opensource/dify
3
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-09-24 23:22:26 +08:00
Code Issues Packages Projects Releases Wiki Activity

chore(api): Upgrade graphon to v0.5.1 (#37168)

Browse Source 
Co-authored-by: Yunlu Wen <yunlu.wen@dify.ai>
Co-authored-by: L1nSn0w <l1nsn0w@qq.com>
This commit is contained in:
QuantumGhost
2026-06-11 07:37:27 +00:00
committed by GitHub
co-authored by Yunlu Wen L1nSn0w
parent fb39df49c8
commit b4c50eb920
25 changed files with 450 additions and 90 deletions
Download Patch File Download Diff File Expand all files Collapse all files
api/tests/fixtures/workflow/response_stream_filter_issue_170_workflow.yml
Vendored
+138
View File
@@ -0,0 +1,138 @@
app:
description: Response stream ordering fixture matching graphon issue 170.
icon: 🤖
icon_background: '#FFEAD5'
mode: advanced-chat
name: response_stream_filter_issue_170_workflow
use_icon_as_answer_icon: false
dependencies: []
kind: app
version: 0.3.1
workflow:
conversation_variables: []
environment_variables: []
features:
file_upload: {}
opening_statement: ''
retriever_resource:
enabled: true
sensitive_word_avoidance:
enabled: false
speech_to_text:
enabled: false
suggested_questions: []
suggested_questions_after_answer:
enabled: false
text_to_speech:
enabled: false
language: ''
voice: ''
graph:
edges:
- id: start-llm
source: start
sourceHandle: source
target: llm
targetHandle: target
- id: llm-dufu
source: llm
sourceHandle: source
target: dufu
targetHandle: target
- id: dufu-answer
source: dufu
sourceHandle: source
target: answer
targetHandle: target
nodes:
- data:
desc: ''
title: Start
type: start
variables: []
id: start
position:
x: 80
y: 282
sourcePosition: right
targetPosition: left
type: custom
- data:
context:
enabled: false
variable_selector: []
desc: ''
memory:
query_prompt_template: '{{#sys.query#}}'
window:
enabled: false
size: 10
model:
completion_params:
temperature: 0.7
mode: chat
name: gpt-4o-mini
provider: openai
prompt_template:
- role: system
text: Please output a poem by Li Bai
selected: false
title: Li Bai
type: llm
variables: []
vision:
enabled: false
id: llm
position:
x: 380
y: 282
sourcePosition: right
targetPosition: left
type: custom
- data:
context:
enabled: false
variable_selector: []
desc: ''
model:
completion_params:
temperature: 0.7
mode: chat
name: gpt-4o-mini
provider: openai
prompt_template:
- role: system
text: Please output a poem by Du Fu
selected: false
title: Du Fu
type: llm
variables: []
vision:
enabled: false
id: dufu
position:
x: 680
y: 282
sourcePosition: right
targetPosition: left
type: custom
- data:
answer: |-
# Du Fu
{{#dufu.text#}}
# Li Bai
{{#llm.text#}}
desc: ''
title: Answer
type: answer
variables: []
id: answer
position:
x: 980
y: 282
sourcePosition: right
targetPosition: left
type: custom
api/tests/integration_tests/workflow/test_response_stream_filter_integration.py
+75
View File
@@ -0,0 +1,75 @@
"""Integration coverage for Dify's ResponseStreamFilter boundary behavior."""
from core.workflow.workflow_entry import iter_dify_graph_engine_events
from graphon.graph_engine import GraphEngine, GraphEngineConfig
from graphon.graph_engine.command_channels import InMemoryChannel
from graphon.graph_events import GraphRunSucceededEvent, NodeRunStreamChunkEvent
from tests.unit_tests.core.workflow.graph_engine.test_mock_config import MockConfigBuilder
from tests.unit_tests.core.workflow.graph_engine.test_table_runner import WorkflowRunner
def _build_issue_170_mock_config():
runner = WorkflowRunner()
mock_config = (
MockConfigBuilder()
.with_node_output(
"llm",
{
"text": "Quiet Night Thought",
"usage": {
"prompt_tokens": 10,
"completion_tokens": 5,
"total_tokens": 15,
},
"finish_reason": "stop",
},
)
.with_node_output(
"dufu",
{
"text": "Spring View",
"usage": {
"prompt_tokens": 10,
"completion_tokens": 5,
"total_tokens": 15,
},
"finish_reason": "stop",
},
)
.build()
)
return runner, mock_config
def test_dify_response_stream_filter_handles_issue_170_shape() -> None:
runner, mock_config = _build_issue_170_mock_config()
fixture_data = runner.load_fixture("response_stream_filter_issue_170_workflow")
graph, graph_runtime_state = runner.create_graph_from_fixture(
fixture_data=fixture_data,
query="1",
use_mock_factory=True,
mock_config=mock_config,
)
expected_answer = "# Du Fu\n\nSpring View\n\n# Li Bai\n\nQuiet Night Thought"
engine = GraphEngine(
workflow_id="test_workflow",
graph=graph,
graph_runtime_state=graph_runtime_state,
command_channel=InMemoryChannel(),
config=GraphEngineConfig(),
)
events = list(iter_dify_graph_engine_events(engine))
stream_chunk_events = [event for event in events if isinstance(event, NodeRunStreamChunkEvent)]
success_events = [event for event in events if isinstance(event, GraphRunSucceededEvent)]
assert success_events
assert stream_chunk_events
actual_answer = "".join(event.chunk for event in stream_chunk_events)
assert actual_answer.strip() == expected_answer
assert stream_chunk_events[-1].is_final is True
assert success_events[-1].outputs["answer"].strip() == expected_answer
assert actual_answer.strip() == success_events[-1].outputs["answer"].strip()
api/tests/unit_tests/core/workflow/graph_engine/test_parallel_human_input_join_resume.py
+9 -5
View File
@@ -9,7 +9,7 @@ from core.repositories.human_input_repository import (
HumanInputFormEntity,
HumanInputFormRepository,
)
from core.workflow.node_runtime import DifyFileReferenceFactory, DifyHumanInputNodeRuntime
from core.workflow.node_runtime import DifyHumanInputNodeRuntime
from core.workflow.system_variables import build_system_variables
from graphon.entities import WorkflowStartReason
from graphon.file import File, FileTransferMethod, FileType
@@ -186,25 +186,29 @@ def _build_graph(runtime_state: GraphRuntimeState, repo: HumanInputFormRepositor
)
human_a_config = {"id": "human_a", "data": human_data.model_dump()}
human_a_runtime = DifyHumanInputNodeRuntime(graph_init_params.run_context)
human_a_runtime._file_reference_factory = _TestFileReferenceFactory() # type: ignore[attr-defined]
human_a = HumanInputNode(
node_id=human_a_config["id"],
data=human_data,
graph_init_params=graph_init_params,
graph_runtime_state=runtime_state,
form_repository=repo,
file_reference_factory=DifyFileReferenceFactory(graph_init_params.run_context),
runtime=DifyHumanInputNodeRuntime(graph_init_params.run_context),
file_reference_factory=_TestFileReferenceFactory(),
runtime=human_a_runtime,
)
human_b_config = {"id": "human_b", "data": human_data.model_dump()}
human_b_runtime = DifyHumanInputNodeRuntime(graph_init_params.run_context)
human_b_runtime._file_reference_factory = _TestFileReferenceFactory() # type: ignore[attr-defined]
human_b = HumanInputNode(
node_id=human_b_config["id"],
data=human_data,
graph_init_params=graph_init_params,
graph_runtime_state=runtime_state,
form_repository=repo,
file_reference_factory=DifyFileReferenceFactory(graph_init_params.run_context),
runtime=DifyHumanInputNodeRuntime(graph_init_params.run_context),
file_reference_factory=_TestFileReferenceFactory(),
runtime=human_b_runtime,
)
end_data = EndNodeData(
api/tests/unit_tests/core/workflow/graph_engine/test_table_runner.py
+2 -1
View File
@@ -24,6 +24,7 @@ from core.tools.utils.yaml_utils import _load_yaml_file
from core.workflow.node_factory import DifyNodeFactory, get_default_root_node_id
from core.workflow.system_variables import build_bootstrap_variables, build_system_variables
from core.workflow.variable_pool_initializer import add_node_inputs_to_pool, add_variables_to_pool
from core.workflow.workflow_entry import iter_dify_graph_engine_events
from graphon.entities import GraphInitParams
from graphon.graph import Graph
from graphon.graph_engine import GraphEngine, GraphEngineConfig
@@ -386,7 +387,7 @@ class TableTestRunner:
# Execute and collect events
events: list[GraphEngineEvent] = []
for event in engine.run():
for event in iter_dify_graph_engine_events(engine):
events.append(event)
# Check execution success
api/tests/unit_tests/core/workflow/graph_engine/test_tool_in_chatflow.py
+7 -9
View File
@@ -1,3 +1,4 @@
from core.workflow.workflow_entry import iter_dify_graph_engine_events
from graphon.graph_engine import GraphEngine, GraphEngineConfig
from graphon.graph_engine.command_channels import InMemoryChannel
from graphon.graph_events import (
@@ -31,20 +32,17 @@ def test_tool_in_chatflow():
config=GraphEngineConfig(),
)
events = list(engine.run())
events = list(iter_dify_graph_engine_events(engine))
# Check for successful completion
success_events = [e for e in events if isinstance(e, GraphRunSucceededEvent)]
assert len(success_events) > 0, "Workflow should complete successfully"
# Check for streaming events
stream_chunk_events = [e for e in events if isinstance(e, NodeRunStreamChunkEvent)]
stream_chunk_count = len(stream_chunk_events)
assert stream_chunk_count == 1, f"Expected 1 streaming events, but got {stream_chunk_count}"
assert stream_chunk_events[0].chunk == "hello, dify!", (
f"Expected chunk to be 'hello, dify!', but got {stream_chunk_events[0].chunk}"
)
assert len(stream_chunk_events) > 0
assert "".join(event.chunk for event in stream_chunk_events) == "hello, dify!"
assert stream_chunk_events[-1].is_final is True
assert success_events[-1].outputs["answer"] == "hello, dify!"
def test_answer_can_render_llm_structured_output_in_chatflow():
@@ -88,7 +86,7 @@ def test_answer_can_render_llm_structured_output_in_chatflow():
config=GraphEngineConfig(),
)
events = list(engine.run())
events = list(iter_dify_graph_engine_events(engine))
success_events = [e for e in events if isinstance(e, GraphRunSucceededEvent)]
assert success_events, "Workflow should complete successfully"
api/tests/unit_tests/core/workflow/nodes/human_input/test_entities.py
+3 -2
View File
@@ -30,7 +30,7 @@ from core.workflow.human_input_adapter import (
WebAppDeliveryMethod,
_WebAppDeliveryConfig,
)
from core.workflow.node_runtime import DifyFileReferenceFactory, DifyHumanInputNodeRuntime
from core.workflow.node_runtime import DifyHumanInputNodeRuntime
from core.workflow.system_variables import build_system_variables
from graphon.entities import GraphInitParams
from graphon.file import File, FileTransferMethod, FileType
@@ -171,12 +171,13 @@ def _build_human_input_node(
typed_node_data = (
node_data if isinstance(node_data, HumanInputNodeData) else HumanInputNodeData.model_validate(node_data)
)
runtime._file_reference_factory = _TestFileReferenceFactory() # type: ignore[attr-defined]
return HumanInputNode(
node_id=node_id,
data=typed_node_data,
graph_init_params=graph_init_params,
graph_runtime_state=graph_runtime_state,
file_reference_factory=DifyFileReferenceFactory(graph_init_params.run_context),
file_reference_factory=_TestFileReferenceFactory(),
runtime=runtime,
)
api/tests/unit_tests/core/workflow/nodes/human_input/test_human_input_form_filled_event.py
+5 -3
View File
@@ -4,7 +4,7 @@ from types import SimpleNamespace
from typing import Any
from core.app.entities.app_invoke_entities import DIFY_RUN_CONTEXT_KEY, InvokeFrom, UserFrom
from core.workflow.node_runtime import DifyFileReferenceFactory, DifyHumanInputNodeRuntime
from core.workflow.node_runtime import DifyHumanInputNodeRuntime
from core.workflow.system_variables import default_system_variables
from graphon.entities import GraphInitParams
from graphon.enums import BuiltinNodeTypes
@@ -67,14 +67,16 @@ def _create_human_input_node(
if isinstance(config["data"], HumanInputNodeData)
else HumanInputNodeData.model_validate(config["data"])
)
runtime = DifyHumanInputNodeRuntime(graph_init_params.run_context)
runtime._file_reference_factory = _TestFileReferenceFactory() # type: ignore[attr-defined]
return HumanInputNode(
node_id=config["id"],
data=node_data,
graph_init_params=graph_init_params,
graph_runtime_state=graph_runtime_state,
form_repository=repo,
file_reference_factory=DifyFileReferenceFactory(graph_init_params.run_context),
runtime=DifyHumanInputNodeRuntime(graph_init_params.run_context),
file_reference_factory=_TestFileReferenceFactory(),
runtime=runtime,
)
api/tests/unit_tests/core/workflow/nodes/llm/test_node.py
+15 -9
View File
@@ -76,6 +76,7 @@ from graphon.nodes.llm.node import (
_render_jinja2_message,
)
from graphon.nodes.llm.protocols import CredentialsProvider, ModelFactory
from graphon.nodes.llm.reasoning import split_reasoning
from graphon.nodes.llm.runtime_protocols import PromptMessageSerializerProtocol
from graphon.runtime import GraphRuntimeState, VariablePool
from graphon.template_rendering import TemplateRenderError
@@ -1271,7 +1272,10 @@ class TestLLMNodeSaveMultiModalImageOutput:
assert llm_node._file_outputs == [mock_file]
assert file == mock_file
mock_file_saver.save_binary_string.assert_called_once_with(
data=b"test-data", mime_type="image/png", file_type=FileType.IMAGE
data=b"test-data",
mime_type="image/png",
file_type=FileType.IMAGE,
extension_override=".png",
)
def test_llm_node_save_url_output(self, llm_node_for_multimodal: tuple[LLMNode, LLMFileSaver]):
@@ -1305,8 +1309,9 @@ class TestLLMNodeSaveMultiModalImageOutput:
def test_llm_node_image_file_to_markdown(llm_node: LLMNode):
mock_file = mock.MagicMock(spec=File)
mock_file.type = FileType.IMAGE
mock_file.generate_url.return_value = "https://example.com/image.png"
markdown = llm_node._image_file_to_markdown(mock_file)
markdown = llm_node._saved_file_to_markdown(mock_file)
assert markdown == "![](https://example.com/image.png)"
@@ -1378,6 +1383,7 @@ class TestSaveMultimodalOutputAndConvertResultToMarkdown:
data=image_raw_data,
mime_type="image/png",
file_type=FileType.IMAGE,
extension_override=".png",
)
assert mock_saved_file in llm_node._file_outputs
@@ -1425,7 +1431,7 @@ class TestReasoningFormat:
</think>Dify is an open source AI platform.
"""
clean_text, reasoning_content = LLMNode._split_reasoning(text_with_think, "separated")
clean_text, reasoning_content = split_reasoning(text_with_think, "separated")
assert clean_text == "Dify is an open source AI platform."
assert reasoning_content == "I need to explain what Dify is. It's an open source AI platform."
@@ -1438,7 +1444,7 @@ class TestReasoningFormat:
</think>Dify is an open source AI platform.
"""
clean_text, reasoning_content = LLMNode._split_reasoning(text_with_think, "tagged")
clean_text, reasoning_content = split_reasoning(text_with_think, "tagged")
# Original text unchanged
assert clean_text == text_with_think
@@ -1450,7 +1456,7 @@ class TestReasoningFormat:
text_without_think = "This is a simple answer without any thinking blocks."
clean_text, reasoning_content = LLMNode._split_reasoning(text_without_think, "separated")
clean_text, reasoning_content = split_reasoning(text_without_think, "separated")
assert clean_text == text_without_think
assert reasoning_content == ""
@@ -1471,7 +1477,7 @@ class TestReasoningFormat:
<think>I need to explain what Dify is. It's an open source AI platform.
</think>Dify is an open source AI platform.
"""
clean_text, reasoning_content = LLMNode._split_reasoning(text_with_think, node_data.reasoning_format)
clean_text, reasoning_content = split_reasoning(text_with_think, node_data.reasoning_format)
assert clean_text == text_with_think
assert reasoning_content == ""
@@ -1569,10 +1575,10 @@ def test_handle_invoke_result_streaming_collects_text_metrics_and_structured_out
)
assert events[0] == first_chunk
assert events[1] == StreamChunkEvent(selector=["node-1", "text"], chunk="<think>plan</think>", is_final=False)
assert events[2] == StreamChunkEvent(selector=["node-1", "text"], chunk="answer", is_final=False)
completed = events[3]
assert events[1] == StreamChunkEvent(selector=["node-1", "text"], chunk="answer", is_final=False)
completed = events[2]
assert isinstance(completed, ModelInvokeCompletedEvent)
assert completed.text == "answer"
assert completed.reasoning_content == "plan"
api/tests/unit_tests/core/workflow/test_workflow_entry_helpers.py
+46
View File
@@ -338,6 +338,52 @@ class TestWorkflowEntryRun:
assert list(entry.run()) == []
def test_iter_dify_graph_engine_events_applies_response_stream_filter(self):
graph_engine = MagicMock()
graph_engine.run.return_value = iter([sentinel.raw_event])
with (
patch.object(
workflow_entry.GraphEventFilterContext,
"from_engine",
return_value=sentinel.filter_context,
) as from_engine,
patch.object(
workflow_entry,
"ResponseStreamFilter",
return_value=sentinel.response_stream_filter,
) as response_stream_filter_cls,
patch.object(
workflow_entry,
"filter_graph_events",
return_value=iter([sentinel.filtered_event]),
) as filter_graph_events,
):
events = list(workflow_entry.iter_dify_graph_engine_events(graph_engine))
assert events == [sentinel.filtered_event]
from_engine.assert_called_once_with(graph_engine)
response_stream_filter_cls.assert_called_once_with()
filter_graph_events.assert_called_once_with(
graph_engine.run.return_value,
context=sentinel.filter_context,
filters=[sentinel.response_stream_filter],
)
def test_run_delegates_to_dify_event_iterator(self):
entry = object.__new__(workflow_entry.WorkflowEntry)
entry.graph_engine = sentinel.graph_engine
with patch.object(
workflow_entry,
"iter_dify_graph_engine_events",
return_value=iter([sentinel.filtered_event]),
) as iter_dify_graph_engine_events:
events = list(entry.run())
assert events == [sentinel.filtered_event]
iter_dify_graph_engine_events.assert_called_once_with(sentinel.graph_engine)
def test_run_emits_failed_event_for_unexpected_errors(self):
entry = object.__new__(workflow_entry.WorkflowEntry)
entry.graph_engine = MagicMock()
api/tests/unit_tests/services/test_model_load_balancing_service.py
+23
View File
@@ -112,6 +112,29 @@ def test_enable_disable_model_load_balancing_should_call_provider_configuration_
)
@pytest.mark.parametrize(
("method_name", "expected_provider_method"),
[
("enable_model_load_balancing", "enable_model_load_balancing"),
("disable_model_load_balancing", "disable_model_load_balancing"),
],
)
def test_enable_disable_model_load_balancing_uses_model_type_constructor_directly(
method_name: str,
expected_provider_method: str,
service: ModelLoadBalancingService,
monkeypatch: pytest.MonkeyPatch,
) -> None:
provider_configuration = _build_provider_configuration(provider_schema=_build_provider_credential_schema())
service.provider_manager.get_configurations.return_value = {"openai": provider_configuration}
getattr(service, method_name)("tenant-1", "openai", "gpt-4o-mini", "text-generation")
getattr(provider_configuration, expected_provider_method).assert_called_once_with(
model="gpt-4o-mini", model_type=ModelType.LLM
)
@pytest.mark.parametrize(
"method_name",
["enable_model_load_balancing", "disable_model_load_balancing"],
api/tests/unit_tests/services/test_model_provider_service.py
+64
View File
@@ -368,6 +368,70 @@ class TestModelProviderServiceDelegation:
if method_name == "get_model_credential":
assert result == {"api_key": "x"}
@pytest.mark.parametrize(
("method_name", "method_kwargs", "provider_method_name", "expected_kwargs"),
[
(
"get_model_credential",
{
"tenant_id": "tenant-1",
"provider": "openai",
"model_type": "text-generation",
"model": "gpt-4o",
"credential_id": "cred-1",
},
"get_custom_model_credential",
{"model_type": ModelType.LLM, "model": "gpt-4o", "credential_id": "cred-1"},
),
(
"create_model_credential",
{
"tenant_id": "tenant-1",
"provider": "openai",
"model_type": "text-generation",
"model": "gpt-4o",
"credentials": {"api_key": "x"},
"credential_name": "cred-a",
},
"create_custom_model_credential",
{
"model_type": ModelType.LLM,
"model": "gpt-4o",
"credentials": {"api_key": "x"},
"credential_name": "cred-a",
},
),
(
"remove_model",
{
"tenant_id": "tenant-1",
"provider": "openai",
"model_type": "text-generation",
"model": "gpt-4o",
},
"delete_custom_model",
{"model_type": ModelType.LLM, "model": "gpt-4o"},
),
],
)
def test_custom_model_methods_use_model_type_constructor_directly(
self,
method_name: str,
method_kwargs: dict[str, Any],
provider_method_name: str,
expected_kwargs: dict[str, Any],
monkeypatch: pytest.MonkeyPatch,
) -> None:
service = ModelProviderService()
provider_configuration = MagicMock()
get_provider_config_mock = MagicMock(return_value=provider_configuration)
monkeypatch.setattr(service, "_get_provider_configuration", get_provider_config_mock)
getattr(service, method_name)(**method_kwargs)
get_provider_config_mock.assert_called_once_with("tenant-1", "openai")
getattr(provider_configuration, provider_method_name).assert_called_once_with(**expected_kwargs)
class TestModelProviderServiceListingsAndDefaults:
def test_get_models_by_model_type_should_group_active_non_deprecated_models(self) -> None: