Merge pull request openMetadataInitiative#81 from apdavison/improved-by-name

Improvements to the `by_name()` method; fixes openMetadataInitiative#69

Author: Raphael-Gazzotti

pipeline/src/additional_methods/by_name.py.txt

@@ -3,12 +3,52 @@
         return [value for value in cls.__dict__.values() if isinstance(value, cls)]
 
     @classmethod
-    def by_name(cls, name):
+    def by_name(
+        cls,
+        name: str,
+        match: str = "equals",
+        all: bool = False,
+    ):
+        """
+        Search for instances in the openMINDS instance library based on their name.
+
+        This includes properties "name", "lookup_label", "family_name", "full_name", "short_name", "abbreviation", and "synonyms".
+
+        Note that not all metadata classes have a name.
+
+        Args:
+            name (str): a string to search for.
+            match (str, optional): either "equals" (exact match - default) or "contains".
+            all (bool, optional): Whether to return all objects that match the name, or only the first. Defaults to False.
+        """
+        namelike_properties = ("name", "lookup_label", "family_name", "full_name", "short_name", "abbreviation")
         if cls._instance_lookup is None:
             cls._instance_lookup = {}
             for instance in cls.instances():
-                cls._instance_lookup[instance.name] = instance
-                if instance.synonyms:
-                    for synonym in instance.synonyms:
-                        cls._instance_lookup[synonym] = instance
-        return cls._instance_lookup[name]
+                keys = []
+                for prop_name in namelike_properties:
+                    if hasattr(instance, prop_name):
+                        keys.append(getattr(instance, prop_name))
+                if hasattr(instance, "synonyms"):
+                    for synonym in instance.synonyms or []:
+                        keys.append(synonym)
+                for key in keys:
+                    if key in cls._instance_lookup:
+                        cls._instance_lookup[key].append(instance)
+                    else:
+                        cls._instance_lookup[key] = [instance]
+        if match == "equals":
+            matches = cls._instance_lookup.get(name, None)
+        elif match == "contains":
+            matches = []
+            for key, instances in cls._instance_lookup.items():
+                if name in key:
+                    matches.extend(instances)
+        else:
+            raise ValueError("'match' must be either 'equals' or 'contains'")
+        if all:
+            return matches
+        elif len(matches) > 0:
+            return matches[0]
+        else:
+            return None

pipeline/tests/test_regressions.py

@@ -300,3 +300,34 @@ def test_issue0073b(om):
     ds1.is_variant_of = ds2
 
     failures = ds1.validate()
+
+
+@pytest.mark.parametrize("om", [openminds.latest])
+def test_issue0069(om):
+    # https://github.com/openMetadataInitiative/openMINDS_Python/issues/69
+    # The License class has a classmethod "by_name()" which assumes License is a controlled term
+    # (i.e., it has properties "name" and "synonyms").
+    # However License does not have these properties, it has "short_name" and "full_name".
+
+    # Test with default arguments (single result, exact match)
+    result = om.core.License.by_name("CC-BY-4.0")
+    assert result.short_name == "CC-BY-4.0"
+
+    result = om.sands.ParcellationEntity.by_name("NODa,b")
+    assert result.abbreviation == "NODa,b"
+
+    result = om.sands.CommonCoordinateSpace.by_name("MEBRAINS population-based monkey brain template")
+    assert result.full_name == "MEBRAINS population-based monkey brain template"
+
+    assert om.controlled_terms.BiologicalOrder.by_name("rodents") == om.controlled_terms.BiologicalOrder.by_name("Rodentia") != None
+
+    # Test with "all=True"
+    results = om.sands.BrainAtlasVersion.by_name("Julich-Brain Atlas", all=True)
+    assert len(results) == 30
+    assert all(r.short_name == "Julich-Brain Atlas" for r in results)
+    assert len(set(r.id for r in results)) == len(results)
+
+    # Test with "match='contains'"
+    results = om.core.License.by_name("Creative Commons", all=True, match="contains")
+    assert len(results) == 7
+    assert all("CC" in r.short_name for r in results)