Fix allOf with multiple refs having same property name generates broken models (#2579)

* Add support for merging schemas with duplicate property names and boolean properties

* Fix normalization of property schema for allOf merging to handle non-serializable types

Author: koxudaxi

src/datamodel_code_generator/parser/jsonschema.py

@@ -7,6 +7,7 @@
 from __future__ import annotations
 
 import enum as _enum
+import json
 from collections import defaultdict
 from contextlib import contextmanager
 from functools import cached_property, lru_cache
@@ -784,6 +785,53 @@ def _deep_merge(self, dict1: dict[Any, Any], dict2: dict[Any, Any]) -> dict[Any,
             result[key] = value
         return result
 
+    def _load_ref_schema_object(self, ref: str) -> JsonSchemaObject:
+        """Load a JsonSchemaObject from a $ref using standard resolve/load pipeline."""
+        resolved_ref = self.model_resolver.resolve_ref(ref)
+        file_part, fragment = ([*resolved_ref.split("#", 1), ""])[:2]
+        raw_doc = self._get_ref_body(file_part) if file_part else self.raw_obj
+
+        target_schema: dict[str, YamlValue] | YamlValue = raw_doc
+        if fragment:
+            pointer = [p for p in fragment.split("/") if p]
+            target_schema = get_model_by_path(raw_doc, pointer)
+
+        return self.SCHEMA_OBJECT_TYPE.parse_obj(target_schema)
+
+    def _schema_signature(self, prop_schema: JsonSchemaObject | bool) -> str | bool:  # noqa: FBT001, PLR6301
+        """Normalize property schema for comparison across allOf items."""
+        if isinstance(prop_schema, bool):
+            return prop_schema
+        return json.dumps(prop_schema.dict(exclude_unset=True, by_alias=True), sort_keys=True, default=repr)
+
+    def _merge_all_of_object(self, obj: JsonSchemaObject) -> JsonSchemaObject | None:
+        """Merge allOf items when they share object properties to avoid duplicate models."""
+        resolved_items: list[JsonSchemaObject] = []
+        property_signatures: dict[str, set[str | bool]] = {}
+        for item in obj.allOf:
+            resolved_item = self._load_ref_schema_object(item.ref) if item.ref else item
+            resolved_items.append(resolved_item)
+            if resolved_item.properties:
+                for prop_name, prop_schema in resolved_item.properties.items():
+                    property_signatures.setdefault(prop_name, set()).add(self._schema_signature(prop_schema))
+
+        if obj.properties:
+            for prop_name, prop_schema in obj.properties.items():
+                property_signatures.setdefault(prop_name, set()).add(self._schema_signature(prop_schema))
+
+        if not any(len(signatures) > 1 for signatures in property_signatures.values()):
+            return None
+
+        merged_schema: dict[str, Any] = obj.dict(exclude={"allOf"}, exclude_unset=True, by_alias=True)
+        for resolved_item in resolved_items:
+            merged_schema = self._deep_merge(merged_schema, resolved_item.dict(exclude_unset=True, by_alias=True))
+
+        if "required" in merged_schema and isinstance(merged_schema["required"], list):
+            merged_schema["required"] = list(dict.fromkeys(merged_schema["required"]))
+
+        merged_schema.pop("allOf", None)
+        return self.SCHEMA_OBJECT_TYPE.parse_obj(merged_schema)
+
     def parse_combined_schema(
         self,
         name: str,
@@ -986,6 +1034,10 @@ def parse_all_of(
                 and get_model_by_path(self.raw_obj, single_obj.ref[2:].split("/")).get("enum")
             ):
                 return self.get_ref_data_type(single_obj.ref)
+
+        merged_all_of_obj = self._merge_all_of_object(obj)
+        if merged_all_of_obj:
+            return self._parse_object_common_part(name, merged_all_of_obj, path, ignore_duplicate_model, [], [], [])
         fields: list[DataModelFieldBase] = []
         base_classes: list[Reference] = []
         required: list[str] = []

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

@@ -0,0 +1,25 @@
+# generated by datamodel-codegen:
+#   filename:  all_of_merge_boolean_property.json
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Any, Optional
+
+from pydantic import BaseModel
+
+
+class Model(BaseModel):
+    data: Optional[Any] = None
+
+
+class Data(BaseModel):
+    value: Optional[str] = None
+
+
+class Base1(BaseModel):
+    data: Optional[Data] = None
+
+
+class Base2(BaseModel):
+    data: Optional[Any] = None

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

@@ -0,0 +1,34 @@
+# generated by datamodel-codegen:
+#   filename:  all_of_merge_same_property.json
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from typing import Optional
+
+from pydantic import BaseModel
+
+
+class Links(BaseModel):
+    self: Optional[str] = None
+    collection: Optional[str] = None
+
+
+class Model(BaseModel):
+    links: Links
+
+
+class Links1(BaseModel):
+    self: Optional[str] = None
+
+
+class SelfLink(BaseModel):
+    links: Links1
+
+
+class Links2(BaseModel):
+    collection: Optional[str] = None
+
+
+class CollectionLink(BaseModel):
+    links: Links2

tests/data/jsonschema/all_of_merge_boolean_property.json

@@ -0,0 +1,25 @@
+{
+  "$defs": {
+    "Base1": {
+      "type": "object",
+      "properties": {
+        "data": {
+          "type": "object",
+          "properties": {
+            "value": { "type": "string" }
+          }
+        }
+      }
+    },
+    "Base2": {
+      "type": "object",
+      "properties": {
+        "data": false
+      }
+    }
+  },
+  "allOf": [
+    { "$ref": "#/$defs/Base1" },
+    { "$ref": "#/$defs/Base2" }
+  ]
+}

tests/data/jsonschema/all_of_merge_same_property.json

@@ -0,0 +1,32 @@
+{
+  "$defs": {
+    "SelfLink": {
+      "type": "object",
+      "required": ["links"],
+      "properties": {
+        "links": {
+          "type": "object",
+          "properties": {
+            "self": { "type": "string" }
+          }
+        }
+      }
+    },
+    "CollectionLink": {
+      "type": "object",
+      "required": ["links"],
+      "properties": {
+        "links": {
+          "type": "object",
+          "properties": {
+            "collection": { "type": "string" }
+          }
+        }
+      }
+    }
+  },
+  "allOf": [
+    { "$ref": "#/$defs/SelfLink" },
+    { "$ref": "#/$defs/CollectionLink" }
+  ]
+}

tests/main/jsonschema/test_main_jsonschema.py

@@ -908,6 +908,32 @@ def test_main_all_of_with_object(output_file: Path) -> None:
         )
 
 
+def test_main_all_of_merge_same_property(output_file: Path) -> None:
+    """Test allOf merging when duplicate property names exist across refs."""
+    with chdir(JSON_SCHEMA_DATA_PATH):
+        run_main_and_assert(
+            input_path=Path("all_of_merge_same_property.json"),
+            output_path=output_file,
+            input_file_type="jsonschema",
+            assert_func=assert_file_content,
+            expected_file="all_of_merge_same_property.py",
+            extra_args=["--class-name", "Model"],
+        )
+
+
+def test_main_all_of_merge_boolean_property(output_file: Path) -> None:
+    """Test allOf merging when a property has a boolean schema (false)."""
+    with chdir(JSON_SCHEMA_DATA_PATH):
+        run_main_and_assert(
+            input_path=Path("all_of_merge_boolean_property.json"),
+            output_path=output_file,
+            input_file_type="jsonschema",
+            assert_func=assert_file_content,
+            expected_file="all_of_merge_boolean_property.py",
+            extra_args=["--class-name", "Model"],
+        )
+
+
 @pytest.mark.skipif(
     black.__version__.split(".")[0] >= "24",
     reason="Installed black doesn't support the old style",