        except StopIteration:
+            break
 
     koza_app.previous_row = current_row
     return rows

tests/unit/bgee/test_bgee_gene_to_expression.py

@@ -67,7 +67,15 @@ def get_koza_rows(mock_koza: KozaApp, n_rows: int) -> List[Dict]:
     """
     rows = []
     for i in range(0, n_rows):
-        rows.append(mock_koza.get_row())
+        row = mock_koza.get_row()
+        if " ∩ " in row['Anatomical entity ID']:
+            entities = [entity.strip().replace('"','') for entity in row['Anatomical entity ID'].split(" ∩ ")]
+            for entity in entities:
+                split_row = row.copy()
+                split_row['Anatomical entity ID'] = entity.strip()
+                rows.append(split_row)
+        else:
+            rows.append(row)
     return rows
 
 
@@ -87,14 +95,20 @@ def bgee_test_output_format() -> str:
 
 
 @pytest.fixture
-def bgee_test_files() -> List[str]:
+def bgee_test_file() -> List[str]:
     return ["tests/unit/bgee/test_bgee.tsv.gz"]
 
+@pytest.fixture
+def bgee_test_file_2() -> List[str]:
+    return ["tests/unit/bgee/test_bgee_2.tsv.gz"]
 
 @pytest.fixture
-def bgee_mock_koza_rows(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_files) -> KozaApp:
-    return get_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_files)
+def bgee_mock_koza_rows(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file) -> KozaApp:
+    return get_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file)
 
+@pytest.fixture
+def bgee_mock_koza_rows_2(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file_2) -> KozaApp:
+    return get_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file_2)
 
 @pytest.fixture
 def row_group_1(bgee_mock_koza_rows) -> List[Dict]:
@@ -106,6 +120,10 @@ def row_group_2(bgee_mock_koza_rows) -> List[Dict]:
     _ = get_koza_rows(bgee_mock_koza_rows, 5)
     return get_koza_rows(bgee_mock_koza_rows, 22)
 
+@pytest.fixture
+def row_group_3(bgee_mock_koza_rows_2) -> List[Dict]:
+    return get_koza_rows(bgee_mock_koza_rows_2, 3)
+
 
 @pytest.fixture
 def filter_col() -> str:
@@ -146,8 +164,12 @@ def test_filter_group_by_rank_long(row_group_2, filter_col, smallest_n):
 
 
 @pytest.fixture
-def bgee_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_files) -> KozaApp:
-    return get_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_files)
+def bgee_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file) -> KozaApp:
+    return get_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file)
+
+@pytest.fixture
+def bgee_mock_koza_2(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file_2) -> KozaApp:
+    return get_mock_koza(bgee_yaml, global_table, bgee_test_output, bgee_test_output_format, bgee_test_file_2)
 
 
 def test_write_group(row_group_1, bgee_mock_koza):
@@ -165,17 +187,38 @@ def test_write_group(row_group_1, bgee_mock_koza):
         assert item.subject == 'ENSEMBL:ENSSSCG00000000002'
         assert item.object == object_list[index]
 
+def test_write_group_2(row_group_3, bgee_mock_koza_2):
+    write_group(row_group_3, bgee_mock_koza_2)
+    write_result: list[GeneToExpressionSiteAssociation] = bgee_mock_koza_2._entities
+    assert len(write_result) == 5
+    object_list = ['UBERON:0000473', 'CL:0000089', 'UBERON:0000123', 'UBERON:0000473', 'CL:0000089']
+    predicate_ist = ['ENSEMBL:ENSSSCG00000000419', 'ENSEMBL:ENSSSCG00000000419', 'ENSEMBL:ENSSSCG00000000419', 'ENSEMBL:ENSSSCG00000000457', 'ENSEMBL:ENSSSCG00000000457']
+    prev_uuid = 0
+    for index, item in enumerate(write_result):
+        assert isinstance(item, GeneToExpressionSiteAssociation)
+        assert item.id != prev_uuid
+        prev_uuid = item.id
+        assert item.category == ['biolink:GeneToExpressionSiteAssociation']
+        assert item.predicate == 'biolink:expressed_in'
+        assert item.subject == predicate_ist[index]
+        assert item.object == object_list[index]
 
-def test_get_row_group(bgee_mock_koza, row_group_1, filter_col) -> List:
+def test_get_row_group(bgee_mock_koza, row_group_1, filter_col):
     row_group = get_row_group(bgee_mock_koza)
-
     assert isinstance(row_group, list)
     assert len(row_group) == 5
     for i in row_group:
         assert isinstance(i, dict)
 
     assert row_group == row_group_1
 
+def test_get_row_group_inter_split(bgee_mock_koza_2, row_group_3, filter_col) -> List:
+    row_group = get_row_group(bgee_mock_koza_2)
+    assert isinstance(row_group, list)
+    assert len(row_group) == 3
+    assert len(row_group_3) == 5
+    for i in row_group:
+        assert isinstance(i, dict)
 
 # Ignoring process_koza_sources for now as it depends completely on above tested functions but goes deeper into Koza.
 # def test_process_koza_source(bgee_mock_koza):