vellum-ai 0.14.41__py3-none-any.whl → 1.11.2__py3-none-any.whl

vellum/client/resources/ad_hoc/client.py

@@ -1,25 +1,18 @@
 # This file was auto-generated by Fern from our API Definition.
 
 import typing
-from ...core.client_wrapper import SyncClientWrapper
-from ...types.prompt_request_input import PromptRequestInput
-from ...types.vellum_variable import VellumVariable
-from ...types.prompt_parameters import PromptParameters
-from ...types.prompt_block import PromptBlock
-from ...types.prompt_settings import PromptSettings
-from ...types.function_definition import FunctionDefinition
-from ...types.ad_hoc_expand_meta import AdHocExpandMeta
+
+from ...core.client_wrapper import AsyncClientWrapper, SyncClientWrapper
 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
-from ...errors.bad_request_error import BadRequestError
-from ...errors.forbidden_error import ForbiddenError
-from ...errors.internal_server_error import InternalServerError
-from json.decoder import JSONDecodeError
-from ...core.api_error import ApiError
-import json
-from ...core.client_wrapper import AsyncClientWrapper
+from ...types.ad_hoc_expand_meta import AdHocExpandMeta
+from ...types.deprecated_prompt_request_input import DeprecatedPromptRequestInput
+from ...types.function_definition import FunctionDefinition
+from ...types.prompt_block import PromptBlock
+from ...types.prompt_parameters import PromptParameters
+from ...types.prompt_settings import PromptSettings
+from ...types.vellum_variable import VellumVariable
+from .raw_client import AsyncRawAdHocClient, RawAdHocClient
 
 # this is used as the default value for optional parameters
 OMIT = typing.cast(typing.Any, ...)
@@ -27,13 +20,28 @@ OMIT = typing.cast(typing.Any, ...)
 
 class AdHocClient:
     def __init__(self, *, client_wrapper: SyncClientWrapper):
-        self._client_wrapper = client_wrapper
+        self._raw_client = RawAdHocClient(client_wrapper=client_wrapper)
+
+    @property
+    def _client_wrapper(self) -> SyncClientWrapper:
+        return self._raw_client._client_wrapper
+
+    @property
+    def with_raw_response(self) -> RawAdHocClient:
+        """
+        Retrieves a raw implementation of this client that returns raw responses.
+
+        Returns
+        -------
+        RawAdHocClient
+        """
+        return self._raw_client
 
     def adhoc_execute_prompt(
         self,
         *,
         ml_model: str,
-        input_values: typing.Sequence[PromptRequestInput],
+        input_values: typing.Sequence[DeprecatedPromptRequestInput],
         input_variables: typing.Sequence[VellumVariable],
         parameters: PromptParameters,
         blocks: typing.Sequence[PromptBlock],
@@ -47,7 +55,7 @@ class AdHocClient:
         ----------
         ml_model : str
 
-        input_values : typing.Sequence[PromptRequestInput]
+        input_values : typing.Sequence[DeprecatedPromptRequestInput]
 
         input_variables : typing.Sequence[VellumVariable]
 
@@ -74,117 +82,68 @@ class AdHocClient:
         from vellum import (
             JinjaPromptBlock,
             PromptParameters,
-            PromptRequestStringInput,
+            StringInput,
             Vellum,
             VellumVariable,
         )
 
         client = Vellum(
+            api_version="YOUR_API_VERSION",
             api_key="YOUR_API_KEY",
         )
         client.ad_hoc.adhoc_execute_prompt(
-            ml_model="ml_model",
+            ml_model="x",
             input_values=[
-                PromptRequestStringInput(
-                    key="key",
+                StringInput(
+                    name="x",
                     value="value",
-                )
+                ),
+                StringInput(
+                    name="x",
+                    value="value",
+                ),
             ],
             input_variables=[
                 VellumVariable(
-                    id="id",
+                    id="x",
+                    key="key",
+                    type="STRING",
+                ),
+                VellumVariable(
+                    id="x",
                     key="key",
                     type="STRING",
-                )
+                ),
             ],
             parameters=PromptParameters(),
             blocks=[
                 JinjaPromptBlock(
                     template="template",
-                )
+                ),
+                JinjaPromptBlock(
+                    template="template",
+                ),
             ],
         )
         """
-        _response = self._client_wrapper.httpx_client.request(
-            "v1/ad-hoc/execute-prompt",
-            base_url=self._client_wrapper.get_environment().default,
-            method="POST",
-            json={
-                "ml_model": ml_model,
-                "input_values": convert_and_respect_annotation_metadata(
-                    object_=input_values, annotation=typing.Sequence[PromptRequestInput], direction="write"
-                ),
-                "input_variables": convert_and_respect_annotation_metadata(
-                    object_=input_variables, annotation=typing.Sequence[VellumVariable], direction="write"
-                ),
-                "parameters": convert_and_respect_annotation_metadata(
-                    object_=parameters, annotation=PromptParameters, direction="write"
-                ),
-                "settings": convert_and_respect_annotation_metadata(
-                    object_=settings, annotation=PromptSettings, direction="write"
-                ),
-                "blocks": convert_and_respect_annotation_metadata(
-                    object_=blocks, annotation=typing.Sequence[PromptBlock], direction="write"
-                ),
-                "functions": convert_and_respect_annotation_metadata(
-                    object_=functions, annotation=typing.Sequence[FunctionDefinition], direction="write"
-                ),
-                "expand_meta": convert_and_respect_annotation_metadata(
-                    object_=expand_meta, annotation=AdHocExpandMeta, direction="write"
-                ),
-            },
+        _response = self._raw_client.adhoc_execute_prompt(
+            ml_model=ml_model,
+            input_values=input_values,
+            input_variables=input_variables,
+            parameters=parameters,
+            blocks=blocks,
+            settings=settings,
+            functions=functions,
+            expand_meta=expand_meta,
             request_options=request_options,
-            omit=OMIT,
         )
-        try:
-            if 200 <= _response.status_code < 300:
-                return typing.cast(
-                    AdHocExecutePromptEvent,
-                    parse_obj_as(
-                        type_=AdHocExecutePromptEvent,  # 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 == 403:
-                raise ForbiddenError(
-                    typing.cast(
-                        typing.Optional[typing.Any],
-                        parse_obj_as(
-                            type_=typing.Optional[typing.Any],  # type: ignore
-                            object_=_response.json(),
-                        ),
-                    )
-                )
-            if _response.status_code == 500:
-                raise InternalServerError(
-                    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)
+        return _response.data
 
     def adhoc_execute_prompt_stream(
         self,
         *,
         ml_model: str,
-        input_values: typing.Sequence[PromptRequestInput],
+        input_values: typing.Sequence[DeprecatedPromptRequestInput],
         input_variables: typing.Sequence[VellumVariable],
         parameters: PromptParameters,
         blocks: typing.Sequence[PromptBlock],
@@ -198,7 +157,7 @@ class AdHocClient:
         ----------
         ml_model : str
 
-        input_values : typing.Sequence[PromptRequestInput]
+        input_values : typing.Sequence[DeprecatedPromptRequestInput]
 
         input_variables : typing.Sequence[VellumVariable]
 
@@ -225,132 +184,86 @@ class AdHocClient:
         from vellum import (
             JinjaPromptBlock,
             PromptParameters,
-            PromptRequestStringInput,
+            StringInput,
             Vellum,
             VellumVariable,
         )
 
         client = Vellum(
+            api_version="YOUR_API_VERSION",
             api_key="YOUR_API_KEY",
         )
         response = client.ad_hoc.adhoc_execute_prompt_stream(
-            ml_model="ml_model",
+            ml_model="x",
             input_values=[
-                PromptRequestStringInput(
-                    key="key",
+                StringInput(
+                    name="x",
+                    value="value",
+                ),
+                StringInput(
+                    name="x",
                     value="value",
-                )
+                ),
             ],
             input_variables=[
                 VellumVariable(
-                    id="id",
+                    id="x",
                     key="key",
                     type="STRING",
-                )
+                ),
+                VellumVariable(
+                    id="x",
+                    key="key",
+                    type="STRING",
+                ),
             ],
             parameters=PromptParameters(),
             blocks=[
                 JinjaPromptBlock(
                     template="template",
-                )
+                ),
+                JinjaPromptBlock(
+                    template="template",
+                ),
             ],
         )
         for chunk in response:
             yield chunk
         """
-        with self._client_wrapper.httpx_client.stream(
-            "v1/ad-hoc/execute-prompt-stream",
-            base_url=self._client_wrapper.get_environment().predict,
-            method="POST",
-            json={
-                "ml_model": ml_model,
-                "input_values": convert_and_respect_annotation_metadata(
-                    object_=input_values, annotation=typing.Sequence[PromptRequestInput], direction="write"
-                ),
-                "input_variables": convert_and_respect_annotation_metadata(
-                    object_=input_variables, annotation=typing.Sequence[VellumVariable], direction="write"
-                ),
-                "parameters": convert_and_respect_annotation_metadata(
-                    object_=parameters, annotation=PromptParameters, direction="write"
-                ),
-                "settings": convert_and_respect_annotation_metadata(
-                    object_=settings, annotation=PromptSettings, direction="write"
-                ),
-                "blocks": convert_and_respect_annotation_metadata(
-                    object_=blocks, annotation=typing.Sequence[PromptBlock], direction="write"
-                ),
-                "functions": convert_and_respect_annotation_metadata(
-                    object_=functions, annotation=typing.Sequence[FunctionDefinition], direction="write"
-                ),
-                "expand_meta": convert_and_respect_annotation_metadata(
-                    object_=expand_meta, annotation=AdHocExpandMeta, direction="write"
-                ),
-            },
+        with self._raw_client.adhoc_execute_prompt_stream(
+            ml_model=ml_model,
+            input_values=input_values,
+            input_variables=input_variables,
+            parameters=parameters,
+            blocks=blocks,
+            settings=settings,
+            functions=functions,
+            expand_meta=expand_meta,
             request_options=request_options,
-            omit=OMIT,
-        ) as _response:
-            try:
-                if 200 <= _response.status_code < 300:
-                    for _text in _response.iter_lines():
-                        try:
-                            if len(_text) == 0:
-                                continue
-                            yield typing.cast(
-                                AdHocExecutePromptEvent,
-                                parse_obj_as(
-                                    type_=AdHocExecutePromptEvent,  # type: ignore
-                                    object_=json.loads(_text),
-                                ),
-                            )
-                        except:
-                            pass
-                    return
-                _response.read()
-                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 == 403:
-                    raise ForbiddenError(
-                        typing.cast(
-                            typing.Optional[typing.Any],
-                            parse_obj_as(
-                                type_=typing.Optional[typing.Any],  # type: ignore
-                                object_=_response.json(),
-                            ),
-                        )
-                    )
-                if _response.status_code == 500:
-                    raise InternalServerError(
-                        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)
+        ) as r:
+            yield from r.data
 
 
 class AsyncAdHocClient:
     def __init__(self, *, client_wrapper: AsyncClientWrapper):
-        self._client_wrapper = client_wrapper
+        self._raw_client = AsyncRawAdHocClient(client_wrapper=client_wrapper)
+
+    @property
+    def with_raw_response(self) -> AsyncRawAdHocClient:
+        """
+        Retrieves a raw implementation of this client that returns raw responses.
+
+        Returns
+        -------
+        AsyncRawAdHocClient
+        """
+        return self._raw_client
 
     async def adhoc_execute_prompt(
         self,
         *,
         ml_model: str,
-        input_values: typing.Sequence[PromptRequestInput],
+        input_values: typing.Sequence[DeprecatedPromptRequestInput],
         input_variables: typing.Sequence[VellumVariable],
         parameters: PromptParameters,
         blocks: typing.Sequence[PromptBlock],
@@ -364,7 +277,7 @@ class AsyncAdHocClient:
         ----------
         ml_model : str
 
-        input_values : typing.Sequence[PromptRequestInput]
+        input_values : typing.Sequence[DeprecatedPromptRequestInput]
 
         input_variables : typing.Sequence[VellumVariable]
 
@@ -394,122 +307,73 @@ class AsyncAdHocClient:
             AsyncVellum,
             JinjaPromptBlock,
             PromptParameters,
-            PromptRequestStringInput,
+            StringInput,
             VellumVariable,
         )
 
         client = AsyncVellum(
+            api_version="YOUR_API_VERSION",
             api_key="YOUR_API_KEY",
         )
 
 
         async def main() -> None:
             await client.ad_hoc.adhoc_execute_prompt(
-                ml_model="ml_model",
+                ml_model="x",
                 input_values=[
-                    PromptRequestStringInput(
-                        key="key",
+                    StringInput(
+                        name="x",
                         value="value",
-                    )
+                    ),
+                    StringInput(
+                        name="x",
+                        value="value",
+                    ),
                 ],
                 input_variables=[
                     VellumVariable(
-                        id="id",
+                        id="x",
+                        key="key",
+                        type="STRING",
+                    ),
+                    VellumVariable(
+                        id="x",
                         key="key",
                         type="STRING",
-                    )
+                    ),
                 ],
                 parameters=PromptParameters(),
                 blocks=[
                     JinjaPromptBlock(
                         template="template",
-                    )
+                    ),
+                    JinjaPromptBlock(
+                        template="template",
+                    ),
                 ],
             )
 
 
         asyncio.run(main())
         """
-        _response = await self._client_wrapper.httpx_client.request(
-            "v1/ad-hoc/execute-prompt",
-            base_url=self._client_wrapper.get_environment().default,
-            method="POST",
-            json={
-                "ml_model": ml_model,
-                "input_values": convert_and_respect_annotation_metadata(
-                    object_=input_values, annotation=typing.Sequence[PromptRequestInput], direction="write"
-                ),
-                "input_variables": convert_and_respect_annotation_metadata(
-                    object_=input_variables, annotation=typing.Sequence[VellumVariable], direction="write"
-                ),
-                "parameters": convert_and_respect_annotation_metadata(
-                    object_=parameters, annotation=PromptParameters, direction="write"
-                ),
-                "settings": convert_and_respect_annotation_metadata(
-                    object_=settings, annotation=PromptSettings, direction="write"
-                ),
-                "blocks": convert_and_respect_annotation_metadata(
-                    object_=blocks, annotation=typing.Sequence[PromptBlock], direction="write"
-                ),
-                "functions": convert_and_respect_annotation_metadata(
-                    object_=functions, annotation=typing.Sequence[FunctionDefinition], direction="write"
-                ),
-                "expand_meta": convert_and_respect_annotation_metadata(
-                    object_=expand_meta, annotation=AdHocExpandMeta, direction="write"
-                ),
-            },
+        _response = await self._raw_client.adhoc_execute_prompt(
+            ml_model=ml_model,
+            input_values=input_values,
+            input_variables=input_variables,
+            parameters=parameters,
+            blocks=blocks,
+            settings=settings,
+            functions=functions,
+            expand_meta=expand_meta,
             request_options=request_options,
-            omit=OMIT,
         )
-        try:
-            if 200 <= _response.status_code < 300:
-                return typing.cast(
-                    AdHocExecutePromptEvent,
-                    parse_obj_as(
-                        type_=AdHocExecutePromptEvent,  # 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 == 403:
-                raise ForbiddenError(
-                    typing.cast(
-                        typing.Optional[typing.Any],
-                        parse_obj_as(
-                            type_=typing.Optional[typing.Any],  # type: ignore
-                            object_=_response.json(),
-                        ),
-                    )
-                )
-            if _response.status_code == 500:
-                raise InternalServerError(
-                    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)
+        return _response.data
 
     async def adhoc_execute_prompt_stream(
         self,
         *,
         ml_model: str,
-        input_values: typing.Sequence[PromptRequestInput],
+        input_values: typing.Sequence[DeprecatedPromptRequestInput],
         input_variables: typing.Sequence[VellumVariable],
         parameters: PromptParameters,
         blocks: typing.Sequence[PromptBlock],
@@ -523,7 +387,7 @@ class AsyncAdHocClient:
         ----------
         ml_model : str
 
-        input_values : typing.Sequence[PromptRequestInput]
+        input_values : typing.Sequence[DeprecatedPromptRequestInput]
 
         input_variables : typing.Sequence[VellumVariable]
 
@@ -553,36 +417,49 @@ class AsyncAdHocClient:
             AsyncVellum,
             JinjaPromptBlock,
             PromptParameters,
-            PromptRequestStringInput,
+            StringInput,
             VellumVariable,
         )
 
         client = AsyncVellum(
+            api_version="YOUR_API_VERSION",
             api_key="YOUR_API_KEY",
         )
 
 
         async def main() -> None:
             response = await client.ad_hoc.adhoc_execute_prompt_stream(
-                ml_model="ml_model",
+                ml_model="x",
                 input_values=[
-                    PromptRequestStringInput(
-                        key="key",
+                    StringInput(
+                        name="x",
+                        value="value",
+                    ),
+                    StringInput(
+                        name="x",
                         value="value",
-                    )
+                    ),
                 ],
                 input_variables=[
                     VellumVariable(
-                        id="id",
+                        id="x",
                         key="key",
                         type="STRING",
-                    )
+                    ),
+                    VellumVariable(
+                        id="x",
+                        key="key",
+                        type="STRING",
+                    ),
                 ],
                 parameters=PromptParameters(),
                 blocks=[
                     JinjaPromptBlock(
                         template="template",
-                    )
+                    ),
+                    JinjaPromptBlock(
+                        template="template",
+                    ),
                 ],
             )
             async for chunk in response:
@@ -591,85 +468,16 @@ class AsyncAdHocClient:
 
         asyncio.run(main())
         """
-        async with self._client_wrapper.httpx_client.stream(
-            "v1/ad-hoc/execute-prompt-stream",
-            base_url=self._client_wrapper.get_environment().predict,
-            method="POST",
-            json={
-                "ml_model": ml_model,
-                "input_values": convert_and_respect_annotation_metadata(
-                    object_=input_values, annotation=typing.Sequence[PromptRequestInput], direction="write"
-                ),
-                "input_variables": convert_and_respect_annotation_metadata(
-                    object_=input_variables, annotation=typing.Sequence[VellumVariable], direction="write"
-                ),
-                "parameters": convert_and_respect_annotation_metadata(
-                    object_=parameters, annotation=PromptParameters, direction="write"
-                ),
-                "settings": convert_and_respect_annotation_metadata(
-                    object_=settings, annotation=PromptSettings, direction="write"
-                ),
-                "blocks": convert_and_respect_annotation_metadata(
-                    object_=blocks, annotation=typing.Sequence[PromptBlock], direction="write"
-                ),
-                "functions": convert_and_respect_annotation_metadata(
-                    object_=functions, annotation=typing.Sequence[FunctionDefinition], direction="write"
-                ),
-                "expand_meta": convert_and_respect_annotation_metadata(
-                    object_=expand_meta, annotation=AdHocExpandMeta, direction="write"
-                ),
-            },
+        async with self._raw_client.adhoc_execute_prompt_stream(
+            ml_model=ml_model,
+            input_values=input_values,
+            input_variables=input_variables,
+            parameters=parameters,
+            blocks=blocks,
+            settings=settings,
+            functions=functions,
+            expand_meta=expand_meta,
             request_options=request_options,
-            omit=OMIT,
-        ) as _response:
-            try:
-                if 200 <= _response.status_code < 300:
-                    async for _text in _response.aiter_lines():
-                        try:
-                            if len(_text) == 0:
-                                continue
-                            yield typing.cast(
-                                AdHocExecutePromptEvent,
-                                parse_obj_as(
-                                    type_=AdHocExecutePromptEvent,  # type: ignore
-                                    object_=json.loads(_text),
-                                ),
-                            )
-                        except:
-                            pass
-                    return
-                await _response.aread()
-                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 == 403:
-                    raise ForbiddenError(
-                        typing.cast(
-                            typing.Optional[typing.Any],
-                            parse_obj_as(
-                                type_=typing.Optional[typing.Any],  # type: ignore
-                                object_=_response.json(),
-                            ),
-                        )
-                    )
-                if _response.status_code == 500:
-                    raise InternalServerError(
-                        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)
+        ) as r:
+            async for _chunk in r.data:
+                yield _chunk