vellum-ai 0.8.13__py3-none-any.whl → 0.8.15__py3-none-any.whl

vellum/__init__.py

@@ -14,9 +14,12 @@ from .types import (
     ArrayChatMessageContentItem,
     ArrayChatMessageContentItemRequest,
     ArrayChatMessageContentRequest,
+    ArrayVariableValue,
     ArrayVariableValueItem,
+    ArrayVellumValue,
     ArrayVellumValueItem,
     ArrayVellumValueItemRequest,
+    ArrayVellumValueRequest,
     BasicVectorizerIntfloatMultilingualE5Large,
     BasicVectorizerIntfloatMultilingualE5LargeRequest,
     BasicVectorizerSentenceTransformersMultiQaMpnetBaseCosV1,
@@ -24,6 +27,9 @@ from .types import (
     BasicVectorizerSentenceTransformersMultiQaMpnetBaseDotV1,
     BasicVectorizerSentenceTransformersMultiQaMpnetBaseDotV1Request,
     ChatHistoryInputRequest,
+    ChatHistoryVariableValue,
+    ChatHistoryVellumValue,
+    ChatHistoryVellumValueRequest,
     ChatMessage,
     ChatMessageContent,
     ChatMessageContentRequest,
@@ -258,6 +264,9 @@ from .types import (
     SearchResultMeta,
     SearchResultMetaRequest,
     SearchResultRequest,
+    SearchResultsVariableValue,
+    SearchResultsVellumValue,
+    SearchResultsVellumValueRequest,
     SearchWeightsRequest,
     SentenceChunkerConfig,
     SentenceChunkerConfigRequest,
@@ -452,9 +461,12 @@ __all__ = [
     "ArrayChatMessageContentItem",
     "ArrayChatMessageContentItemRequest",
     "ArrayChatMessageContentRequest",
+    "ArrayVariableValue",
     "ArrayVariableValueItem",
+    "ArrayVellumValue",
     "ArrayVellumValueItem",
     "ArrayVellumValueItemRequest",
+    "ArrayVellumValueRequest",
     "AsyncVellum",
     "BadRequestError",
     "BasicVectorizerIntfloatMultilingualE5Large",
@@ -464,6 +476,9 @@ __all__ = [
     "BasicVectorizerSentenceTransformersMultiQaMpnetBaseDotV1",
     "BasicVectorizerSentenceTransformersMultiQaMpnetBaseDotV1Request",
     "ChatHistoryInputRequest",
+    "ChatHistoryVariableValue",
+    "ChatHistoryVellumValue",
+    "ChatHistoryVellumValueRequest",
     "ChatMessage",
     "ChatMessageContent",
     "ChatMessageContentRequest",
@@ -703,6 +718,9 @@ __all__ = [
     "SearchResultMeta",
     "SearchResultMetaRequest",
     "SearchResultRequest",
+    "SearchResultsVariableValue",
+    "SearchResultsVellumValue",
+    "SearchResultsVellumValueRequest",
     "SearchWeightsRequest",
     "SentenceChunkerConfig",
     "SentenceChunkerConfigRequest",

vellum/client.py

@@ -19,6 +19,7 @@ from .types.prompt_deployment_expand_meta_request import PromptDeploymentExpandM
 from .types.raw_prompt_execution_overrides_request import RawPromptExecutionOverridesRequest
 from .core.request_options import RequestOptions
 from .types.execute_prompt_response import ExecutePromptResponse
+from .core.serialization import convert_and_respect_annotation_metadata
 from .core.pydantic_utilities import parse_obj_as
 from .errors.bad_request_error import BadRequestError
 from .errors.forbidden_error import ForbiddenError
@@ -197,13 +198,19 @@ class Vellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[PromptDeploymentInputRequest], direction="write"
+                ),
                 "prompt_deployment_id": prompt_deployment_id,
                 "prompt_deployment_name": prompt_deployment_name,
                 "release_tag": release_tag,
                 "external_id": external_id,
-                "expand_meta": expand_meta,
-                "raw_overrides": raw_overrides,
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=PromptDeploymentExpandMetaRequest, direction="write"
+                ),
+                "raw_overrides": convert_and_respect_annotation_metadata(
+                    object_=raw_overrides, annotation=RawPromptExecutionOverridesRequest, direction="write"
+                ),
                 "expand_raw": expand_raw,
                 "metadata": metadata,
             },
@@ -366,13 +373,19 @@ class Vellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[PromptDeploymentInputRequest], direction="write"
+                ),
                 "prompt_deployment_id": prompt_deployment_id,
                 "prompt_deployment_name": prompt_deployment_name,
                 "release_tag": release_tag,
                 "external_id": external_id,
-                "expand_meta": expand_meta,
-                "raw_overrides": raw_overrides,
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=PromptDeploymentExpandMetaRequest, direction="write"
+                ),
+                "raw_overrides": convert_and_respect_annotation_metadata(
+                    object_=raw_overrides, annotation=RawPromptExecutionOverridesRequest, direction="write"
+                ),
                 "expand_raw": expand_raw,
                 "metadata": metadata,
             },
@@ -508,8 +521,12 @@ class Vellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
-                "expand_meta": expand_meta,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[WorkflowRequestInputRequest], direction="write"
+                ),
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=WorkflowExpandMetaRequest, direction="write"
+                ),
                 "workflow_deployment_id": workflow_deployment_id,
                 "workflow_deployment_name": workflow_deployment_name,
                 "release_tag": release_tag,
@@ -649,8 +666,12 @@ class Vellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
-                "expand_meta": expand_meta,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[WorkflowRequestInputRequest], direction="write"
+                ),
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=WorkflowExpandMetaRequest, direction="write"
+                ),
                 "workflow_deployment_id": workflow_deployment_id,
                 "workflow_deployment_name": workflow_deployment_name,
                 "release_tag": release_tag,
@@ -772,8 +793,12 @@ class Vellum:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "requests": requests,
-                "options": options,
+                "requests": convert_and_respect_annotation_metadata(
+                    object_=requests, annotation=typing.Sequence[GenerateRequest], direction="write"
+                ),
+                "options": convert_and_respect_annotation_metadata(
+                    object_=options, annotation=GenerateOptionsRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -915,8 +940,12 @@ class Vellum:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "requests": requests,
-                "options": options,
+                "requests": convert_and_respect_annotation_metadata(
+                    object_=requests, annotation=typing.Sequence[GenerateRequest], direction="write"
+                ),
+                "options": convert_and_respect_annotation_metadata(
+                    object_=options, annotation=GenerateOptionsRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -1036,7 +1065,9 @@ class Vellum:
                 "index_id": index_id,
                 "index_name": index_name,
                 "query": query,
-                "options": options,
+                "options": convert_and_respect_annotation_metadata(
+                    object_=options, annotation=SearchRequestOptionsRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -1132,7 +1163,9 @@ class Vellum:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "actuals": actuals,
+                "actuals": convert_and_respect_annotation_metadata(
+                    object_=actuals, annotation=typing.Sequence[SubmitCompletionActualRequest], direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -1222,7 +1255,9 @@ class Vellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "actuals": actuals,
+                "actuals": convert_and_respect_annotation_metadata(
+                    object_=actuals, annotation=typing.Sequence[SubmitWorkflowExecutionActualRequest], direction="write"
+                ),
                 "execution_id": execution_id,
                 "external_id": external_id,
             },
@@ -1386,13 +1421,19 @@ class AsyncVellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[PromptDeploymentInputRequest], direction="write"
+                ),
                 "prompt_deployment_id": prompt_deployment_id,
                 "prompt_deployment_name": prompt_deployment_name,
                 "release_tag": release_tag,
                 "external_id": external_id,
-                "expand_meta": expand_meta,
-                "raw_overrides": raw_overrides,
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=PromptDeploymentExpandMetaRequest, direction="write"
+                ),
+                "raw_overrides": convert_and_respect_annotation_metadata(
+                    object_=raw_overrides, annotation=RawPromptExecutionOverridesRequest, direction="write"
+                ),
                 "expand_raw": expand_raw,
                 "metadata": metadata,
             },
@@ -1563,13 +1604,19 @@ class AsyncVellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[PromptDeploymentInputRequest], direction="write"
+                ),
                 "prompt_deployment_id": prompt_deployment_id,
                 "prompt_deployment_name": prompt_deployment_name,
                 "release_tag": release_tag,
                 "external_id": external_id,
-                "expand_meta": expand_meta,
-                "raw_overrides": raw_overrides,
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=PromptDeploymentExpandMetaRequest, direction="write"
+                ),
+                "raw_overrides": convert_and_respect_annotation_metadata(
+                    object_=raw_overrides, annotation=RawPromptExecutionOverridesRequest, direction="write"
+                ),
                 "expand_raw": expand_raw,
                 "metadata": metadata,
             },
@@ -1713,8 +1760,12 @@ class AsyncVellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
-                "expand_meta": expand_meta,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[WorkflowRequestInputRequest], direction="write"
+                ),
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=WorkflowExpandMetaRequest, direction="write"
+                ),
                 "workflow_deployment_id": workflow_deployment_id,
                 "workflow_deployment_name": workflow_deployment_name,
                 "release_tag": release_tag,
@@ -1862,8 +1913,12 @@ class AsyncVellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "inputs": inputs,
-                "expand_meta": expand_meta,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[WorkflowRequestInputRequest], direction="write"
+                ),
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=WorkflowExpandMetaRequest, direction="write"
+                ),
                 "workflow_deployment_id": workflow_deployment_id,
                 "workflow_deployment_name": workflow_deployment_name,
                 "release_tag": release_tag,
@@ -1993,8 +2048,12 @@ class AsyncVellum:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "requests": requests,
-                "options": options,
+                "requests": convert_and_respect_annotation_metadata(
+                    object_=requests, annotation=typing.Sequence[GenerateRequest], direction="write"
+                ),
+                "options": convert_and_respect_annotation_metadata(
+                    object_=options, annotation=GenerateOptionsRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -2144,8 +2203,12 @@ class AsyncVellum:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "requests": requests,
-                "options": options,
+                "requests": convert_and_respect_annotation_metadata(
+                    object_=requests, annotation=typing.Sequence[GenerateRequest], direction="write"
+                ),
+                "options": convert_and_respect_annotation_metadata(
+                    object_=options, annotation=GenerateOptionsRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -2273,7 +2336,9 @@ class AsyncVellum:
                 "index_id": index_id,
                 "index_name": index_name,
                 "query": query,
-                "options": options,
+                "options": convert_and_respect_annotation_metadata(
+                    object_=options, annotation=SearchRequestOptionsRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -2377,7 +2442,9 @@ class AsyncVellum:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "actuals": actuals,
+                "actuals": convert_and_respect_annotation_metadata(
+                    object_=actuals, annotation=typing.Sequence[SubmitCompletionActualRequest], direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -2475,7 +2542,9 @@ class AsyncVellum:
             base_url=self._client_wrapper.get_environment().predict,
             method="POST",
             json={
-                "actuals": actuals,
+                "actuals": convert_and_respect_annotation_metadata(
+                    object_=actuals, annotation=typing.Sequence[SubmitWorkflowExecutionActualRequest], direction="write"
+                ),
                 "execution_id": execution_id,
                 "external_id": external_id,
             },

vellum/core/client_wrapper.py

@@ -17,7 +17,7 @@ class BaseClientWrapper:
         headers: typing.Dict[str, str] = {
             "X-Fern-Language": "Python",
             "X-Fern-SDK-Name": "vellum-ai",
-            "X-Fern-SDK-Version": "0.8.13",
+            "X-Fern-SDK-Version": "0.8.15",
         }
         headers["X_API_KEY"] = self.api_key
         return headers

vellum/core/http_client.py

@@ -224,7 +224,9 @@ class HttpClient:
             json=json_body,
             data=data_body,
             content=content,
-            files=convert_file_dict_to_httpx_tuples(remove_none_from_dict(files)) if files is not None else None,
+            files=convert_file_dict_to_httpx_tuples(remove_none_from_dict(files))
+            if (files is not None and files is not omit)
+            else None,
             timeout=timeout,
         )
 
@@ -306,7 +308,9 @@ class HttpClient:
             json=json_body,
             data=data_body,
             content=content,
-            files=convert_file_dict_to_httpx_tuples(remove_none_from_dict(files)) if files is not None else None,
+            files=convert_file_dict_to_httpx_tuples(remove_none_from_dict(files))
+            if (files is not None and files is not omit)
+            else None,
             timeout=timeout,
         ) as stream:
             yield stream

vellum/core/pydantic_utilities.py

@@ -10,6 +10,7 @@ import typing_extensions
 import pydantic
 
 from .datetime_utils import serialize_datetime
+from .serialization import convert_and_respect_annotation_metadata
 
 IS_PYDANTIC_V2 = pydantic.VERSION.startswith("2.")
 
@@ -56,11 +57,12 @@ Model = typing.TypeVar("Model", bound=pydantic.BaseModel)
 
 
 def parse_obj_as(type_: typing.Type[T], object_: typing.Any) -> T:
+    dealiased_object = convert_and_respect_annotation_metadata(object_=object_, annotation=type_, direction="read")
     if IS_PYDANTIC_V2:
         adapter = pydantic.TypeAdapter(type_)  # type: ignore # Pydantic v2
-        return adapter.validate_python(object_)
+        return adapter.validate_python(dealiased_object)
     else:
-        return pydantic.parse_obj_as(type_, object_)
+        return pydantic.parse_obj_as(type_, dealiased_object)
 
 
 def to_jsonable_with_fallback(
@@ -75,13 +77,40 @@ def to_jsonable_with_fallback(
 
 
 class UniversalBaseModel(pydantic.BaseModel):
-    class Config:
-        populate_by_name = True
-        smart_union = True
-        allow_population_by_field_name = True
-        json_encoders = {dt.datetime: serialize_datetime}
-        # Allow fields begining with `model_` to be used in the model
-        protected_namespaces = ()
+    if IS_PYDANTIC_V2:
+        model_config: typing.ClassVar[pydantic.ConfigDict] = pydantic.ConfigDict(
+            # Allow fields begining with `model_` to be used in the model
+            protected_namespaces=(),
+        )  # type: ignore # Pydantic v2
+
+        @pydantic.model_serializer(mode="wrap", when_used="json")  # type: ignore # Pydantic v2
+        def serialize_model(self, handler: pydantic.SerializerFunctionWrapHandler) -> typing.Any:  # type: ignore # Pydantic v2
+            serialized = handler(self)
+            data = {k: serialize_datetime(v) if isinstance(v, dt.datetime) else v for k, v in serialized.items()}
+            return data
+
+    else:
+
+        class Config:
+            smart_union = True
+            json_encoders = {dt.datetime: serialize_datetime}
+
+    @classmethod
+    def model_construct(
+        cls: typing.Type["Model"], _fields_set: typing.Optional[typing.Set[str]] = None, **values: typing.Any
+    ) -> "Model":
+        dealiased_object = convert_and_respect_annotation_metadata(object_=values, annotation=cls, direction="read")
+        return cls.construct(_fields_set, **dealiased_object)
+
+    @classmethod
+    def construct(
+        cls: typing.Type["Model"], _fields_set: typing.Optional[typing.Set[str]] = None, **values: typing.Any
+    ) -> "Model":
+        dealiased_object = convert_and_respect_annotation_metadata(object_=values, annotation=cls, direction="read")
+        if IS_PYDANTIC_V2:
+            return super().model_construct(_fields_set, **dealiased_object)  # type: ignore # Pydantic v2
+        else:
+            return super().construct(_fields_set, **dealiased_object)
 
     def json(self, **kwargs: typing.Any) -> str:
         kwargs_with_defaults: typing.Any = {
@@ -117,7 +146,7 @@ class UniversalBaseModel(pydantic.BaseModel):
                 "exclude_none": True,
                 "exclude_unset": False,
             }
-            return deep_union_pydantic_dicts(
+            dict_dump = deep_union_pydantic_dicts(
                 super().model_dump(**kwargs_with_defaults_exclude_unset),  # type: ignore # Pydantic v2
                 super().model_dump(**kwargs_with_defaults_exclude_none),  # type: ignore # Pydantic v2
             )
@@ -143,7 +172,9 @@ class UniversalBaseModel(pydantic.BaseModel):
                 **kwargs,
             }
 
-            return super().dict(**kwargs_with_defaults_exclude_unset_include_fields)
+            dict_dump = super().dict(**kwargs_with_defaults_exclude_unset_include_fields)
+
+        return convert_and_respect_annotation_metadata(object_=dict_dump, annotation=self.__class__, direction="write")
 
 
 def deep_union_pydantic_dicts(

vellum/core/serialization.py

@@ -71,6 +71,24 @@ def convert_and_respect_annotation_metadata(
     if typing_extensions.is_typeddict(clean_type) and isinstance(object_, typing.Mapping):
         return _convert_mapping(object_, clean_type, direction)
 
+    if (
+        typing_extensions.get_origin(clean_type) == typing.Dict
+        or typing_extensions.get_origin(clean_type) == dict
+        or clean_type == typing.Dict
+    ) and isinstance(object_, typing.Dict):
+        key_type = typing_extensions.get_args(clean_type)[0]
+        value_type = typing_extensions.get_args(clean_type)[1]
+
+        return {
+            key: convert_and_respect_annotation_metadata(
+                object_=value,
+                annotation=annotation,
+                inner_type=value_type,
+                direction=direction,
+            )
+            for key, value in object_.items()
+        }
+
     # If you're iterating on a string, do not bother to coerce it to a sequence.
     if not isinstance(object_, str):
         if (

vellum/resources/ad_hoc/client.py

@@ -9,6 +9,7 @@ from ...types.prompt_block_request import PromptBlockRequest
 from ...types.ad_hoc_expand_meta_request import AdHocExpandMetaRequest
 from ...core.request_options import RequestOptions
 from ...types.ad_hoc_execute_prompt_event import AdHocExecutePromptEvent
+from ...core.serialization import convert_and_respect_annotation_metadata
 from ...core.pydantic_utilities import parse_obj_as
 import json
 from ...errors.bad_request_error import BadRequestError
@@ -133,11 +134,21 @@ class AdHocClient:
             method="POST",
             json={
                 "ml_model": ml_model,
-                "input_values": input_values,
-                "input_variables": input_variables,
-                "parameters": parameters,
-                "blocks": blocks,
-                "expand_meta": expand_meta,
+                "input_values": convert_and_respect_annotation_metadata(
+                    object_=input_values, annotation=typing.Sequence[PromptRequestInputRequest], direction="write"
+                ),
+                "input_variables": convert_and_respect_annotation_metadata(
+                    object_=input_variables, annotation=typing.Sequence[VellumVariableRequest], direction="write"
+                ),
+                "parameters": convert_and_respect_annotation_metadata(
+                    object_=parameters, annotation=PromptParametersRequest, direction="write"
+                ),
+                "blocks": convert_and_respect_annotation_metadata(
+                    object_=blocks, annotation=typing.Sequence[PromptBlockRequest], direction="write"
+                ),
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=AdHocExpandMetaRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -314,11 +325,21 @@ class AsyncAdHocClient:
             method="POST",
             json={
                 "ml_model": ml_model,
-                "input_values": input_values,
-                "input_variables": input_variables,
-                "parameters": parameters,
-                "blocks": blocks,
-                "expand_meta": expand_meta,
+                "input_values": convert_and_respect_annotation_metadata(
+                    object_=input_values, annotation=typing.Sequence[PromptRequestInputRequest], direction="write"
+                ),
+                "input_variables": convert_and_respect_annotation_metadata(
+                    object_=input_variables, annotation=typing.Sequence[VellumVariableRequest], direction="write"
+                ),
+                "parameters": convert_and_respect_annotation_metadata(
+                    object_=parameters, annotation=PromptParametersRequest, direction="write"
+                ),
+                "blocks": convert_and_respect_annotation_metadata(
+                    object_=blocks, annotation=typing.Sequence[PromptBlockRequest], direction="write"
+                ),
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=AdHocExpandMetaRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,

vellum/resources/deployments/client.py

@@ -14,6 +14,7 @@ from ...types.deployment_release_tag_read import DeploymentReleaseTagRead
 from ...types.prompt_deployment_input_request import PromptDeploymentInputRequest
 from ...types.compile_prompt_deployment_expand_meta_request import CompilePromptDeploymentExpandMetaRequest
 from ...types.deployment_provider_payload_response import DeploymentProviderPayloadResponse
+from ...core.serialization import convert_and_respect_annotation_metadata
 from ...errors.bad_request_error import BadRequestError
 from ...errors.forbidden_error import ForbiddenError
 from ...errors.not_found_error import NotFoundError
@@ -334,9 +335,13 @@ class DeploymentsClient:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "inputs": inputs,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[PromptDeploymentInputRequest], direction="write"
+                ),
                 "release_tag": release_tag,
-                "expand_meta": expand_meta,
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=CompilePromptDeploymentExpandMetaRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,
@@ -746,9 +751,13 @@ class AsyncDeploymentsClient:
             json={
                 "deployment_id": deployment_id,
                 "deployment_name": deployment_name,
-                "inputs": inputs,
+                "inputs": convert_and_respect_annotation_metadata(
+                    object_=inputs, annotation=typing.Sequence[PromptDeploymentInputRequest], direction="write"
+                ),
                 "release_tag": release_tag,
-                "expand_meta": expand_meta,
+                "expand_meta": convert_and_respect_annotation_metadata(
+                    object_=expand_meta, annotation=CompilePromptDeploymentExpandMetaRequest, direction="write"
+                ),
             },
             request_options=request_options,
             omit=OMIT,

vellum/resources/document_indexes/client.py

@@ -12,6 +12,7 @@ from ...types.document_index_indexing_config_request import DocumentIndexIndexin
 from ...types.entity_status import EntityStatus
 from ...types.environment_enum import EnvironmentEnum
 from ...types.document_index_read import DocumentIndexRead
+from ...core.serialization import convert_and_respect_annotation_metadata
 from ...core.jsonable_encoder import jsonable_encoder
 from ...core.client_wrapper import AsyncClientWrapper
 
@@ -191,7 +192,9 @@ class DocumentIndexesClient:
                 "name": name,
                 "status": status,
                 "environment": environment,
-                "indexing_config": indexing_config,
+                "indexing_config": convert_and_respect_annotation_metadata(
+                    object_=indexing_config, annotation=DocumentIndexIndexingConfigRequest, direction="write"
+                ),
                 "copy_documents_from_index_id": copy_documents_from_index_id,
             },
             request_options=request_options,
@@ -740,7 +743,9 @@ class AsyncDocumentIndexesClient:
                 "name": name,
                 "status": status,
                 "environment": environment,
-                "indexing_config": indexing_config,
+                "indexing_config": convert_and_respect_annotation_metadata(
+                    object_=indexing_config, annotation=DocumentIndexIndexingConfigRequest, direction="write"
+                ),
                 "copy_documents_from_index_id": copy_documents_from_index_id,
             },
             request_options=request_options,

vellum/resources/documents/client.py

@@ -261,10 +261,10 @@ class DocumentsClient:
         *,
         label: str,
         contents: core.File,
-        add_to_index_names: typing.Optional[typing.List[str]] = None,
-        external_id: typing.Optional[str] = None,
-        keywords: typing.Optional[typing.List[str]] = None,
-        metadata: typing.Optional[str] = None,
+        add_to_index_names: typing.Optional[typing.List[str]] = OMIT,
+        external_id: typing.Optional[str] = OMIT,
+        keywords: typing.Optional[typing.List[str]] = OMIT,
+        metadata: typing.Optional[str] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> UploadDocumentResponse:
         """
@@ -653,10 +653,10 @@ class AsyncDocumentsClient:
         *,
         label: str,
         contents: core.File,
-        add_to_index_names: typing.Optional[typing.List[str]] = None,
-        external_id: typing.Optional[str] = None,
-        keywords: typing.Optional[typing.List[str]] = None,
-        metadata: typing.Optional[str] = None,
+        add_to_index_names: typing.Optional[typing.List[str]] = OMIT,
+        external_id: typing.Optional[str] = OMIT,
+        keywords: typing.Optional[typing.List[str]] = OMIT,
+        metadata: typing.Optional[str] = OMIT,
         request_options: typing.Optional[RequestOptions] = None,
     ) -> UploadDocumentResponse:
         """