Add bundled schema with $id and update local fragment resolution logic (#2646)

Author: koxudaxi

src/datamodel_code_generator/reference.py

@@ -585,7 +585,7 @@ def resolve_ref(self, path: Sequence[str] | str) -> str:  # noqa: PLR0911, PLR09
         else:
             if "#" not in joined_path:
                 joined_path += "#"
-            elif joined_path[0] == "#" and not self.base_url:
+            elif joined_path[0] == "#" and self.current_root:
                 joined_path = f"{'/'.join(self.current_root)}{joined_path}"
 
             file_path, fragment = joined_path.split("#", 1)

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

@@ -0,0 +1,31 @@
+# generated by datamodel-codegen:
+#   filename:  bundled_schema_with_id.json
+#   timestamp: 2019-07-26T00:00:00+00:00
+
+from __future__ import annotations
+
+from enum import Enum
+from typing import Optional
+
+from pydantic import BaseModel
+
+
+class Species(Enum):
+    dog = 'dog'
+    cat = 'cat'
+    bird = 'bird'
+
+
+class Pet(BaseModel):
+    name: str
+    species: Species
+
+
+class User(BaseModel):
+    name: str
+    pet: Optional[Pet] = None
+
+
+class BundledSchema(BaseModel):
+    user: Optional[User] = None
+    pet: Optional[Pet] = None

tests/data/jsonschema/bundled_schema_with_id.json

@@ -0,0 +1,45 @@
+{
+  "$schema": "http://json-schema.org/draft-07/schema#",
+  "$id": "https://example.org/schemas/bundled.json",
+  "title": "Bundled Schema",
+  "description": "A bundled schema with $id that differs from fetch URL (Issue #1798)",
+  "type": "object",
+  "properties": {
+    "user": {
+      "$ref": "#/definitions/user"
+    },
+    "pet": {
+      "$ref": "#/definitions/pet"
+    }
+  },
+  "definitions": {
+    "user": {
+      "type": "object",
+      "properties": {
+        "name": {
+          "type": "string"
+        },
+        "pet": {
+          "$ref": "#/definitions/pet"
+        }
+      },
+      "required": ["name"]
+    },
+    "pet": {
+      "type": "object",
+      "properties": {
+        "name": {
+          "type": "string"
+        },
+        "species": {
+          "$ref": "#/definitions/species"
+        }
+      },
+      "required": ["name", "species"]
+    },
+    "species": {
+      "type": "string",
+      "enum": ["dog", "cat", "bird"]
+    }
+  }
+}

tests/main/jsonschema/test_main_jsonschema.py

@@ -3610,3 +3610,59 @@ def test_main_jsonschema_reserved_field_name_pydantic(output_file: Path) -> None
             "3.11",
         ],
     )
+
+
+@pytest.mark.benchmark
+@LEGACY_BLACK_SKIP
+def test_main_bundled_schema_with_id_local_file(output_file: Path) -> None:
+    """Test bundled schema with $id using local file input (Issue #1798)."""
+    run_main_and_assert(
+        input_path=JSON_SCHEMA_DATA_PATH / "bundled_schema_with_id.json",
+        output_path=output_file,
+        input_file_type="jsonschema",
+        assert_func=assert_file_content,
+        expected_file="bundled_schema_with_id.py",
+        extra_args=[
+            "--output-model-type",
+            "pydantic_v2.BaseModel",
+        ],
+    )
+
+
+@pytest.mark.benchmark
+@LEGACY_BLACK_SKIP
+def test_main_bundled_schema_with_id_url(mocker: MockerFixture, output_file: Path) -> None:
+    """Test bundled schema with $id using URL input produces same output as local file."""
+    schema_path = JSON_SCHEMA_DATA_PATH / "bundled_schema_with_id.json"
+
+    mock_response = mocker.Mock()
+    mock_response.text = schema_path.read_text()
+
+    httpx_get_mock = mocker.patch(
+        "httpx.get",
+        return_value=mock_response,
+    )
+
+    run_main_url_and_assert(
+        url="https://cdn.example.com/schemas/bundled_schema_with_id.json",
+        output_path=output_file,
+        input_file_type="jsonschema",
+        assert_func=assert_file_content,
+        expected_file="bundled_schema_with_id.py",
+        extra_args=[
+            "--output-model-type",
+            "pydantic_v2.BaseModel",
+        ],
+        transform=lambda s: s.replace(
+            "#   filename:  https://cdn.example.com/schemas/bundled_schema_with_id.json",
+            "#   filename:  bundled_schema_with_id.json",
+        ),
+    )
+
+    httpx_get_mock.assert_called_once_with(
+        "https://cdn.example.com/schemas/bundled_schema_with_id.json",
+        headers=None,
+        verify=True,
+        follow_redirects=True,
+        params=None,
+    )

tests/test_reference.py

@@ -325,3 +325,34 @@ def test_url_ref_matches_local_id_preserves_empty_json_pointer_token() -> None:
     result = resolver.resolve_ref("https://example.org/types#/items//child")
 
     assert result == "#/$defs/types/items//child"
+
+
+def test_resolve_ref_local_fragment_with_base_url_and_current_root() -> None:
+    """Local fragment refs should resolve to current_root when it's set, even with base_url (Issue #1798)."""
+    resolver = ModelResolver(base_url="https://raw.githubusercontent.com/user/repo/schema.json")
+    resolver.set_root_id("https://cveproject.github.io/schema/schema.json")
+    resolver.set_current_root(["https://raw.githubusercontent.com/user/repo/schema.json"])
+
+    result = resolver.resolve_ref("#/definitions/Foo")
+
+    assert result == "https://raw.githubusercontent.com/user/repo/schema.json#/definitions/Foo"
+
+
+def test_resolve_ref_local_fragment_with_different_host_base_url_and_root_id() -> None:
+    """Local fragment refs should resolve correctly when base_url and root_id have different hosts (Issue #1798)."""
+    resolver = ModelResolver(base_url="https://raw.githubusercontent.com/user/repo/schema.json")
+    resolver.set_root_id("https://cveproject.github.io/schema/schema.json")
+    resolver.set_current_root(["https://raw.githubusercontent.com/user/repo/schema.json"])
+
+    result = resolver.resolve_ref("#/definitions/product/properties/url")
+
+    assert result == "https://raw.githubusercontent.com/user/repo/schema.json#/definitions/product/properties/url"
+
+
+def test_resolve_ref_local_fragment_without_current_root_falls_back_to_url() -> None:
+    """Local fragment refs without current_root should fall back to URL resolution (Issue #1798)."""
+    resolver = ModelResolver(base_url="https://example.com/schemas/main.json")
+
+    result = resolver.resolve_ref("#/definitions/Foo")
+
+    assert result == "https://example.com/schemas/main.json#/definitions/Foo"