Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
test: use unbound_session for legacy agent node validator test (#44996)
hjeesum86f0bf1a8b37
diff --git a/api/tests/unit_tests/core/workflow/nodes/agent_v2/test_validators.py b/api/tests/unit_tests/core/workflow/nodes/agent_v2/test_validators.pyindex d3526c85ee1..f9b2fc7e074 100644--- a/api/tests/unit_tests/core/workflow/nodes/agent_v2/test_validators.py+++ b/api/tests/unit_tests/core/workflow/nodes/agent_v2/test_validators.py@@ -137,16 +137,13 @@ def _tool_graph(tool_data: dict) -> dict:}-def test_historical_agent_version_two_is_not_validated_as_dify_agent() -> None:+def test_historical_agent_version_two_is_not_validated_as_dify_agent(unbound_session: Session) -> None:graph = {"nodes": [{"id": "legacy-agent", "data": {"type": "agent", "version": "2"}}],"edges": [],}- session = Mock()-- WorkflowAgentNodeValidator.validate_published_workflow(session=session, workflow=_workflow(graph))- session.scalar.assert_not_called()+ WorkflowAgentNodeValidator.validate_published_workflow(session=unbound_session, workflow=_workflow(graph))def _persist_validation_scope(