vellum-ai 0.14.42__py3-none-any.whl → 0.14.43__py3-none-any.whl

vellum/__init__.py

@@ -310,6 +310,7 @@ from .types import (
     PromptNodeResultData,
     PromptOutput,
     PromptParameters,
+    PromptPushResponse,
     PromptRequestChatHistoryInput,
     PromptRequestInput,
     PromptRequestJsonInput,
@@ -939,6 +940,7 @@ __all__ = [
     "PromptNodeResultData",
     "PromptOutput",
     "PromptParameters",
+    "PromptPushResponse",
     "PromptRequestChatHistoryInput",
     "PromptRequestInput",
     "PromptRequestJsonInput",

vellum/client/core/client_wrapper.py

@@ -18,9 +18,9 @@ class BaseClientWrapper:
         headers: typing.Dict[str, str] = {
             "X-Fern-Language": "Python",
             "X-Fern-SDK-Name": "vellum-ai",
-            "X-Fern-SDK-Version": "0.14.42",
+            "X-Fern-SDK-Version": "0.14.43",
         }
-        headers["X_API_KEY"] = self.api_key
+        headers["X-API-KEY"] = self.api_key
         return headers
 
     def get_environment(self) -> VellumEnvironment:

vellum/client/reference.md

@@ -4266,6 +4266,113 @@ client.prompts.pull(
 </dl>
 
 
+</dd>
+</dl>
+</details>
+
+<details><summary><code>client.prompts.<a href="src/vellum/resources/prompts/client.py">push</a>(...)</code></summary>
+<dl>
+<dd>
+
+#### 📝 Description
+
+<dl>
+<dd>
+
+<dl>
+<dd>
+
+Used to push updates to a Prompt in Vellum.
+</dd>
+</dl>
+</dd>
+</dl>
+
+#### 🔌 Usage
+
+<dl>
+<dd>
+
+<dl>
+<dd>
+
+```python
+from vellum import (
+    JinjaPromptBlock,
+    PromptExecConfig,
+    PromptParameters,
+    Vellum,
+    VellumVariable,
+)
+
+client = Vellum(
+    api_key="YOUR_API_KEY",
+)
+client.prompts.push(
+    exec_config=PromptExecConfig(
+        ml_model="ml_model",
+        input_variables=[
+            VellumVariable(
+                id="id",
+                key="key",
+                type="STRING",
+            )
+        ],
+        parameters=PromptParameters(),
+        blocks=[
+            JinjaPromptBlock(
+                template="template",
+            )
+        ],
+    ),
+)
+
+```
+</dd>
+</dl>
+</dd>
+</dl>
+
+#### ⚙️ Parameters
+
+<dl>
+<dd>
+
+<dl>
+<dd>
+
+**exec_config:** `PromptExecConfig` 
+    
+</dd>
+</dl>
+
+<dl>
+<dd>
+
+**prompt_variant_id:** `typing.Optional[str]` 
+    
+</dd>
+</dl>
+
+<dl>
+<dd>
+
+**prompt_sandbox_id:** `typing.Optional[str]` 
+    
+</dd>
+</dl>
+
+<dl>
+<dd>
+
+**request_options:** `typing.Optional[RequestOptions]` — Request-specific configuration.
+    
+</dd>
+</dl>
+</dd>
+</dl>
+
+
 </dd>
 </dl>
 </details>

vellum/client/resources/prompts/client.py

@@ -1,7 +1,7 @@
 # This file was auto-generated by Fern from our API Definition.
 
-from ...core.client_wrapper import SyncClientWrapper
 import typing
+from ...core.client_wrapper import SyncClientWrapper
 from ...core.request_options import RequestOptions
 from ...types.prompt_exec_config import PromptExecConfig
 from ...core.jsonable_encoder import jsonable_encoder
@@ -10,8 +10,13 @@ from ...errors.bad_request_error import BadRequestError
 from ...errors.not_found_error import NotFoundError
 from json.decoder import JSONDecodeError
 from ...core.api_error import ApiError
+from ...types.prompt_push_response import PromptPushResponse
+from ...core.serialization import convert_and_respect_annotation_metadata
 from ...core.client_wrapper import AsyncClientWrapper
 
+# this is used as the default value for optional parameters
+OMIT = typing.cast(typing.Any, ...)
+
 
 class PromptsClient:
     def __init__(self, *, client_wrapper: SyncClientWrapper):
@@ -100,6 +105,113 @@ class PromptsClient:
             raise ApiError(status_code=_response.status_code, body=_response.text)
         raise ApiError(status_code=_response.status_code, body=_response_json)
 
+    def push(
+        self,
+        *,
+        exec_config: PromptExecConfig,
+        prompt_variant_id: typing.Optional[str] = OMIT,
+        prompt_sandbox_id: typing.Optional[str] = OMIT,
+        request_options: typing.Optional[RequestOptions] = None,
+    ) -> PromptPushResponse:
+        """
+        Used to push updates to a Prompt in Vellum.
+
+        Parameters
+        ----------
+        exec_config : PromptExecConfig
+
+        prompt_variant_id : typing.Optional[str]
+
+        prompt_sandbox_id : typing.Optional[str]
+
+        request_options : typing.Optional[RequestOptions]
+            Request-specific configuration.
+
+        Returns
+        -------
+        PromptPushResponse
+
+
+        Examples
+        --------
+        from vellum import (
+            JinjaPromptBlock,
+            PromptExecConfig,
+            PromptParameters,
+            Vellum,
+            VellumVariable,
+        )
+
+        client = Vellum(
+            api_key="YOUR_API_KEY",
+        )
+        client.prompts.push(
+            exec_config=PromptExecConfig(
+                ml_model="ml_model",
+                input_variables=[
+                    VellumVariable(
+                        id="id",
+                        key="key",
+                        type="STRING",
+                    )
+                ],
+                parameters=PromptParameters(),
+                blocks=[
+                    JinjaPromptBlock(
+                        template="template",
+                    )
+                ],
+            ),
+        )
+        """
+        _response = self._client_wrapper.httpx_client.request(
+            "v1/prompts/push",
+            base_url=self._client_wrapper.get_environment().default,
+            method="POST",
+            json={
+                "exec_config": convert_and_respect_annotation_metadata(
+                    object_=exec_config, annotation=PromptExecConfig, direction="write"
+                ),
+                "prompt_variant_id": prompt_variant_id,
+                "prompt_sandbox_id": prompt_sandbox_id,
+            },
+            request_options=request_options,
+            omit=OMIT,
+        )
+        try:
+            if 200 <= _response.status_code < 300:
+                return typing.cast(
+                    PromptPushResponse,
+                    parse_obj_as(
+                        type_=PromptPushResponse,  # type: ignore
+                        object_=_response.json(),
+                    ),
+                )
+            if _response.status_code == 400:
+                raise BadRequestError(
+                    typing.cast(
+                        typing.Optional[typing.Any],
+                        parse_obj_as(
+                            type_=typing.Optional[typing.Any],  # type: ignore
+                            object_=_response.json(),
+                        ),
+                    )
+                )
+            if _response.status_code == 404:
+                raise NotFoundError(
+                    typing.cast(
+                        typing.Optional[typing.Any],
+                        parse_obj_as(
+                            type_=typing.Optional[typing.Any],  # type: ignore
+                            object_=_response.json(),
+                        ),
+                    )
+                )
+            _response_json = _response.json()
+        except JSONDecodeError:
+            raise ApiError(status_code=_response.status_code, body=_response.text)
+        raise ApiError(status_code=_response.status_code, body=_response_json)
+
 
 class AsyncPromptsClient:
     def __init__(self, *, client_wrapper: AsyncClientWrapper):
@@ -195,3 +307,118 @@ class AsyncPromptsClient:
         except JSONDecodeError:
             raise ApiError(status_code=_response.status_code, body=_response.text)
         raise ApiError(status_code=_response.status_code, body=_response_json)
+
+    async def push(
+        self,
+        *,
+        exec_config: PromptExecConfig,
+        prompt_variant_id: typing.Optional[str] = OMIT,
+        prompt_sandbox_id: typing.Optional[str] = OMIT,
+        request_options: typing.Optional[RequestOptions] = None,
+    ) -> PromptPushResponse:
+        """
+        Used to push updates to a Prompt in Vellum.
+
+        Parameters
+        ----------
+        exec_config : PromptExecConfig
+
+        prompt_variant_id : typing.Optional[str]
+
+        prompt_sandbox_id : typing.Optional[str]
+
+        request_options : typing.Optional[RequestOptions]
+            Request-specific configuration.
+
+        Returns
+        -------
+        PromptPushResponse
+
+
+        Examples
+        --------
+        import asyncio
+
+        from vellum import (
+            AsyncVellum,
+            JinjaPromptBlock,
+            PromptExecConfig,
+            PromptParameters,
+            VellumVariable,
+        )
+
+        client = AsyncVellum(
+            api_key="YOUR_API_KEY",
+        )
+
+
+        async def main() -> None:
+            await client.prompts.push(
+                exec_config=PromptExecConfig(
+                    ml_model="ml_model",
+                    input_variables=[
+                        VellumVariable(
+                            id="id",
+                            key="key",
+                            type="STRING",
+                        )
+                    ],
+                    parameters=PromptParameters(),
+                    blocks=[
+                        JinjaPromptBlock(
+                            template="template",
+                        )
+                    ],
+                ),
+            )
+
+
+        asyncio.run(main())
+        """
+        _response = await self._client_wrapper.httpx_client.request(
+            "v1/prompts/push",
+            base_url=self._client_wrapper.get_environment().default,
+            method="POST",
+            json={
+                "exec_config": convert_and_respect_annotation_metadata(
+                    object_=exec_config, annotation=PromptExecConfig, direction="write"
+                ),
+                "prompt_variant_id": prompt_variant_id,
+                "prompt_sandbox_id": prompt_sandbox_id,
+            },
+            request_options=request_options,
+            omit=OMIT,
+        )
+        try:
+            if 200 <= _response.status_code < 300:
+                return typing.cast(
+                    PromptPushResponse,
+                    parse_obj_as(
+                        type_=PromptPushResponse,  # type: ignore
+                        object_=_response.json(),
+                    ),
+                )
+            if _response.status_code == 400:
+                raise BadRequestError(
+                    typing.cast(
+                        typing.Optional[typing.Any],
+                        parse_obj_as(
+                            type_=typing.Optional[typing.Any],  # type: ignore
+                            object_=_response.json(),
+                        ),
+                    )
+                )
+            if _response.status_code == 404:
+                raise NotFoundError(
+                    typing.cast(
+                        typing.Optional[typing.Any],
+                        parse_obj_as(
+                            type_=typing.Optional[typing.Any],  # type: ignore
+                            object_=_response.json(),
+                        ),
+                    )
+                )
+            _response_json = _response.json()
+        except JSONDecodeError:
+            raise ApiError(status_code=_response.status_code, body=_response.text)
+        raise ApiError(status_code=_response.status_code, body=_response_json)

vellum/client/types/__init__.py

@@ -318,6 +318,7 @@ from .prompt_node_result import PromptNodeResult
 from .prompt_node_result_data import PromptNodeResultData
 from .prompt_output import PromptOutput
 from .prompt_parameters import PromptParameters
+from .prompt_push_response import PromptPushResponse
 from .prompt_request_chat_history_input import PromptRequestChatHistoryInput
 from .prompt_request_input import PromptRequestInput
 from .prompt_request_json_input import PromptRequestJsonInput
@@ -919,6 +920,7 @@ __all__ = [
     "PromptNodeResultData",
     "PromptOutput",
     "PromptParameters",
+    "PromptPushResponse",
     "PromptRequestChatHistoryInput",
     "PromptRequestInput",
     "PromptRequestJsonInput",

vellum/client/types/deployment_read.py

@@ -50,9 +50,9 @@ class DeploymentRead(UniversalBaseModel):
     A human-readable description of the deployment
     """
 
-    active_model_version_ids: typing.List[str] = pydantic.Field()
+    active_model_version_ids: typing.Optional[typing.List[str]] = pydantic.Field(default=None)
     """
-    Deprecated. The Prompt execution endpoints return a `prompt_version_id` that could be used instead.
+    Deprecated. This now always returns a null value.
     """
 
     last_deployed_history_item_id: str = pydantic.Field()

vellum/client/types/prompt_push_response.py

@@ -0,0 +1,20 @@
+# This file was auto-generated by Fern from our API Definition.
+
+from ..core.pydantic_utilities import UniversalBaseModel
+from ..core.pydantic_utilities import IS_PYDANTIC_V2
+import typing
+import pydantic
+
+
+class PromptPushResponse(UniversalBaseModel):
+    prompt_variant_id: str
+    prompt_sandbox_id: str
+
+    if IS_PYDANTIC_V2:
+        model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(extra="allow", frozen=True)  # type: ignore # Pydantic v2
+    else:
+
+        class Config:
+            frozen = True
+            smart_union = True
+            extra = pydantic.Extra.allow

vellum/types/prompt_push_response.py

@@ -0,0 +1,3 @@
+# WARNING: This file will be removed in a future release. Please import from "vellum.client" instead.
+
+from vellum.client.types.prompt_push_response import *

vellum/workflows/nodes/bases/base.py

@@ -41,9 +41,6 @@ def is_nested_class(nested: Any, parent: Type) -> bool:
 
 class BaseNodeMeta(type):
     def __new__(mcs, name: str, bases: Tuple[Type, ...], dct: Dict[str, Any]) -> Any:
-        # TODO: Inherit the inner Output classes from every base class.
-        #   https://app.shortcut.com/vellum/story/4007/support-auto-inheriting-parent-node-outputs
-
         if "Outputs" in dct:
             outputs_class = dct["Outputs"]
             if not any(issubclass(base, BaseOutputs) for base in outputs_class.__bases__):

vellum/workflows/nodes/bases/tests/test_base_adornment_node.py

@@ -0,0 +1,31 @@
+from typing import Callable, Type
+
+from vellum.workflows.nodes.bases.base import BaseNode
+from vellum.workflows.nodes.bases.base_adornment_node import BaseAdornmentNode
+from vellum.workflows.nodes.utils import create_adornment
+
+
+def test_base_adornment_node__output_references_of_same_name():
+    # GIVEN a custom adornment node
+    class CustomAdornmentNode(BaseAdornmentNode):
+        @classmethod
+        def wrap(cls) -> Callable[..., Type["CustomAdornmentNode"]]:
+            return create_adornment(cls)
+
+    # AND two nodes wrapped by the adornment with the same output
+    @CustomAdornmentNode.wrap()
+    class AppleNode(BaseNode):
+        class Outputs(BaseNode.Outputs):
+            fruit: str
+
+    @CustomAdornmentNode.wrap()
+    class OrangeNode(BaseNode):
+        class Outputs(BaseNode.Outputs):
+            fruit: str
+
+    # WHEN get output references of these outputs
+    apple_output_reference = AppleNode.Outputs.fruit
+    orange_output_reference = OrangeNode.Outputs.fruit
+
+    # THEN the output references should not be equal
+    assert apple_output_reference != orange_output_reference, "Output references should not be equal"

vellum/workflows/nodes/displayable/bases/inline_prompt_node/node.py

@@ -1,6 +1,6 @@
 import json
 from uuid import uuid4
-from typing import Callable, ClassVar, Generic, Iterator, List, Optional, Set, Tuple, Union
+from typing import Callable, ClassVar, Generator, Generic, Iterator, List, Optional, Set, Tuple, Union
 
 from vellum import (
     AdHocExecutePromptEvent,
@@ -8,6 +8,7 @@ from vellum import (
     ChatMessage,
     FunctionDefinition,
     PromptBlock,
+    PromptOutput,
     PromptParameters,
     PromptRequestChatHistoryInput,
     PromptRequestInput,
@@ -15,17 +16,19 @@ from vellum import (
     PromptRequestStringInput,
     VellumVariable,
 )
-from vellum.client import RequestOptions
+from vellum.client import ApiError, RequestOptions
 from vellum.client.types.chat_message_request import ChatMessageRequest
 from vellum.client.types.prompt_settings import PromptSettings
 from vellum.client.types.rich_text_child_block import RichTextChildBlock
 from vellum.workflows.constants import OMIT
 from vellum.workflows.context import get_execution_context
 from vellum.workflows.errors import WorkflowErrorCode
+from vellum.workflows.errors.types import vellum_error_to_workflow_error
 from vellum.workflows.events.types import default_serializer
 from vellum.workflows.exceptions import NodeException
 from vellum.workflows.nodes.displayable.bases.base_prompt_node import BasePromptNode
 from vellum.workflows.nodes.displayable.bases.inline_prompt_node.constants import DEFAULT_PROMPT_PARAMETERS
+from vellum.workflows.outputs import BaseOutput
 from vellum.workflows.types import MergeBehavior
 from vellum.workflows.types.generics import StateType
 from vellum.workflows.utils.functions import compile_function_definition
@@ -103,17 +106,63 @@ class BaseInlinePromptNode(BasePromptNode[StateType], Generic[StateType]):
             else None
         )
 
-        return self._context.vellum_client.ad_hoc.adhoc_execute_prompt_stream(
-            ml_model=self.ml_model,
-            input_values=input_values,
-            input_variables=input_variables,
-            parameters=self.parameters,
-            blocks=self.blocks,
-            settings=self.settings,
-            functions=normalized_functions,
-            expand_meta=self.expand_meta,
-            request_options=request_options,
-        )
+        if self.settings and not self.settings.stream_enabled:
+            # This endpoint is returning a single event, so we need to wrap it in a generator
+            # to match the existing interface.
+            response = self._context.vellum_client.ad_hoc.adhoc_execute_prompt(
+                ml_model=self.ml_model,
+                input_values=input_values,
+                input_variables=input_variables,
+                parameters=self.parameters,
+                blocks=self.blocks,
+                settings=self.settings,
+                functions=normalized_functions,
+                expand_meta=self.expand_meta,
+                request_options=request_options,
+            )
+            return iter([response])
+        else:
+            return self._context.vellum_client.ad_hoc.adhoc_execute_prompt_stream(
+                ml_model=self.ml_model,
+                input_values=input_values,
+                input_variables=input_variables,
+                parameters=self.parameters,
+                blocks=self.blocks,
+                settings=self.settings,
+                functions=normalized_functions,
+                expand_meta=self.expand_meta,
+                request_options=request_options,
+            )
+
+    def _process_prompt_event_stream(self) -> Generator[BaseOutput, None, Optional[List[PromptOutput]]]:
+        self._validate()
+        try:
+            prompt_event_stream = self._get_prompt_event_stream()
+        except ApiError as e:
+            self._handle_api_error(e)
+
+        if not self.settings or (self.settings and self.settings.stream_enabled):
+            # We don't use the INITIATED event anyway, so we can just skip it
+            # and use the exception handling to catch other api level errors
+            try:
+                next(prompt_event_stream)
+            except ApiError as e:
+                self._handle_api_error(e)
+
+        outputs: Optional[List[PromptOutput]] = None
+        for event in prompt_event_stream:
+            if event.state == "INITIATED":
+                continue
+            elif event.state == "STREAMING":
+                yield BaseOutput(name="results", delta=event.output.value)
+            elif event.state == "FULFILLED":
+                outputs = event.outputs
+                yield BaseOutput(name="results", value=event.outputs)
+            elif event.state == "REJECTED":
+                workflow_error = vellum_error_to_workflow_error(event.error)
+                raise NodeException.of(workflow_error)
+
+        return outputs
 
     def _compile_prompt_inputs(self) -> Tuple[List[VellumVariable], List[PromptRequestInput]]:
         input_variables: List[VellumVariable] = []

vellum/workflows/nodes/displayable/bases/inline_prompt_node/tests/test_inline_prompt_node.py

@@ -5,11 +5,14 @@ from uuid import uuid4
 from typing import Any, Iterator, List
 
 from vellum import (
+    AdHocExecutePromptEvent,
     ChatMessagePromptBlock,
+    FulfilledAdHocExecutePromptEvent,
     JinjaPromptBlock,
     PlainTextPromptBlock,
     PromptBlock,
     PromptParameters,
+    PromptSettings,
     RichTextPromptBlock,
     VariablePromptBlock,
 )
@@ -296,3 +299,177 @@ def test_inline_prompt_node__json_output(vellum_adhoc_prompt_client):
         request_options=mock.ANY,
         settings=None,
     )
+
+
+def test_inline_prompt_node__streaming_disabled(vellum_adhoc_prompt_client):
+    # GIVEN an InlinePromptNode
+    class Inputs(BaseInputs):
+        input: str
+
+    class State(BaseState):
+        pass
+
+    # AND it has streaming disabled
+    class MyInlinePromptNode(InlinePromptNode):
+        ml_model = "gpt-4o"
+        blocks = []
+        parameters = PromptParameters(
+            stop=[],
+            temperature=0.0,
+            max_tokens=4096,
+            top_p=1.0,
+            top_k=0,
+            frequency_penalty=0.0,
+            presence_penalty=0.0,
+            logit_bias=None,
+            custom_parameters={},
+        )
+        settings = PromptSettings(stream_enabled=False)
+
+    # AND a known response from invoking an inline prompt
+    expected_output: list[PromptOutput] = [StringVellumValue(value="Hello, world!")]
+
+    def generate_prompt_event(*args: Any, **kwargs: Any) -> AdHocExecutePromptEvent:
+        execution_id = str(uuid4())
+        return FulfilledAdHocExecutePromptEvent(
+            execution_id=execution_id,
+            outputs=expected_output,
+        )
+
+    vellum_adhoc_prompt_client.adhoc_execute_prompt.side_effect = generate_prompt_event
+
+    # WHEN the node is run
+    node = MyInlinePromptNode()
+    outputs = [o for o in node.run()]
+
+    # THEN the node should have produced the outputs we expect
+    result_output = outputs[0]
+    assert result_output.name == "results"
+    assert result_output.value == expected_output
+
+    # AND we should have made the expected call to Vellum search
+    vellum_adhoc_prompt_client.adhoc_execute_prompt.assert_called_once_with(
+        blocks=[],
+        expand_meta=Ellipsis,
+        functions=None,
+        input_values=[],
+        input_variables=[],
+        ml_model="gpt-4o",
+        parameters=PromptParameters(
+            stop=[],
+            temperature=0.0,
+            max_tokens=4096,
+            top_p=1.0,
+            top_k=0,
+            frequency_penalty=0.0,
+            presence_penalty=0.0,
+            logit_bias=None,
+            custom_parameters={},
+        ),
+        request_options=mock.ANY,
+        settings=PromptSettings(stream_enabled=False),
+    )
+
+
+def test_inline_prompt_node__json_output_with_streaming_disabled(vellum_adhoc_prompt_client):
+    # GIVEN an InlinePromptNode
+    class Inputs(BaseInputs):
+        input: str
+
+    class State(BaseState):
+        pass
+
+    class MyInlinePromptNode(InlinePromptNode):
+        ml_model = "gpt-4o"
+        blocks = []
+        parameters = PromptParameters(
+            stop=[],
+            temperature=0.0,
+            max_tokens=4096,
+            top_p=1.0,
+            top_k=0,
+            frequency_penalty=0.0,
+            presence_penalty=0.0,
+            logit_bias=None,
+            custom_parameters={
+                "json_mode": False,
+                "json_schema": {
+                    "name": "get_result",
+                    "schema": {
+                        "type": "object",
+                        "required": ["result"],
+                        "properties": {"result": {"type": "string", "description": ""}},
+                    },
+                },
+            },
+        )
+        settings = PromptSettings(stream_enabled=False)
+
+    # AND a known JSON response from invoking an inline prompt
+    expected_json = {"result": "Hello, world!"}
+    expected_outputs: List[PromptOutput] = [
+        StringVellumValue(value=json.dumps(expected_json)),
+    ]
+
+    def generate_prompt_event(*args: Any, **kwargs: Any) -> AdHocExecutePromptEvent:
+        execution_id = str(uuid4())
+        return FulfilledAdHocExecutePromptEvent(
+            execution_id=execution_id,
+            outputs=expected_outputs,
+        )
+
+    vellum_adhoc_prompt_client.adhoc_execute_prompt.side_effect = generate_prompt_event
+
+    # WHEN the node is run
+    node = MyInlinePromptNode(
+        state=State(
+            meta=StateMeta(workflow_inputs=Inputs(input="Generate JSON.")),
+        )
+    )
+    outputs = [o for o in node.run()]
+
+    # THEN the node should have produced the outputs we expect
+    results_output = outputs[0]
+    assert results_output.name == "results"
+    assert results_output.value == expected_outputs
+
+    text_output = outputs[1]
+    assert text_output.name == "text"
+    assert text_output.value == '{"result": "Hello, world!"}'
+
+    json_output = outputs[2]
+    assert json_output.name == "json"
+    assert json_output.value == expected_json
+
+    # AND we should have made the expected call to Vellum search
+    vellum_adhoc_prompt_client.adhoc_execute_prompt.assert_called_once_with(
+        blocks=[],
+        expand_meta=Ellipsis,
+        functions=None,
+        input_values=[],
+        input_variables=[],
+        ml_model="gpt-4o",
+        parameters=PromptParameters(
+            stop=[],
+            temperature=0.0,
+            max_tokens=4096,
+            top_p=1.0,
+            top_k=0,
+            frequency_penalty=0.0,
+            presence_penalty=0.0,
+            logit_bias=None,
+            custom_parameters={
+                "json_mode": False,
+                "json_schema": {
+                    "name": "get_result",
+                    "schema": {
+                        "type": "object",
+                        "required": ["result"],
+                        "properties": {"result": {"type": "string", "description": ""}},
+                    },
+                },
+            },
+        ),
+        request_options=mock.ANY,
+        settings=PromptSettings(stream_enabled=False),
+    )

vellum/workflows/nodes/experimental/tool_calling_node/node.py

@@ -8,11 +8,7 @@ from vellum.workflows.exceptions import NodeException
 from vellum.workflows.graph.graph import Graph
 from vellum.workflows.inputs.base import BaseInputs
 from vellum.workflows.nodes.bases import BaseNode
-from vellum.workflows.nodes.experimental.tool_calling_node.utils import (
-    ToolRouterNode,
-    create_function_node,
-    create_tool_router_node,
-)
+from vellum.workflows.nodes.experimental.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
 from vellum.workflows.state.context import WorkflowContext
@@ -69,7 +65,7 @@ class ToolCallingNode(BaseNode):
                 graph = self._graph
 
                 class Outputs(BaseWorkflow.Outputs):
-                    text: str = ToolRouterNode.Outputs.text
+                    text: str = self.tool_router_node.Outputs.text
                     chat_history: List[ChatMessage] = ToolCallingState.chat_history
 
             subworkflow = ToolCallingWorkflow(
@@ -107,6 +103,7 @@ class ToolCallingNode(BaseNode):
         self._function_nodes = {
             function.__name__: create_function_node(
                 function=function,
+                tool_router_node=self.tool_router_node,
             )
             for function in self.functions
         }

vellum/workflows/nodes/experimental/tool_calling_node/utils.py

@@ -59,8 +59,8 @@ def create_tool_router_node(
         function_name = function.__name__
         port_condition = LazyReference(
             lambda: (
-                ToolRouterNode.Outputs.results[0]["type"].equals("FUNCTION_CALL")
-                & ToolRouterNode.Outputs.results[0]["value"]["name"].equals(function_name)
+                node.Outputs.results[0]["type"].equals("FUNCTION_CALL")
+                & node.Outputs.results[0]["value"]["name"].equals(function_name)
             )
         )
         port = Port.on_if(port_condition)
@@ -78,23 +78,26 @@ def create_tool_router_node(
         )
     )
 
-    node = type(
-        "ToolRouterNode",
-        (ToolRouterNode,),
-        {
-            "ml_model": ml_model,
-            "blocks": blocks,
-            "functions": functions,
-            "prompt_inputs": prompt_inputs,
-            "Ports": Ports,
-            "__module__": __name__,
-        },
+    node = cast(
+        Type[ToolRouterNode],
+        type(
+            "ToolRouterNode",
+            (ToolRouterNode,),
+            {
+                "ml_model": ml_model,
+                "blocks": blocks,
+                "functions": functions,
+                "prompt_inputs": prompt_inputs,
+                "Ports": Ports,
+                "__module__": __name__,
+            },
+        ),
     )
 
     return node
 
 
-def create_function_node(function: Callable[..., Any]) -> Type[FunctionNode]:
+def create_function_node(function: Callable[..., Any], tool_router_node: Type[ToolRouterNode]) -> Type[FunctionNode]:
     """
     Create a FunctionNode class for a given function.
 
@@ -103,7 +106,7 @@ def create_function_node(function: Callable[..., Any]) -> Type[FunctionNode]:
 
     # Create a class-level wrapper that calls the original function
     def execute_function(self) -> BaseNode.Outputs:
-        outputs = self.state.meta.node_outputs.get(ToolRouterNode.Outputs.text)
+        outputs = self.state.meta.node_outputs.get(tool_router_node.Outputs.text)
         # first parse into json
         outputs = json.loads(outputs)
         arguments = outputs["arguments"]

vellum/workflows/references/output.py

@@ -78,7 +78,7 @@ class OutputReference(BaseDescriptor[_OutputType], Generic[_OutputType]):
     def __eq__(self, other: object) -> bool:
         if not isinstance(other, type(self)):
             return False
-        return super().__eq__(other) and self._outputs_class == other._outputs_class
+        return super().__eq__(other) and id(self._outputs_class) == id(other._outputs_class)
 
     def __hash__(self) -> int:
         return hash((self._outputs_class, self._name))

{vellum_ai-0.14.42.dist-info → vellum_ai-0.14.43.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vellum-ai
-Version: 0.14.42
+Version: 0.14.43
 Summary: 
 License: MIT
 Requires-Python: >=3.9,<4.0

{vellum_ai-0.14.42.dist-info → vellum_ai-0.14.43.dist-info}/RECORD

@@ -7,7 +7,7 @@ vellum_cli/image_push.py,sha256=8DDvRDJEZ-FukUCqGW1827bg1ybF4xBbx9WyqWYQE-g,6816
 vellum_cli/init.py,sha256=WpnMXPItPmh0f0bBGIer3p-e5gu8DUGwSArT_FuoMEw,5093
 vellum_cli/logger.py,sha256=PuRFa0WCh4sAGFS5aqWB0QIYpS6nBWwPJrIXpWxugV4,1022
 vellum_cli/ping.py,sha256=p_BCCRjgPhng6JktuECtkDQLbhopt6JpmrtGoLnLJT8,1161
-vellum_cli/pull.py,sha256=27Mh10aQ8H1OkTmQOJcOuJ5cQcYbNjkkuQrBmtkRe0o,12262
+vellum_cli/pull.py,sha256=2hSJGeqooevMb--mcvRLQ1GYT-9290cI7VdSRifzmTg,12561
 vellum_cli/push.py,sha256=0bHAhp6B67JUFuiL9sSekrLOw-x2Dp6VsrzhFbOy7e4,9508
 vellum_cli/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 vellum_cli/tests/conftest.py,sha256=AFYZryKA2qnUuCPBxBKmHLFoPiE0WhBFFej9tNwSHdc,1526
@@ -16,7 +16,7 @@ vellum_cli/tests/test_image_push.py,sha256=i3lJuW8nFRwL1M1OF6752IZYvGAFgKmkB2hd_
 vellum_cli/tests/test_init.py,sha256=8UOc_ThfouR4ja5cCl_URuLk7ohr9JXfCnG4yka1OUQ,18754
 vellum_cli/tests/test_main.py,sha256=qDZG-aQauPwBwM6A2DIu1494n47v3pL28XakTbLGZ-k,272
 vellum_cli/tests/test_ping.py,sha256=3ucVRThEmTadlV9LrJdCCrr1Ofj3rOjG6ue0BNR2UC0,2523
-vellum_cli/tests/test_pull.py,sha256=5nIDX5yXLyi6p-siLvrzfaepI4RWbaoMwgFoQFUEhI0,46692
+vellum_cli/tests/test_pull.py,sha256=iTxVbJGuehvgNt8Vp9W3Y5Bvaocfws8bl8LMGEbc_qQ,47508
 vellum_cli/tests/test_push.py,sha256=j22l7p_cy1KXdcvQKhWiM2bpu-3WL1q5IJquRm84mxE,25580
 vellum_ee/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 vellum_ee/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -125,12 +125,12 @@ vellum_ee/workflows/tests/local_workflow/workflow.py,sha256=A4qOzOPNwePYxWbcAgIP
 vellum_ee/workflows/tests/test_display_meta.py,sha256=C25dErwghPNXio49pvSRxyOuc96srH6eYEwTAWdE2zY,2258
 vellum_ee/workflows/tests/test_server.py,sha256=SsOkS6sGO7uGC4mxvk4iv8AtcXs058P9hgFHzTWmpII,14519
 vellum_ee/workflows/tests/test_virtual_files.py,sha256=TJEcMR0v2S8CkloXNmCHA0QW0K6pYNGaIjraJz7sFvY,2762
-vellum/__init__.py,sha256=dGOb8Oe4E3jff14Y51TevTM2wx5z4C0nPCBrxW2E-Yg,41708
+vellum/__init__.py,sha256=4UwLVTlKBQ0cTIJ3NTqxw51aEFNUPfmJ-rVYP1yVUU0,41758
 vellum/client/README.md,sha256=JkCJjmMZl4jrPj46pkmL9dpK4gSzQQmP5I7z4aME4LY,4749
 vellum/client/__init__.py,sha256=Z-JHK2jGxhtTtmkLeOaUGGJWIUNYGNVBLvUewC6lp6w,118148
 vellum/client/core/__init__.py,sha256=SQ85PF84B9MuKnBwHNHWemSGuy-g_515gFYNFhvEE0I,1438
 vellum/client/core/api_error.py,sha256=RE8LELok2QCjABadECTvtDp7qejA1VmINCh6TbqPwSE,426
-vellum/client/core/client_wrapper.py,sha256=HDYrFzioATPcqeXlA9AJXpStw-Rw1q_cI0rzgpCOpNU,1869
+vellum/client/core/client_wrapper.py,sha256=gOQ_e1s2iQI79ioggNAaNo71Ue9CGaAkWTvlVyZxaMI,1869
 vellum/client/core/datetime_utils.py,sha256=nBys2IsYrhPdszxGKCNRPSOCwa-5DWOHG95FB8G9PKo,1047
 vellum/client/core/file.py,sha256=X9IbmkZmB2bB_DpmZAO3crWdXagOakAyn6UCOCImCPg,2322
 vellum/client/core/http_client.py,sha256=R0pQpCppnEtxccGvXl4uJ76s7ro_65Fo_erlNNLp_AI,19228
@@ -146,7 +146,7 @@ vellum/client/errors/bad_request_error.py,sha256=_EbO8mWqN9kFZPvIap8qa1lL_EWkRcs
 vellum/client/errors/forbidden_error.py,sha256=QO1kKlhClAPES6zsEK7g9pglWnxn3KWaOCAawWOg6Aw,263
 vellum/client/errors/internal_server_error.py,sha256=8USCagXyJJ1MOm9snpcXIUt6eNXvrd_aq7Gfcu1vlOI,268
 vellum/client/errors/not_found_error.py,sha256=tBVCeBC8n3C811WHRj_n-hs3h8MqwR5gp0vLiobk7W8,262
-vellum/client/reference.md,sha256=D7qqNgorznQIlZHVqW6UMjTmPSI2zSF2QuMx5KFn4F4,89585
+vellum/client/reference.md,sha256=23-W_ajXfNLsPKA25ICSXpSxxnI0sRQgYCAk6D2w9LM,90890
 vellum/client/resources/__init__.py,sha256=XgQao4rJxyYu71j64RFIsshz4op9GE8-i-C5GCv-KVE,1555
 vellum/client/resources/ad_hoc/__init__.py,sha256=FTtvy8EDg9nNNg9WCatVgKTRYV8-_v1roeGPAKoa_pw,65
 vellum/client/resources/ad_hoc/client.py,sha256=93FXK-Wpvh72G8ji2__2Dmc5OYl9G5GRHiknyGIjeX4,25557
@@ -174,7 +174,7 @@ vellum/client/resources/ml_models/client.py,sha256=XIYapTEY6GRNr7V0Kjy5bEeKmrhv9
 vellum/client/resources/organizations/__init__.py,sha256=FTtvy8EDg9nNNg9WCatVgKTRYV8-_v1roeGPAKoa_pw,65
 vellum/client/resources/organizations/client.py,sha256=Uye92moqjAcOCs4astmuFpT92QdC5SLMunA-C8_G-gA,3675
 vellum/client/resources/prompts/__init__.py,sha256=FTtvy8EDg9nNNg9WCatVgKTRYV8-_v1roeGPAKoa_pw,65
-vellum/client/resources/prompts/client.py,sha256=_rNTUjhl_ZF3vyQa_M1BSTrX4DlFXU_SXkwwCEYKD2s,6598
+vellum/client/resources/prompts/client.py,sha256=C9t17kHuVsgarw1dEk8NSEaRXb-WOi0VT8huT2jySlk,14044
 vellum/client/resources/release_reviews/__init__.py,sha256=FTtvy8EDg9nNNg9WCatVgKTRYV8-_v1roeGPAKoa_pw,65
 vellum/client/resources/release_reviews/client.py,sha256=VLXcmw1o8cYYtdTJQpajJWE2ve1z40_IXIbQRQIhqpY,9395
 vellum/client/resources/sandboxes/__init__.py,sha256=FTtvy8EDg9nNNg9WCatVgKTRYV8-_v1roeGPAKoa_pw,65
@@ -198,7 +198,7 @@ vellum/client/resources/workspace_secrets/__init__.py,sha256=FTtvy8EDg9nNNg9WCat
 vellum/client/resources/workspace_secrets/client.py,sha256=h7UzXLyTttPq1t-JZGMg1BWxypxJvBGUdqg7KGT7MK4,8027
 vellum/client/resources/workspaces/__init__.py,sha256=FTtvy8EDg9nNNg9WCatVgKTRYV8-_v1roeGPAKoa_pw,65
 vellum/client/resources/workspaces/client.py,sha256=RthwzN1o-Jxwg5yyNNodavFyNUSxfLoTv26w3mRR5g8,3595
-vellum/client/types/__init__.py,sha256=JanJFC8Ha3zl3jT4kZgUzIbj2oo4AY1sONJH1nM7r_o,63143
+vellum/client/types/__init__.py,sha256=DODWRvaxjqdVCvpr4b6kxku_2WZKSvl7lEl6tSr_gXU,63222
 vellum/client/types/ad_hoc_execute_prompt_event.py,sha256=bCjujA2XsOgyF3bRZbcEqV2rOIymRgsLoIRtZpB14xg,607
 vellum/client/types/ad_hoc_expand_meta.py,sha256=1gv-NCsy_6xBYupLvZH979yf2VMdxAU-l0y0ynMKZaw,1331
 vellum/client/types/ad_hoc_fulfilled_prompt_execution_meta.py,sha256=oDG60TpwK1YNSKhRsBbiP2O3ZF9PKR-M9chGIfKw4R4,1004
@@ -269,7 +269,7 @@ vellum/client/types/create_test_suite_test_case_request.py,sha256=SYUz7_aZMQlin_
 vellum/client/types/deployment_history_item.py,sha256=YfcHo4X5OjHXsffndZoAjShYncUN19ZwIm96qKE0G7o,1310
 vellum/client/types/deployment_provider_payload_response.py,sha256=b0lkt0rK88ARQaMWn9MAHeWtMBsZKofDMlOAUsQvv7g,818
 vellum/client/types/deployment_provider_payload_response_payload.py,sha256=xHLQnWFN0AZRZdrOiKawwpoKK7BTmnZfp0P7FCc2ZqE,188
-vellum/client/types/deployment_read.py,sha256=e1Z3vHwtJ1AsNE83PqXte9aFV2LhqEK4zRUeMSVJPGA,2160
+vellum/client/types/deployment_read.py,sha256=GwegkT3HzTJL5_awWKn8rFdpILgtwWEZIPE_YCqXwWg,2139
 vellum/client/types/deployment_release_tag_deployment_history_item.py,sha256=df4qKHT1f-z0jnRS4UmP8MQe6u3PwYej_d8KDF7EL88,631
 vellum/client/types/deployment_release_tag_read.py,sha256=dUrTOz9LH1gAvC_ktMB_7NztkeBnlNSX_9x15Ld3D3I,1278
 vellum/client/types/docker_service_token.py,sha256=T0icNHBKsIs6TrEiDRjckM_f37hcF1DMwEE8161tTvY,614
@@ -505,6 +505,7 @@ vellum/client/types/prompt_node_result.py,sha256=3jewO-nPodoXTq_5RxgwhKfDZrvoPjR
 vellum/client/types/prompt_node_result_data.py,sha256=fNOxBfK3ablDBxkUWVVstJMYaGdHGgu27WxP87E6UQ4,872
 vellum/client/types/prompt_output.py,sha256=NpDGJNIYIivzQJnBeoJLpJlCk7gqBESLwv5Qtn_20qQ,398
 vellum/client/types/prompt_parameters.py,sha256=Vkwh4zI9gX1DuGQxrWiUUa1TshTfnPlS7_yRrziD5qg,1046
+vellum/client/types/prompt_push_response.py,sha256=Un61pv4kVH0Omd8OKVfTk1yZ1RWrwuquAgl1vanb04Y,601
 vellum/client/types/prompt_request_chat_history_input.py,sha256=DB2io5piMSyA89f5lnIVYO4MLZoNALNSufx8Y-oOwOE,790
 vellum/client/types/prompt_request_input.py,sha256=brEdYhYm74Ac8XjK9wF0rKOLgnqd_Cg19yMS7VfB4qQ,400
 vellum/client/types/prompt_request_json_input.py,sha256=vLhwvCWL_yjVfDzT4921xK4Ql92OkvG-ruvOC_uppFI,739
@@ -1178,6 +1179,7 @@ vellum/types/prompt_node_result.py,sha256=9ootTTh8lscQ-0WE0-bqdmn7XFvpP7uavO-g7m
 vellum/types/prompt_node_result_data.py,sha256=2ivj-nvfHUB85fCsQHJV8nACA6lsqHWR2PFf5AqycBI,161
 vellum/types/prompt_output.py,sha256=t5Ue9qqNCyDJ54i4YPF4kQ4nbbzgrXT1srqv_PEdtQU,151
 vellum/types/prompt_parameters.py,sha256=MnKA48eOcybD-gSTp5jLmzib-U2KeSQ0G1VJpPg5qJc,155
+vellum/types/prompt_push_response.py,sha256=VZq4LnRDeMWdMIW5WqaU2v1jDUKKakerxakox781118,158
 vellum/types/prompt_request_chat_history_input.py,sha256=Ngf_R0NgpFWmLKP1zQsXANaRB8P9LqJuk5a0Dlm-78U,171
 vellum/types/prompt_request_input.py,sha256=kjHHbam1HcDm9_1br9iqlZhrke3nRJYh4UNZs1h5qy4,158
 vellum/types/prompt_request_json_input.py,sha256=OlXiUPchxe184SWbmIvbmARpY9YWPi8yPqWaVC8xoBU,163
@@ -1534,9 +1536,10 @@ vellum/workflows/inputs/tests/test_inputs.py,sha256=lioA8917mFLYq7Ml69UNkqUjcWbb
 vellum/workflows/logging.py,sha256=_a217XogktV4Ncz6xKFz7WfYmZAzkfVRVuC0rWob8ls,437
 vellum/workflows/nodes/__init__.py,sha256=aVdQVv7Y3Ro3JlqXGpxwaU2zrI06plDHD2aumH5WUIs,1157
 vellum/workflows/nodes/bases/__init__.py,sha256=cniHuz_RXdJ4TQgD8CBzoiKDiPxg62ErdVpCbWICX64,58
-vellum/workflows/nodes/bases/base.py,sha256=eW-3RSkBgtuGY8x2nmbHYiUg_HXS5U57n3k6Fh-dJ9s,15330
+vellum/workflows/nodes/bases/base.py,sha256=tX3xOQIPfnspQPNaOMwGoZ93gM5d5dWeK1YUukzkINI,15158
 vellum/workflows/nodes/bases/base_adornment_node.py,sha256=Ao2opOW4kgNoYXFF9Pk7IMpVZdy6luwrjcqEwU5Q9V0,3404
 vellum/workflows/nodes/bases/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+vellum/workflows/nodes/bases/tests/test_base_adornment_node.py,sha256=fXZI9KqpS4XMBrBnIEkK3foHaBVvyHwYcQWWDKay7ic,1148
 vellum/workflows/nodes/bases/tests/test_base_node.py,sha256=6J85q-vtfG-NHzCndMKLk5_sEEDiI52sTGaxefcUCOU,7892
 vellum/workflows/nodes/core/__init__.py,sha256=5zDMCmyt1v0HTJzlUBwq3U9L825yZGZhT9JL18-mRR4,455
 vellum/workflows/nodes/core/error_node/__init__.py,sha256=g7RRnlHhqu4qByfLjBwCunmgGA8dI5gNsjS3h6TwlSI,60
@@ -1572,9 +1575,9 @@ vellum/workflows/nodes/displayable/bases/base_prompt_node/__init__.py,sha256=Org
 vellum/workflows/nodes/displayable/bases/base_prompt_node/node.py,sha256=amBXi7Tv50AbGLhfWbwX83PlOdV1XyYRyQmpa6_afE4,3511
 vellum/workflows/nodes/displayable/bases/inline_prompt_node/__init__.py,sha256=Hl35IAoepRpE-j4cALaXVJIYTYOF3qszyVbxTj4kS1s,82
 vellum/workflows/nodes/displayable/bases/inline_prompt_node/constants.py,sha256=fnjiRWLoRlC4Puo5oQcpZD5Hd-EesxsAo9l5tGAkpZQ,270
-vellum/workflows/nodes/displayable/bases/inline_prompt_node/node.py,sha256=rga24gkK9_STRhFwhBwGL7oHhTTZvLWS_rXHHrp85p4,8386
+vellum/workflows/nodes/displayable/bases/inline_prompt_node/node.py,sha256=wqN1EjyjTL6McUmlkHWu3GXVzcNaqDjavvmKUHDaVqg,10623
 vellum/workflows/nodes/displayable/bases/inline_prompt_node/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-vellum/workflows/nodes/displayable/bases/inline_prompt_node/tests/test_inline_prompt_node.py,sha256=ZCXCZs-_OyPk4nqCpuWY-vw87lg92TDZ2tK_gckJ7mg,10450
+vellum/workflows/nodes/displayable/bases/inline_prompt_node/tests/test_inline_prompt_node.py,sha256=YPOFoaEBENfOzE_qWo3WdQ_E1dQk78aLCWk8gOMvTjg,16042
 vellum/workflows/nodes/displayable/bases/prompt_deployment_node.py,sha256=lwH7mfiHcRKFxU1Y9IPQVgb3o5trssuhwlKnA30rTWk,9777
 vellum/workflows/nodes/displayable/bases/search_node.py,sha256=3UtbqY3QO4kzfJHbmUNZGnEEfJmaoiF892u8H6TGjp8,5381
 vellum/workflows/nodes/displayable/bases/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -1629,8 +1632,8 @@ vellum/workflows/nodes/experimental/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQ
 vellum/workflows/nodes/experimental/openai_chat_completion_node/__init__.py,sha256=lsyD9laR9p7kx5-BXGH2gUTM242UhKy8SMV0SR6S2iE,90
 vellum/workflows/nodes/experimental/openai_chat_completion_node/node.py,sha256=cKI2Ls25L-JVt4z4a2ozQa-YBeVy21Z7BQ32Sj7iBPE,10460
 vellum/workflows/nodes/experimental/tool_calling_node/__init__.py,sha256=S7OzT3I4cyOU5Beoz87nPwCejCMP2FsHBFL8OcVmxJ4,118
-vellum/workflows/nodes/experimental/tool_calling_node/node.py,sha256=w73v2pfpqFaTOYbtHkH7nFuGgqQebLO6dr_Mxn1n7Dc,4848
-vellum/workflows/nodes/experimental/tool_calling_node/utils.py,sha256=hwJ1GjoNOEWoCdT1R0b6gVRAAGWGTUlKutTgBF7GRP4,4664
+vellum/workflows/nodes/experimental/tool_calling_node/node.py,sha256=NUC7VZj2D86IDQzjCq_a3-Xeqj_b3BE8T1kOMIfN7V8,4878
+vellum/workflows/nodes/experimental/tool_calling_node/utils.py,sha256=_b4xqs2jEQY9aWCCJsFvZZrvXo74NeYiIkD7uJ9RHeU,4781
 vellum/workflows/nodes/mocks.py,sha256=a1FjWEIocseMfjzM-i8DNozpUsaW0IONRpZmXBoWlyc,10455
 vellum/workflows/nodes/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 vellum/workflows/nodes/tests/test_mocks.py,sha256=mfPvrs75PKcsNsbJLQAN6PDFoVqs9TmQxpdyFKDdO60,7837
@@ -1650,7 +1653,7 @@ vellum/workflows/references/external_input.py,sha256=c_4SojTpykCSbGS1Pjmx9FfquyY
 vellum/workflows/references/input.py,sha256=3INu-TLTi4dziWmva6LO3WvgDlPzsjayUx61cVvqLJA,325
 vellum/workflows/references/lazy.py,sha256=jgUYmgt-yAybzPf_R-74MzdU8VuNwMYI8EQqrj9lVR0,2948
 vellum/workflows/references/node.py,sha256=LP854wDVs-9I_aZ7-nkbwXqL2H7W2_3LED2e9FixNS8,1418
-vellum/workflows/references/output.py,sha256=QWrqLGpeo6mFxC_fKmL3LHIfrcDD9SzOhQdHvG99I9Y,3379
+vellum/workflows/references/output.py,sha256=Odpjqnw2uY6lbmt49sUwDclBPZMndYxgtFAGn1iKj8k,3387
 vellum/workflows/references/state_value.py,sha256=bInUF0A3Pt4-zhA0f6LdSuyv8tz7n5QRkHAEn4gsmqI,711
 vellum/workflows/references/tests/test_lazy.py,sha256=0s50-LizMTlSTBQahpK0fg_xqCucA8YTp6QmIMqPvMk,919
 vellum/workflows/references/vellum_secret.py,sha256=Od4d19a5yletWMqNfJR5d_mZQUkVcFzj29mE-T9J7yE,480
@@ -1695,8 +1698,8 @@ vellum/workflows/workflows/event_filters.py,sha256=GSxIgwrX26a1Smfd-6yss2abGCnad
 vellum/workflows/workflows/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 vellum/workflows/workflows/tests/test_base_workflow.py,sha256=8P5YIsNMO78_CR1NNK6wkEdkMB4b3Q_Ni1qxh78OnHo,20481
 vellum/workflows/workflows/tests/test_context.py,sha256=VJBUcyWVtMa_lE5KxdhgMu0WYNYnUQUDvTF7qm89hJ0,2333
-vellum_ai-0.14.42.dist-info/LICENSE,sha256=hOypcdt481qGNISA784bnAGWAE6tyIf9gc2E78mYC3E,1574
-vellum_ai-0.14.42.dist-info/METADATA,sha256=J5skujRAzUMKICz1RgTzBNlCpPUYQgSpgFEuNkYCuf0,5484
-vellum_ai-0.14.42.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-vellum_ai-0.14.42.dist-info/entry_points.txt,sha256=HCH4yc_V3J_nDv3qJzZ_nYS8llCHZViCDP1ejgCc5Ak,42
-vellum_ai-0.14.42.dist-info/RECORD,,
+vellum_ai-0.14.43.dist-info/LICENSE,sha256=hOypcdt481qGNISA784bnAGWAE6tyIf9gc2E78mYC3E,1574
+vellum_ai-0.14.43.dist-info/METADATA,sha256=_cI_OM2v1FvIfRXuOWWaVpZdA0DFUSezO_ENVULlBbI,5484
+vellum_ai-0.14.43.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+vellum_ai-0.14.43.dist-info/entry_points.txt,sha256=HCH4yc_V3J_nDv3qJzZ_nYS8llCHZViCDP1ejgCc5Ak,42
+vellum_ai-0.14.43.dist-info/RECORD,,

vellum_cli/pull.py

@@ -179,6 +179,13 @@ def pull_command(
     except ApiError as e:
         if e.status_code == 401 or e.status_code == 403:
             raise Exception("Please make sure your `VELLUM_API_KEY` environment variable is set correctly.")
+
+        if e.status_code == 500:
+            raise Exception(
+                "The Pull API failed with an unexpected error. Please try again later and contact support if the problem persists."  # noqa: E501
+            )
+
+        # TODO: We should return an Origin header in to validate this case
         raise Exception(
             "The API we tried to pull is invalid. Please make sure your `VELLUM_API_URL` environment variable is set correctly."  # noqa: E501
         )

vellum_cli/tests/test_pull.py

@@ -950,6 +950,29 @@ def test_pull__unauthorized_error_path(vellum_client):
     assert str(result.exception) == "Please make sure your `VELLUM_API_KEY` environment variable is set correctly."
 
 
+def test_pull__unexpected_error_path(vellum_client):
+    workflow_deployment = "test-workflow-deployment-id"
+
+    # GIVEN an unauthorized error with the error message from the API
+    def mock_error_generator():
+        yield b""
+        raise ApiError(status_code=500, body={"detail": "Internal server error"})
+
+    vellum_client.workflows.pull.return_value = mock_error_generator()
+
+    # WHEN the user runs the pull command
+    runner = CliRunner()
+    result = runner.invoke(cli_main, ["workflows", "pull", "--workflow-deployment", workflow_deployment])
+
+    # THEN the command returns an error
+    assert result.exit_code == 1
+    assert (
+        str(result.exception)
+        == """The Pull API failed with an unexpected error. \
+Please try again later and contact support if the problem persists."""
+    )
+
+
 @pytest.mark.parametrize(
     "workflow_deployment",
     [