• X
• Mastodon
• LinkedIn
• Facebook

src/routers/openml/datasets.py (1)

296-298: Optional: defer ontology query until after empty-feature guard.

Line 296 currently performs an extra read even when no features exist and the endpoint will raise immediately.

⚡ Small query-order optimization

-    ontologies = database.datasets.get_feature_ontologies(dataset_id, expdb)
-    for feature in features:
-        feature.ontology = ontologies.get(feature.index)
-
     if not features:
         processing_state = database.datasets.get_latest_processing_update(dataset_id, expdb)
         ...
         raise HTTPException(...)

+    ontologies = database.datasets.get_feature_ontologies(dataset_id, expdb)
+    for feature in features:
+        feature.ontology = ontologies.get(feature.index)

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@src/routers/openml/datasets.py` around lines 296 - 298, The code calls
database.datasets.get_feature_ontologies(dataset_id, expdb) before checking if
there are any features, causing an unnecessary DB read; change the order so you
first check the features collection (e.g., if not features: return or raise as
the endpoint expects) and only call get_feature_ontologies(dataset_id, expdb)
when features is non-empty, then iterate and set feature.ontology =
ontologies.get(feature.index); this uses the existing symbols features,
dataset_id, expdb, and database.datasets.get_feature_ontologies.

tests/routers/openml/migration/datasets_migration_test.py (1)

225-228: Consider deduplicating legacy list-to-scalar normalization.

Line 224 and Line 228 apply the same conversion pattern; extracting a tiny helper would reduce drift in future parity tweaks.

♻️ Minimal refactor

+    def _legacy_scalar_or_list(values: list[str]) -> str | list[str]:
+        return values if len(values) > 1 else values[0]
+
     for feature in python_body:
         for key, value in list(feature.items()):
             if key == "nominal_values":
                 values = feature.pop(key)
-                feature["nominal_value"] = values if len(values) > 1 else values[0]
+                feature["nominal_value"] = _legacy_scalar_or_list(values)
             elif key == "ontology":
                 values = feature.pop(key)
-                feature["ontology"] = values if len(values) > 1 else values[0]
+                feature["ontology"] = _legacy_scalar_or_list(values)

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/routers/openml/migration/datasets_migration_test.py` around lines 225 -
228, Extract the repeated "list-to-scalar" normalization into a small helper
(e.g., normalize_legacy_list(value) or normalize_list_to_scalar) and replace
both in-place patterns that operate on feature["ontology"] (the block that does
values = feature.pop(key); feature["ontology"] = values if len(values) > 1 else
values[0]) with a call to that helper so the code becomes feature["ontology"] =
normalize_legacy_list(feature.pop(key)); ensure the helper accepts the popped
value (either list or str) and returns the list when length > 1 or the single
element otherwise.

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

• X
• Mastodon
• LinkedIn
• Facebook