vellum-ai 0.14.86__tar.gz → 0.14.88__tar.gz

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vellum-ai
-Version: 0.14.86
+Version: 0.14.88
 Summary: 
 License: MIT
 Requires-Python: >=3.9,<4.0

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/pyproject.toml

@@ -3,7 +3,7 @@ name = "vellum-ai"
 
 [tool.poetry]
 name = "vellum-ai"
-version = "0.14.86"
+version = "0.14.88"
 description = ""
 readme = "README.md"
 authors = []

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/client/core/client_wrapper.py

@@ -16,10 +16,10 @@ class BaseClientWrapper:
 
     def get_headers(self) -> typing.Dict[str, str]:
         headers: typing.Dict[str, str] = {
-            "User-Agent": "vellum-ai/0.14.86",
+            "User-Agent": "vellum-ai/0.14.88",
             "X-Fern-Language": "Python",
             "X-Fern-SDK-Name": "vellum-ai",
-            "X-Fern-SDK-Version": "0.14.86",
+            "X-Fern-SDK-Version": "0.14.88",
         }
         headers["X-API-KEY"] = self.api_key
         return headers

vellum_ai-0.14.88/src/vellum/client/types/api_version_enum.py

@@ -0,0 +1,5 @@
+# This file was auto-generated by Fern from our API Definition.
+
+import typing
+
+ApiVersionEnum = typing.Union[typing.Literal["2024-10-25", "2025-07-30"], typing.Any]

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/client/types/execution_thinking_vellum_value.py

@@ -19,7 +19,7 @@ class ExecutionThinkingVellumValue(UniversalBaseModel):
 
     name: str
     type: typing.Literal["THINKING"] = "THINKING"
-    value: typing.List[StringVellumValue]
+    value: StringVellumValue
 
     if IS_PYDANTIC_V2:
         model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow", frozen=True)  # type: ignore # Pydantic v2

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/client/types/thinking_vellum_value.py

@@ -13,7 +13,7 @@ class ThinkingVellumValue(UniversalBaseModel):
     """
 
     type: typing.Literal["THINKING"] = "THINKING"
-    value: typing.List[StringVellumValue]
+    value: StringVellumValue
 
     if IS_PYDANTIC_V2:
         model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow", frozen=True)  # type: ignore # Pydantic v2

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/client/types/thinking_vellum_value_request.py

@@ -13,7 +13,7 @@ class ThinkingVellumValueRequest(UniversalBaseModel):
     """
 
     type: typing.Literal["THINKING"] = "THINKING"
-    value: typing.List[StringVellumValueRequest]
+    value: StringVellumValueRequest
 
     if IS_PYDANTIC_V2:
         model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow", frozen=True)  # type: ignore # Pydantic v2

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/client/types/workflow_event_execution_read.py

@@ -33,6 +33,7 @@ class WorkflowEventExecutionRead(UniversalBaseModel):
     metric_results: typing.List[WorkflowExecutionViewOnlineEvalMetricResult]
     usage_results: typing.Optional[typing.List[WorkflowExecutionUsageResult]] = None
     spans: typing.List[VellumSpan]
+    state: typing.Optional[typing.Dict[str, typing.Optional[typing.Any]]] = None
 
     if IS_PYDANTIC_V2:
         model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow", frozen=True)  # type: ignore # Pydantic v2

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/workflows/descriptors/utils.py

@@ -103,7 +103,7 @@ def is_unresolved(value: Any) -> bool:
         return any(is_unresolved(getattr(value, field.name)) for field in dataclasses.fields(value))
 
     if isinstance(value, BaseModel):
-        return any(is_unresolved(getattr(value, key)) for key in value.model_fields.keys())
+        return any(is_unresolved(getattr(value, key)) for key in value.__class__.model_fields.keys())
 
     if isinstance(value, Mapping):
         return any(is_unresolved(item) for item in value.values())

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/workflows/nodes/displayable/inline_prompt_node/node.py

@@ -64,9 +64,8 @@ class InlinePromptNode(BaseInlinePromptNode[StateType]):
             elif output.type == "FUNCTION_CALL":
                 string_outputs.append(output.value.model_dump_json(indent=4))
             elif output.type == "THINKING":
-                for thinking_item in output.value:
-                    if thinking_item.type == "STRING":
-                        string_outputs.append(thinking_item.value)
+                if output.value.type == "STRING":
+                    string_outputs.append(output.value.value)
             else:
                 string_outputs.append(output.value.message)
 

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/workflows/nodes/displayable/prompt_deployment_node/node.py

@@ -66,9 +66,8 @@ class PromptDeploymentNode(BasePromptDeploymentNode[StateType]):
             elif output.type == "FUNCTION_CALL":
                 string_outputs.append(output.value.model_dump_json(indent=4))
             elif output.type == "THINKING":
-                for thinking_item in output.value:
-                    if thinking_item.type == "STRING":
-                        string_outputs.append(thinking_item.value)
+                if output.value.type == "STRING":
+                    string_outputs.append(output.value.value)
             else:
                 string_outputs.append(output.value.message)
 

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/workflows/nodes/displayable/tool_calling_node/tests/test_node.py

@@ -1,13 +1,19 @@
 import json
-from typing import Any, List
+from uuid import uuid4
+from typing import Any, Iterator, List
 
 from vellum import ChatMessage
+from vellum.client.types.fulfilled_execute_prompt_event import FulfilledExecutePromptEvent
 from vellum.client.types.function_call import FunctionCall
 from vellum.client.types.function_call_vellum_value import FunctionCallVellumValue
+from vellum.client.types.initiated_execute_prompt_event import InitiatedExecutePromptEvent
 from vellum.client.types.string_chat_message_content import StringChatMessageContent
+from vellum.client.types.string_vellum_value import StringVellumValue
+from vellum.client.types.variable_prompt_block import VariablePromptBlock
 from vellum.workflows import BaseWorkflow
 from vellum.workflows.inputs.base import BaseInputs
 from vellum.workflows.nodes.bases import BaseNode
+from vellum.workflows.nodes.displayable.tool_calling_node.node import ToolCallingNode
 from vellum.workflows.nodes.displayable.tool_calling_node.utils import create_function_node, create_tool_router_node
 from vellum.workflows.outputs.base import BaseOutputs
 from vellum.workflows.state.base import BaseState, StateMeta
@@ -140,3 +146,65 @@ def test_deployment_definition_release_tag_defaults_to_latest():
 
     # THEN the release_tag should default to "LATEST"
     assert deployment_config.release_tag == "LATEST"
+
+
+def test_tool_calling_node_with_user_provided_chat_history_block(vellum_adhoc_prompt_client):
+    """
+    Test that ToolCallingNode with user-provided chat history block merges user and node messages.
+    """
+
+    # GIVEN a ToolCallingNode with a user-provided chat history block
+    user_chat_history_block = VariablePromptBlock(
+        block_type="VARIABLE",
+        input_variable="chat_history",
+        state=None,
+        cache_config=None,
+    )
+
+    class TestToolCallingNode(ToolCallingNode):
+        ml_model = "gpt-4o-mini"
+        blocks = [user_chat_history_block]
+        functions = [first_function]
+        prompt_inputs = {"chat_history": [ChatMessage(role="USER", text="Hello from user")]}
+        max_prompt_iterations = 1
+
+    def generate_prompt_events(*args: Any, **kwargs: Any) -> Iterator[Any]:
+        execution_id = str(uuid4())
+        events = [
+            InitiatedExecutePromptEvent(execution_id=execution_id),
+            FulfilledExecutePromptEvent(
+                execution_id=execution_id,
+                outputs=[StringVellumValue(value="Hello! I can help you.")],
+            ),
+        ]
+        yield from events
+
+    vellum_adhoc_prompt_client.adhoc_execute_prompt_stream.side_effect = generate_prompt_events
+
+    # AND a state
+    state = BaseState()
+
+    # WHEN the ToolCallingNode runs
+    node = TestToolCallingNode(state=state)
+    list(node.run())
+
+    # THEN the API should be called with the correct blocks
+    mock_api = vellum_adhoc_prompt_client.adhoc_execute_prompt_stream
+    assert mock_api.call_count >= 1
+
+    # AND the blocks should include the user-provided chat_history block
+    call_kwargs = mock_api.call_args.kwargs
+    blocks = call_kwargs["blocks"]
+
+    chat_history_blocks = [
+        block for block in blocks if block.block_type == "VARIABLE" and block.input_variable == "chat_history"
+    ]
+    assert len(chat_history_blocks) == 1
+
+    # AND the input_values should include the user's chat history
+    input_values = call_kwargs["input_values"]
+    chat_history_inputs = [
+        input_val for input_val in input_values if hasattr(input_val, "key") and input_val.key == "chat_history"
+    ]
+    assert len(chat_history_inputs) == 1
+    assert chat_history_inputs[0].value == [ChatMessage(role="USER", text="Hello from user")]

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/workflows/nodes/displayable/tool_calling_node/utils.py

@@ -8,12 +8,14 @@ from vellum.client.types.function_call_chat_message_content import FunctionCallC
 from vellum.client.types.function_call_chat_message_content_value import FunctionCallChatMessageContentValue
 from vellum.client.types.string_chat_message_content import StringChatMessageContent
 from vellum.client.types.variable_prompt_block import VariablePromptBlock
+from vellum.workflows.context import execution_context, get_parent_context
 from vellum.workflows.errors.types import WorkflowErrorCode
+from vellum.workflows.events.workflow import is_workflow_event
 from vellum.workflows.exceptions import NodeException
 from vellum.workflows.nodes.bases import BaseNode
 from vellum.workflows.nodes.displayable.inline_prompt_node.node import InlinePromptNode
 from vellum.workflows.nodes.displayable.subworkflow_deployment_node.node import SubworkflowDeploymentNode
-from vellum.workflows.outputs.base import BaseOutput
+from vellum.workflows.outputs.base import BaseOutput, BaseOutputs
 from vellum.workflows.ports.port import Port
 from vellum.workflows.references.lazy import LazyReference
 from vellum.workflows.state.context import WorkflowContext
@@ -21,6 +23,9 @@ from vellum.workflows.state.encoder import DefaultStateEncoder
 from vellum.workflows.types.core import EntityInputsInterface, MergeBehavior, Tool
 from vellum.workflows.types.definition import DeploymentDefinition
 from vellum.workflows.types.generics import is_workflow_class
+from vellum.workflows.workflows.event_filters import all_workflow_event_filter
+
+CHAT_HISTORY_VARIABLE = "chat_history"
 
 
 class FunctionNode(BaseNode):
@@ -40,7 +45,10 @@ class ToolRouterNode(InlinePromptNode):
             max_iterations_message = f"Maximum number of prompt iterations `{self.max_prompt_iterations}` reached."
             raise NodeException(message=max_iterations_message, code=WorkflowErrorCode.NODE_EXECUTION)
 
-        self.prompt_inputs = {**self.prompt_inputs, "chat_history": self.state.chat_history}  # type: ignore
+        # Merge user-provided chat history with node's chat history
+        user_chat_history = self.prompt_inputs.get(CHAT_HISTORY_VARIABLE, []) if self.prompt_inputs else []
+        merged_chat_history = user_chat_history + self.state.chat_history
+        self.prompt_inputs = {**self.prompt_inputs, CHAT_HISTORY_VARIABLE: merged_chat_history}  # type: ignore
         generator = super().run()
         for output in generator:
             if output.name == "results" and output.value:
@@ -102,16 +110,21 @@ def create_tool_router_node(
         # If no functions exist, create a simple Ports class with just a default port
         Ports = type("Ports", (), {"default": Port(default=True)})
 
-    # Add a chat history block to blocks
-    blocks.append(
-        VariablePromptBlock(
-            block_type="VARIABLE",
-            input_variable="chat_history",
-            state=None,
-            cache_config=None,
-        )
+    # Add a chat history block to blocks only if one doesn't already exist
+    has_chat_history_block = any(
+        block.block_type == "VARIABLE" and block.input_variable == CHAT_HISTORY_VARIABLE for block in blocks
     )
 
+    if not has_chat_history_block:
+        blocks.append(
+            VariablePromptBlock(
+                block_type="VARIABLE",
+                input_variable=CHAT_HISTORY_VARIABLE,
+                state=None,
+                cache_config=None,
+            )
+        )
+
     node = cast(
         Type[ToolRouterNode],
         type(
@@ -200,26 +213,60 @@ def create_function_node(
     elif is_workflow_class(function):
         # Create a class-level wrapper that calls the original function
         def execute_inline_workflow_function(self) -> BaseNode.Outputs:
-            outputs = self.state.meta.node_outputs.get(tool_router_node.Outputs.text)
-
-            outputs = json.loads(outputs)
-            arguments = outputs["arguments"]
+            function_call_output = self.state.meta.node_outputs.get(tool_router_node.Outputs.results)
+            if function_call_output and len(function_call_output) > 0:
+                function_call = function_call_output[0]
+                arguments = function_call.value.arguments
+            else:
+                arguments = {}
 
             # Call the function based on its type
             inputs_instance = function.get_inputs_class()(**arguments)
-            workflow = function(context=WorkflowContext.create_from(self._context))
-            terminal_event = workflow.run(
-                inputs=inputs_instance,
-            )
-            if terminal_event.name == "workflow.execution.paused":
+
+            with execution_context(parent_context=get_parent_context()):
+                workflow = function(
+                    parent_state=self.state,
+                    context=WorkflowContext.create_from(self._context),
+                )
+                subworkflow_stream = workflow.stream(
+                    inputs=inputs_instance,
+                    event_filter=all_workflow_event_filter,
+                    node_output_mocks=self._context._get_all_node_output_mocks(),
+                )
+
+            outputs: Optional[BaseOutputs] = None
+            exception: Optional[NodeException] = None
+
+            for event in subworkflow_stream:
+                self._context._emit_subworkflow_event(event)
+                if exception:
+                    continue
+
+                if not is_workflow_event(event):
+                    continue
+                if event.workflow_definition != function:
+                    continue
+
+                if event.name == "workflow.execution.fulfilled":
+                    outputs = event.outputs
+                elif event.name == "workflow.execution.rejected":
+                    exception = NodeException.of(event.error)
+                elif event.name == "workflow.execution.paused":
+                    exception = NodeException(
+                        code=WorkflowErrorCode.INVALID_OUTPUTS,
+                        message="Subworkflow unexpectedly paused",
+                    )
+
+            if exception:
+                raise exception
+
+            if outputs is None:
                 raise NodeException(
+                    message="Expected to receive outputs from inline subworkflow",
                     code=WorkflowErrorCode.INVALID_OUTPUTS,
-                    message="Subworkflow unexpectedly paused",
                 )
-            elif terminal_event.name == "workflow.execution.fulfilled":
-                result = terminal_event.outputs
-            elif terminal_event.name == "workflow.execution.rejected":
-                raise NodeException(message=terminal_event.error.message, code=terminal_event.error.code)
+
+            result = outputs
 
             self.state.chat_history.append(
                 ChatMessage(

{vellum_ai-0.14.86 → vellum_ai-0.14.88}/src/vellum/workflows/utils/functions.py

@@ -82,7 +82,7 @@ def _compile_default_value(default: Any) -> Any:
     if isinstance(default, BaseModel):
         return {
             field_name: _compile_default_value(getattr(default, field_name))
-            for field_name in default.model_fields.keys()
+            for field_name in default.__class__.model_fields.keys()
         }
 
     return default

vellum_ai-0.14.86/src/vellum/client/types/api_version_enum.py

@@ -1,5 +0,0 @@
-# This file was auto-generated by Fern from our API Definition.
-
-import typing
-
-ApiVersionEnum = typing.Literal["2024-10-25"]