fix(api): Resolve error encountered when executing `QuestionClassifieNode` (#20829)

The `QuestionClassifierNode` class extends `LLMNode`, meaning that, per the Liskov Substitution Principle, `QuestionClassifierNodeData` **SHOULD** be compatible in contexts where `LLMNodeData` is expected.

However, the absence of the `structured_output_enabled` attribute violates this principle, causing `QuestionClassifierNode` to fail during execution.

This commit implements a quick and temporary workaround. A proper resolution would involve refactoring to decouple `QuestionClassifierNode` from `LLMNode` to address the underlying design issue.

Fixes #20725.

api/core/workflow/nodes/knowledge_retrieval/entities.py

@@ -132,3 +132,12 @@ class KnowledgeRetrievalNodeData(BaseNodeData):
     metadata_model_config: Optional[ModelConfig] = None
     metadata_filtering_conditions: Optional[MetadataFilteringCondition] = None
     vision: VisionConfig = Field(default_factory=VisionConfig)
+
+    @property
+    def structured_output_enabled(self) -> bool:
+        # NOTE(QuantumGhost): Temporary workaround for issue #20725
+        # (https://github.com/langgenius/dify/issues/20725).
+        #
+        # The proper fix would be to make `KnowledgeRetrievalNode` inherit
+        # from `BaseNode` instead of `LLMNode`.
+        return False

api/core/workflow/nodes/question_classifier/entities.py

@@ -19,3 +19,12 @@ class QuestionClassifierNodeData(BaseNodeData):
     instruction: Optional[str] = None
     memory: Optional[MemoryConfig] = None
     vision: VisionConfig = Field(default_factory=VisionConfig)
+
+    @property
+    def structured_output_enabled(self) -> bool:
+        # NOTE(QuantumGhost): Temporary workaround for issue #20725
+        # (https://github.com/langgenius/dify/issues/20725).
+        #
+        # The proper fix would be to make `QuestionClassifierNode` inherit
+        # from `BaseNode` instead of `LLMNode`.
+        return False