fix: use TypeAdapter for union dict defaults in type aliases and inline unions

Author: koxudaxi

src/datamodel_code_generator/model/pydantic_base.py

@@ -141,6 +141,47 @@ def _get_default_as_pydantic_model(self) -> str | None:  # noqa: PLR0911, PLR091
                     f"lambda :[{data_type_child.alias or data_type_child.reference.source.class_name}."
                     f"{self._PARSE_METHOD}(v) for v in {self.default!r}]"
                 )
+        # If the field references a type alias whose underlying type is a union of BaseModels,
+        # use TypeAdapter to validate the default value against the full union.
+        if (
+            not self.data_type.data_types
+            and self.data_type.reference
+            and isinstance(self.data_type.reference.source, DataModel)
+            and getattr(self.data_type.reference.source, "IS_ALIAS", False)
+            and isinstance(self.default, (dict, list))
+        ):
+            alias_source = self.data_type.reference.source
+            if alias_source.fields and alias_source.fields[0].data_type.data_types:
+                has_base_model = any(
+                    dt.reference and isinstance(dt.reference.source, BaseModelBase)
+                    for dt in alias_source.fields[0].data_type.data_types
+                )
+                if has_base_model:
+                    alias_name = self.data_type.alias or alias_source.class_name
+                    self.__dict__["_type_adapter_import_needed"] = True
+                    return f"lambda :TypeAdapter({alias_name}).validate_python({self.default!r})"
+
+        # Inline union with multiple BaseModel candidates: validate against the full union via TypeAdapter
+        if (
+            self.data_type.is_union
+            and isinstance(self.default, dict)
+            and not any(dt.is_dict for dt in self.data_type.data_types)
+        ):
+            base_model_count = sum(
+                1 for dt in self.data_type.data_types if dt.reference and isinstance(dt.reference.source, BaseModelBase)
+            )
+            if base_model_count > 1:
+                type_parts = []
+                for dt in self.data_type.data_types:
+                    hint = dt.type_hint
+                    if hint and hint != "None":
+                        type_parts.append(hint)
+                if type_parts:
+                    union_expr = ", ".join(type_parts)
+                    self.__dict__["_type_adapter_import_needed"] = True
+                    self.__dict__["_union_import_for_type_adapter"] = True
+                    return f"lambda :TypeAdapter(Union[{union_expr}]).validate_python({self.default!r})"
+
         for data_type in self.data_type.data_types or (self.data_type,):
             # TODO: Check nested data_types
             if data_type.is_dict:

src/datamodel_code_generator/model/pydantic_v2/base_model.py

@@ -12,7 +12,7 @@
 
 from pydantic import Field, field_validator, model_validator
 
-from datamodel_code_generator.imports import IMPORT_ANY, Import
+from datamodel_code_generator.imports import IMPORT_ANY, IMPORT_UNION, Import
 from datamodel_code_generator.model.base import ALL_MODEL, UNDEFINED, BaseClassDataType, DataModelFieldBase
 from datamodel_code_generator.model.imports import IMPORT_CLASSVAR
 from datamodel_code_generator.model.pydantic_base import (
@@ -200,6 +200,12 @@ def imports(self) -> tuple[Import, ...]:
             extra_imports.append(IMPORT_ALIAS_CHOICES)
         if self._has_discriminator_in_data_type():
             extra_imports.append(IMPORT_FIELD)
+        if getattr(self, "_type_adapter_import_needed", False):
+            from datamodel_code_generator.model.pydantic_v2.imports import IMPORT_TYPE_ADAPTER  # noqa: PLC0415
+
+            extra_imports.append(IMPORT_TYPE_ADAPTER)
+        if getattr(self, "_union_import_for_type_adapter", False):
+            extra_imports.append(IMPORT_UNION)
         if extra_imports:
             return chain_as_tuple(base_imports, tuple(extra_imports))
         return base_imports

src/datamodel_code_generator/model/pydantic_v2/imports.py

@@ -20,6 +20,7 @@
 IMPORT_SERIALIZE_AS_ANY = Import.from_full_path("pydantic.SerializeAsAny")
 IMPORT_PYDANTIC_DATACLASS = Import.from_full_path("pydantic.dataclasses.dataclass")
 IMPORT_ROOT_MODEL = Import.from_full_path("pydantic.RootModel")
+IMPORT_TYPE_ADAPTER = Import.from_full_path("pydantic.TypeAdapter")
 IMPORT_FIELD_VALIDATOR = Import.from_full_path("pydantic.field_validator")
 IMPORT_VALIDATION_INFO = Import.from_full_path("pydantic.ValidationInfo")
 IMPORT_VALIDATOR_FUNCTION_WRAP_HANDLER = Import.from_full_path("pydantic.ValidatorFunctionWrapHandler")

tests/data/expected/main/jsonschema/pydantic_v2_multi_model_union_default_object.py

@@ -0,0 +1,28 @@
+# generated by datamodel-codegen:
+#   filename:  multi_model_union_default_object.json
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Annotated, Literal, Union
+
+from pydantic import BaseModel, Field, TypeAdapter
+
+
+class A(BaseModel):
+    type: Literal['a']
+
+
+class B(BaseModel):
+    type: Literal['b']
+
+
+class Model(BaseModel):
+    x: Annotated[
+        A | B | None,
+        Field(
+            default_factory=lambda: TypeAdapter(Union[A, B]).validate_python(
+                {'type': 'b'}
+            )
+        ),
+    ]

tests/data/expected/main/jsonschema/pydantic_v2_multi_model_union_default_object_type_alias.py

@@ -0,0 +1,28 @@
+# generated by datamodel-codegen:
+#   filename:  multi_model_union_default_object.json
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Annotated, Literal, Union
+
+from pydantic import BaseModel, Field, TypeAdapter
+
+
+class A(BaseModel):
+    type: Literal['a']
+
+
+class B(BaseModel):
+    type: Literal['b']
+
+
+class Model(BaseModel):
+    x: Annotated[
+        A | B | None,
+        Field(
+            default_factory=lambda: TypeAdapter(Union[A, B]).validate_python(
+                {'type': 'b'}
+            )
+        ),
+    ]

tests/data/expected/main/jsonschema/ref_union_default_object.py

@@ -0,0 +1,27 @@
+# generated by datamodel-codegen:
+#   filename:  ref_union_default_object.json
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Annotated, Literal
+
+from pydantic import BaseModel, Field, TypeAdapter
+
+
+class A(BaseModel):
+    type: Literal['a']
+
+
+class B(BaseModel):
+    type: Literal['b']
+
+
+type U = A | B
+
+
+class Model(BaseModel):
+    x: Annotated[
+        U | None,
+        Field(default_factory=lambda: TypeAdapter(U).validate_python({'type': 'b'})),
+    ]

tests/data/jsonschema/multi_model_union_default_object.json

@@ -0,0 +1,35 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "object",
+  "$defs": {
+    "A": {
+      "type": "object",
+      "properties": {
+        "type": {
+          "const": "a",
+          "type": "string"
+        }
+      },
+      "required": ["type"]
+    },
+    "B": {
+      "type": "object",
+      "properties": {
+        "type": {
+          "const": "b",
+          "type": "string"
+        }
+      },
+      "required": ["type"]
+    }
+  },
+  "properties": {
+    "x": {
+      "anyOf": [
+        { "$ref": "#/$defs/A" },
+        { "$ref": "#/$defs/B" }
+      ],
+      "default": { "type": "b" }
+    }
+  }
+}

tests/data/jsonschema/ref_union_default_object.json

@@ -0,0 +1,38 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "type": "object",
+  "$defs": {
+    "A": {
+      "type": "object",
+      "properties": {
+        "type": {
+          "const": "a",
+          "type": "string"
+        }
+      },
+      "required": ["type"]
+    },
+    "B": {
+      "type": "object",
+      "properties": {
+        "type": {
+          "const": "b",
+          "type": "string"
+        }
+      },
+      "required": ["type"]
+    },
+    "U": {
+      "anyOf": [
+        { "$ref": "#/$defs/A" },
+        { "$ref": "#/$defs/B" }
+      ]
+    }
+  },
+  "properties": {
+    "x": {
+      "$ref": "#/$defs/U",
+      "default": { "type": "b" }
+    }
+  }
+}

tests/main/jsonschema/test_main_jsonschema.py

@@ -5591,6 +5591,74 @@ def test_main_jsonschema_enum_literal_type_alias_default(output_file: Path) -> N
     )
 
 
+@pytest.mark.skipif(
+    int(black.__version__.split(".")[0]) < 23,
+    reason="Installed black doesn't support the new 'type' statement",
+)
+def test_main_jsonschema_ref_union_default_object(output_file: Path) -> None:
+    """TypeAlias ref to union of BaseModels uses TypeAdapter for dict default (#3034)."""
+    run_main_and_assert(
+        input_path=JSON_SCHEMA_DATA_PATH / "ref_union_default_object.json",
+        output_path=output_file,
+        input_file_type=None,
+        assert_func=assert_file_content,
+        expected_file="ref_union_default_object.py",
+        extra_args=[
+            "--use-type-alias",
+            "--use-annotated",
+            "--target-python-version",
+            "3.12",
+            "--output-model-type",
+            "pydantic_v2.BaseModel",
+        ],
+    )
+
+
+@pytest.mark.skipif(
+    int(black.__version__.split(".")[0]) < 23,
+    reason="Installed black doesn't support the new 'type' statement",
+)
+def test_main_jsonschema_pydantic_v2_multi_model_union_default_object_type_alias(output_file: Path) -> None:
+    """Inline union of multiple BaseModels uses TypeAdapter with --use-type-alias (#3035)."""
+    run_main_and_assert(
+        input_path=JSON_SCHEMA_DATA_PATH / "multi_model_union_default_object.json",
+        output_path=output_file,
+        input_file_type=None,
+        assert_func=assert_file_content,
+        expected_file="pydantic_v2_multi_model_union_default_object_type_alias.py",
+        extra_args=[
+            "--use-type-alias",
+            "--use-annotated",
+            "--target-python-version",
+            "3.12",
+            "--output-model-type",
+            "pydantic_v2.BaseModel",
+        ],
+    )
+
+
+@pytest.mark.skipif(
+    int(black.__version__.split(".")[0]) < 23,
+    reason="Installed black doesn't support the new 'type' statement",
+)
+def test_main_jsonschema_pydantic_v2_multi_model_union_default_object(output_file: Path) -> None:
+    """Inline union of multiple BaseModels uses TypeAdapter without --use-type-alias (#3035)."""
+    run_main_and_assert(
+        input_path=JSON_SCHEMA_DATA_PATH / "multi_model_union_default_object.json",
+        output_path=output_file,
+        input_file_type=None,
+        assert_func=assert_file_content,
+        expected_file="pydantic_v2_multi_model_union_default_object.py",
+        extra_args=[
+            "--use-annotated",
+            "--target-python-version",
+            "3.12",
+            "--output-model-type",
+            "pydantic_v2.BaseModel",
+        ],
+    )
+
+
 @pytest.mark.cli_doc(
     options=["--type-mappings"],
     option_description="""Override default type mappings for schema formats.