Fix --use-unique-items-as-set to output set literals for default values (#2672)

* Add support for unique items as set literals in generated models

* Add support for unique items as set literals in generated models

Author: koxudaxi

src/datamodel_code_generator/model/base.py

@@ -48,6 +48,20 @@
 
 ALL_MODEL: str = "#all#"
 
+
+def repr_set_sorted(value: set[Any]) -> str:
+    """Return a repr of a set with elements sorted for consistent output.
+
+    Uses (type_name, repr(x)) as sort key to safely handle any type including
+    Enum, custom classes, or types without __lt__ defined.
+    """
+    if not value:
+        return "set()"
+    # Sort by type name first, then by repr for consistent output
+    sorted_elements = sorted(value, key=lambda x: (type(x).__name__, repr(x)))
+    return "{" + ", ".join(repr(e) for e in sorted_elements) + "}"
+
+
 ConstraintsBaseT = TypeVar("ConstraintsBaseT", bound="ConstraintsBase")
 DataModelFieldBaseT = TypeVar("DataModelFieldBaseT", bound="DataModelFieldBase")
 
@@ -281,6 +295,8 @@ def method(self) -> str | None:
     @property
     def represented_default(self) -> str:
         """Get the repr() string of the default value."""
+        if isinstance(self.default, set):
+            return repr_set_sorted(self.default)
         return repr(self.default)
 
     @property

src/datamodel_code_generator/model/dataclass.py

@@ -150,8 +150,13 @@ def __str__(self) -> str:
         if len(data) == 1 and "default" in data:
             default = data["default"]
 
-            if isinstance(default, (list, dict)):
-                return f"field(default_factory=lambda :{default!r})"
+            if isinstance(default, (list, dict, set)):
+                if default:
+                    from datamodel_code_generator.model.base import repr_set_sorted  # noqa: PLC0415
+
+                    default_repr = repr_set_sorted(default) if isinstance(default, set) else repr(default)
+                    return f"field(default_factory=lambda: {default_repr})"
+                return f"field(default_factory={type(default).__name__})"
             return repr(default)
         kwargs = [f"{k}={v if k == 'default_factory' else repr(v)}" for k, v in data.items()]
         return f"field({', '.join(kwargs)})"

src/datamodel_code_generator/model/msgspec.py

@@ -277,7 +277,10 @@ def __str__(self) -> str:
         if "default" in data and isinstance(data["default"], (list, dict, set)) and "default_factory" not in data:
             default_value = data.pop("default")
             if default_value:
-                data["default_factory"] = f"lambda: {default_value!r}"
+                from datamodel_code_generator.model.base import repr_set_sorted  # noqa: PLC0415
+
+                default_repr = repr_set_sorted(default_value) if isinstance(default_value, set) else repr(default_value)
+                data["default_factory"] = f"lambda: {default_repr}"
             else:
                 data["default_factory"] = type(default_value).__name__
 

src/datamodel_code_generator/model/pydantic/base_model.py

@@ -221,7 +221,10 @@ def __str__(self) -> str:  # noqa: PLR0912
         elif self.required:
             field_arguments = ["...", *field_arguments]
         elif not default_factory:
-            field_arguments = [f"{self.default!r}", *field_arguments]
+            from datamodel_code_generator.model.base import repr_set_sorted  # noqa: PLC0415
+
+            default_repr = repr_set_sorted(self.default) if isinstance(self.default, set) else repr(self.default)
+            field_arguments = [default_repr, *field_arguments]
 
         return f"Field({', '.join(field_arguments)})"
 

src/datamodel_code_generator/parser/base.py

@@ -1390,6 +1390,15 @@ def __replace_unique_list_to_set(self, models: list[DataModel]) -> None:
                     continue
                 set_data_type = self._create_set_from_list(model_field.data_type)
                 if set_data_type:  # pragma: no cover
+                    # Check if default list elements are hashable before converting type
+                    if isinstance(model_field.default, list):
+                        try:
+                            converted_default = set(model_field.default)
+                        except TypeError:
+                            # Elements are not hashable (e.g., contains dicts)
+                            # Skip both type and default conversion to keep consistency
+                            continue
+                        model_field.default = converted_default
                     model_field.replace_data_type(set_data_type)
 
     @classmethod

tests/data/expected/main/openapi/unique_items_default_set_dataclass.py

@@ -0,0 +1,15 @@
+# generated by datamodel-codegen:
+#   filename:  unique_items_default_set.yaml
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Optional, Set
+
+
+@dataclass
+class TestModel:
+    tags: Optional[Set[str]] = field(default_factory=lambda: {'tag1', 'tag2'})
+    empty_tags: Optional[Set[str]] = field(default_factory=set)
+    numbers: Optional[Set[int]] = field(default_factory=lambda: {1, 2, 3})

tests/data/expected/main/openapi/unique_items_default_set_msgspec.py

@@ -0,0 +1,15 @@
+# generated by datamodel-codegen:
+#   filename:  unique_items_default_set.yaml
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Set, Union
+
+from msgspec import Struct, UnsetType, field
+
+
+class TestModel(Struct):
+    tags: Union[Set[str], UnsetType] = field(default_factory=lambda: {'tag1', 'tag2'})
+    empty_tags: Union[Set[str], UnsetType] = field(default_factory=set)
+    numbers: Union[Set[int], UnsetType] = field(default_factory=lambda: {1, 2, 3})

tests/data/expected/main/openapi/unique_items_default_set_pydantic.py

@@ -0,0 +1,15 @@
+# generated by datamodel-codegen:
+#   filename:  unique_items_default_set.yaml
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Optional, Set
+
+from pydantic import BaseModel, Field
+
+
+class TestModel(BaseModel):
+    tags: Optional[Set[str]] = Field({'tag1', 'tag2'}, unique_items=True)
+    empty_tags: Optional[Set[str]] = Field(set(), unique_items=True)
+    numbers: Optional[Set[int]] = Field({1, 2, 3}, unique_items=True)

tests/data/expected/main/openapi/unique_items_default_set_pydantic_v2.py

@@ -0,0 +1,15 @@
+# generated by datamodel-codegen:
+#   filename:  unique_items_default_set.yaml
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Optional, Set
+
+from pydantic import BaseModel
+
+
+class TestModel(BaseModel):
+    tags: Optional[Set[str]] = {'tag1', 'tag2'}
+    empty_tags: Optional[Set[str]] = set()
+    numbers: Optional[Set[int]] = {1, 2, 3}

tests/data/openapi/unique_items_default_set.yaml

@@ -0,0 +1,33 @@
+openapi: "3.0.0"
+info:
+  title: Test API - Unique Items Default as Set
+  version: 1.0.0
+paths: {}
+components:
+  schemas:
+    TestModel:
+      type: object
+      properties:
+        tags:
+          type: array
+          items:
+            type: string
+          uniqueItems: true
+          default:
+            - tag1
+            - tag2
+        empty_tags:
+          type: array
+          items:
+            type: string
+          uniqueItems: true
+          default: []
+        numbers:
+          type: array
+          items:
+            type: integer
+          uniqueItems: true
+          default:
+            - 1
+            - 2
+            - 3