pydantic-ai-slim 0.0.31__py3-none-any.whl → 0.0.33__py3-none-any.whl

pydantic_ai/models/instrumented.py

@@ -6,7 +6,6 @@ from contextlib import asynccontextmanager, contextmanager
 from dataclasses import dataclass, field
 from typing import Any, Callable, Literal
 
-import logfire_api
 from opentelemetry._events import Event, EventLogger, EventLoggerProvider, get_event_logger_provider
 from opentelemetry.trace import Span, Tracer, TracerProvider, get_tracer_provider
 from opentelemetry.util.types import AttributeValue
@@ -59,27 +58,15 @@ class InstrumentedModel(WrapperModel):
         event_logger_provider: EventLoggerProvider | None = None,
         event_mode: Literal['attributes', 'logs'] = 'attributes',
     ):
+        from pydantic_ai import __version__
+
         super().__init__(wrapped)
         tracer_provider = tracer_provider or get_tracer_provider()
         event_logger_provider = event_logger_provider or get_event_logger_provider()
-        self.tracer = tracer_provider.get_tracer('pydantic-ai')
-        self.event_logger = event_logger_provider.get_event_logger('pydantic-ai')
+        self.tracer = tracer_provider.get_tracer('pydantic-ai', __version__)
+        self.event_logger = event_logger_provider.get_event_logger('pydantic-ai', __version__)
         self.event_mode = event_mode
 
-    @classmethod
-    def from_logfire(
-        cls,
-        wrapped: Model | KnownModelName,
-        logfire_instance: logfire_api.Logfire = logfire_api.DEFAULT_LOGFIRE_INSTANCE,
-        event_mode: Literal['attributes', 'logs'] = 'attributes',
-    ) -> InstrumentedModel:
-        if hasattr(logfire_instance.config, 'get_event_logger_provider'):
-            event_provider = logfire_instance.config.get_event_logger_provider()
-        else:
-            event_provider = None
-        tracer_provider = logfire_instance.config.get_tracer_provider()
-        return cls(wrapped, tracer_provider, event_provider, event_mode)
-
     async def request(
         self,
         messages: list[ModelMessage],
@@ -199,19 +186,30 @@ class InstrumentedModel(WrapperModel):
     @staticmethod
     def messages_to_otel_events(messages: list[ModelMessage]) -> list[Event]:
         result: list[Event] = []
-        for message in messages:
+        for message_index, message in enumerate(messages):
+            message_events: list[Event] = []
             if isinstance(message, ModelRequest):
                 for part in message.parts:
                     if hasattr(part, 'otel_event'):
-                        result.append(part.otel_event())
+                        message_events.append(part.otel_event())
             elif isinstance(message, ModelResponse):
-                result.extend(message.otel_events())
+                message_events = message.otel_events()
+            for event in message_events:
+                event.attributes = {
+                    'gen_ai.message.index': message_index,
+                    **(event.attributes or {}),
+                }
+            result.extend(message_events)
         for event in result:
-            try:
-                event.body = ANY_ADAPTER.dump_python(event.body, mode='json')
-            except Exception:
-                try:
-                    event.body = str(event.body)
-                except Exception:
-                    event.body = 'Unable to serialize event body'
+            event.body = InstrumentedModel.serialize_any(event.body)
         return result
+
+    @staticmethod
+    def serialize_any(value: Any) -> str:
+        try:
+            return ANY_ADAPTER.dump_python(value, mode='json')
+        except Exception:
+            try:
+                return str(value)
+            except Exception as e:
+                return f'Unable to serialize: {e}'

pydantic_ai/models/openai.py

@@ -9,7 +9,9 @@ from datetime import datetime, timezone
 from typing import Literal, Union, cast, overload
 
 from httpx import AsyncClient as AsyncHTTPClient
-from typing_extensions import assert_never
+from typing_extensions import assert_never, deprecated
+
+from pydantic_ai.providers import Provider, infer_provider
 
 from .. import ModelHTTPError, UnexpectedModelBehavior, _utils, usage
 from .._utils import guard_tool_call_id as _guard_tool_call_id
@@ -98,10 +100,36 @@ class OpenAIModel(Model):
     _model_name: OpenAIModelName = field(repr=False)
     _system: str | None = field(repr=False)
 
+    @overload
     def __init__(
         self,
         model_name: OpenAIModelName,
         *,
+        provider: Literal['openai', 'deepseek'] | Provider[AsyncOpenAI] = 'openai',
+        system_prompt_role: OpenAISystemPromptRole | None = None,
+        system: str | None = 'openai',
+    ) -> None: ...
+
+    @deprecated('Use the `provider` parameter instead of `base_url`, `api_key`, `openai_client` and `http_client`.')
+    @overload
+    def __init__(
+        self,
+        model_name: OpenAIModelName,
+        *,
+        provider: None = None,
+        base_url: str | None = None,
+        api_key: str | None = None,
+        openai_client: AsyncOpenAI | None = None,
+        http_client: AsyncHTTPClient | None = None,
+        system_prompt_role: OpenAISystemPromptRole | None = None,
+        system: str | None = 'openai',
+    ) -> None: ...
+
+    def __init__(
+        self,
+        model_name: OpenAIModelName,
+        *,
+        provider: Literal['openai', 'deepseek'] | Provider[AsyncOpenAI] | None = None,
         base_url: str | None = None,
         api_key: str | None = None,
         openai_client: AsyncOpenAI | None = None,
@@ -115,6 +143,7 @@ class OpenAIModel(Model):
             model_name: The name of the OpenAI model to use. List of model names available
                 [here](https://github.com/openai/openai-python/blob/v1.54.3/src/openai/types/chat_model.py#L7)
                 (Unfortunately, despite being ask to do so, OpenAI do not provide `.inv` files for their API).
+            provider: The provider to use. Defaults to `'openai'`.
             base_url: The base url for the OpenAI requests. If not provided, the `OPENAI_BASE_URL` environment variable
                 will be used if available. Otherwise, defaults to OpenAI's base url.
             api_key: The API key to use for authentication, if not provided, the `OPENAI_API_KEY` environment variable
@@ -129,20 +158,32 @@ class OpenAIModel(Model):
                 customize the `base_url` and `api_key` to use a different provider.
         """
         self._model_name = model_name
-        # This is a workaround for the OpenAI client requiring an API key, whilst locally served,
-        # openai compatible models do not always need an API key, but a placeholder (non-empty) key is required.
-        if api_key is None and 'OPENAI_API_KEY' not in os.environ and base_url is not None and openai_client is None:
-            api_key = 'api-key-not-set'
-
-        if openai_client is not None:
-            assert http_client is None, 'Cannot provide both `openai_client` and `http_client`'
-            assert base_url is None, 'Cannot provide both `openai_client` and `base_url`'
-            assert api_key is None, 'Cannot provide both `openai_client` and `api_key`'
-            self.client = openai_client
-        elif http_client is not None:
-            self.client = AsyncOpenAI(base_url=base_url, api_key=api_key, http_client=http_client)
-        else:
-            self.client = AsyncOpenAI(base_url=base_url, api_key=api_key, http_client=cached_async_http_client())
+
+        if provider is not None:
+            if isinstance(provider, str):
+                self.client = infer_provider(provider).client
+            else:
+                self.client = provider.client
+        else:  # pragma: no cover
+            # This is a workaround for the OpenAI client requiring an API key, whilst locally served,
+            # openai compatible models do not always need an API key, but a placeholder (non-empty) key is required.
+            if (
+                api_key is None
+                and 'OPENAI_API_KEY' not in os.environ
+                and base_url is not None
+                and openai_client is None
+            ):
+                api_key = 'api-key-not-set'
+
+            if openai_client is not None:
+                assert http_client is None, 'Cannot provide both `openai_client` and `http_client`'
+                assert base_url is None, 'Cannot provide both `openai_client` and `base_url`'
+                assert api_key is None, 'Cannot provide both `openai_client` and `api_key`'
+                self.client = openai_client
+            elif http_client is not None:
+                self.client = AsyncOpenAI(base_url=base_url, api_key=api_key, http_client=http_client)
+            else:
+                self.client = AsyncOpenAI(base_url=base_url, api_key=api_key, http_client=cached_async_http_client())
         self.system_prompt_role = system_prompt_role
         self._system = system
 

pydantic_ai/models/vertexai.py

@@ -1,5 +1,6 @@
 from __future__ import annotations as _annotations
 
+import warnings
 from collections.abc import AsyncIterator
 from contextlib import asynccontextmanager
 from dataclasses import dataclass, field
@@ -8,6 +9,7 @@ from pathlib import Path
 from typing import Literal
 
 from httpx import AsyncClient as AsyncHTTPClient
+from typing_extensions import deprecated
 
 from .. import usage
 from .._utils import run_in_executor
@@ -55,6 +57,7 @@ The template is used thus:
 """
 
 
+@deprecated('Please use `GeminiModel(provider=GoogleVertexProvider(...))` instead.')
 @dataclass(init=False)
 class VertexAIModel(GeminiModel):
     """A model that uses Gemini via the `*-aiplatform.googleapis.com` VertexAI API."""
@@ -103,11 +106,16 @@ class VertexAIModel(GeminiModel):
         self.project_id = project_id
         self.region = region
         self.model_publisher = model_publisher
-        self.http_client = http_client or cached_async_http_client()
+        self.client = http_client or cached_async_http_client()
         self.url_template = url_template
 
         self._auth = None
         self._url = None
+        warnings.warn(
+            'VertexAIModel is deprecated, please use `GeminiModel(provider=GoogleVertexProvider(...))` instead.',
+            DeprecationWarning,
+        )
+        self._provider = None
 
     async def ainit(self) -> None:
         """Initialize the model, setting the URL and auth.

pydantic_ai/providers/__init__.py

@@ -0,0 +1,64 @@
+"""Providers for the API clients.
+
+The providers are in charge of providing an authenticated client to the API.
+"""
+
+from __future__ import annotations as _annotations
+
+from abc import ABC, abstractmethod
+from typing import Any, Generic, TypeVar
+
+InterfaceClient = TypeVar('InterfaceClient')
+
+
+class Provider(ABC, Generic[InterfaceClient]):
+    """Abstract class for a provider.
+
+    The provider is in charge of providing an authenticated client to the API.
+
+    Each provider only supports a specific interface. A interface can be supported by multiple providers.
+
+    For example, the OpenAIModel interface can be supported by the OpenAIProvider and the DeepSeekProvider.
+    """
+
+    _client: InterfaceClient
+
+    @property
+    @abstractmethod
+    def name(self) -> str:
+        """The provider name."""
+        raise NotImplementedError()
+
+    @property
+    @abstractmethod
+    def base_url(self) -> str:
+        """The base URL for the provider API."""
+        raise NotImplementedError()
+
+    @property
+    @abstractmethod
+    def client(self) -> InterfaceClient:
+        """The client for the provider."""
+        raise NotImplementedError()
+
+
+def infer_provider(provider: str) -> Provider[Any]:
+    """Infer the provider from the provider name."""
+    if provider == 'openai':
+        from .openai import OpenAIProvider
+
+        return OpenAIProvider()
+    elif provider == 'deepseek':
+        from .deepseek import DeepSeekProvider
+
+        return DeepSeekProvider()
+    elif provider == 'google-vertex':
+        from .google_vertex import GoogleVertexProvider
+
+        return GoogleVertexProvider()
+    elif provider == 'google-gla':
+        from .google_gla import GoogleGLAProvider
+
+        return GoogleGLAProvider()
+    else:  # pragma: no cover
+        raise ValueError(f'Unknown provider: {provider}')

pydantic_ai/providers/deepseek.py

@@ -0,0 +1,68 @@
+from __future__ import annotations as _annotations
+
+import os
+from typing import overload
+
+from httpx import AsyncClient as AsyncHTTPClient
+from openai import AsyncOpenAI
+
+from pydantic_ai.models import cached_async_http_client
+
+try:
+    from openai import AsyncOpenAI
+except ImportError as _import_error:  # pragma: no cover
+    raise ImportError(
+        'Please install `openai` to use the DeepSeek provider, '
+        "you can use the `openai` optional group — `pip install 'pydantic-ai-slim[openai]'`"
+    ) from _import_error
+
+from . import Provider
+
+
+class DeepSeekProvider(Provider[AsyncOpenAI]):
+    """Provider for DeepSeek API."""
+
+    @property
+    def name(self) -> str:
+        return 'deepseek'
+
+    @property
+    def base_url(self) -> str:
+        return 'https://api.deepseek.com'
+
+    @property
+    def client(self) -> AsyncOpenAI:
+        return self._client
+
+    @overload
+    def __init__(self) -> None: ...
+
+    @overload
+    def __init__(self, *, api_key: str) -> None: ...
+
+    @overload
+    def __init__(self, *, api_key: str, http_client: AsyncHTTPClient) -> None: ...
+
+    @overload
+    def __init__(self, *, openai_client: AsyncOpenAI | None = None) -> None: ...
+
+    def __init__(
+        self,
+        *,
+        api_key: str | None = None,
+        openai_client: AsyncOpenAI | None = None,
+        http_client: AsyncHTTPClient | None = None,
+    ) -> None:
+        api_key = api_key or os.getenv('DEEPSEEK_API_KEY')
+        if api_key is None and openai_client is None:
+            raise ValueError(
+                'Set the `DEEPSEEK_API_KEY` environment variable or pass it via `DeepSeekProvider(api_key=...)`'
+                'to use the DeepSeek provider.'
+            )
+
+        if openai_client is not None:
+            self._client = openai_client
+        elif http_client is not None:
+            self._client = AsyncOpenAI(base_url=self.base_url, api_key=api_key, http_client=http_client)
+        else:
+            self._client = AsyncOpenAI(base_url=self.base_url, api_key=api_key, http_client=cached_async_http_client())

pydantic_ai/providers/google_gla.py

@@ -0,0 +1,44 @@
+from __future__ import annotations as _annotations
+
+import os
+
+import httpx
+
+from pydantic_ai.models import cached_async_http_client
+from pydantic_ai.providers import Provider
+
+
+class GoogleGLAProvider(Provider[httpx.AsyncClient]):
+    """Provider for Google Generative Language AI API."""
+
+    @property
+    def name(self):
+        return 'google-gla'
+
+    @property
+    def base_url(self) -> str:
+        return 'https://generativelanguage.googleapis.com/v1beta/models/'
+
+    @property
+    def client(self) -> httpx.AsyncClient:
+        return self._client
+
+    def __init__(self, api_key: str | None = None, http_client: httpx.AsyncClient | None = None) -> None:
+        """Create a new Google GLA provider.
+
+        Args:
+            api_key: The API key to use for authentication, if not provided, the `GEMINI_API_KEY` environment variable
+                will be used if available.
+            http_client: An existing `httpx.AsyncClient` to use for making HTTP requests.
+        """
+        api_key = api_key or os.environ.get('GEMINI_API_KEY')
+        if api_key is None:
+            raise ValueError(
+                'Set the `GEMINI_API_KEY` environment variable or pass it via `GoogleGLAProvider(api_key=...)`'
+                'to use the Google GLA provider.'
+            )
+
+        self._client = http_client or cached_async_http_client()
+        self._client.base_url = self.base_url
+        # https://cloud.google.com/docs/authentication/api-keys-use#using-with-rest
+        self._client.headers['X-Goog-Api-Key'] = api_key

pydantic_ai/providers/google_vertex.py

@@ -0,0 +1,200 @@
+from __future__ import annotations as _annotations
+
+import functools
+from collections.abc import AsyncGenerator
+from datetime import datetime, timedelta
+from pathlib import Path
+from typing import Literal
+
+import anyio.to_thread
+import httpx
+
+from pydantic_ai.exceptions import UserError
+
+from ..models import cached_async_http_client
+from . import Provider
+
+try:
+    import google.auth
+    from google.auth.credentials import Credentials as BaseCredentials
+    from google.auth.transport.requests import Request
+    from google.oauth2.service_account import Credentials as ServiceAccountCredentials
+except ImportError as _import_error:
+    raise ImportError(
+        'Please install `google-auth` to use the Google Vertex AI provider, '
+        "you can use the `vertexai` optional group — `pip install 'pydantic-ai-slim[vertexai]'`"
+    ) from _import_error
+
+
+__all__ = ('GoogleVertexProvider',)
+
+# default expiry is 3600 seconds
+MAX_TOKEN_AGE = timedelta(seconds=3000)
+
+
+class GoogleVertexProvider(Provider[httpx.AsyncClient]):
+    """Provider for Vertex AI API."""
+
+    @property
+    def name(self) -> str:
+        return 'google-vertex'
+
+    @property
+    def base_url(self) -> str:
+        return (
+            f'https://{self.region}-aiplatform.googleapis.com/v1'
+            f'/projects/{self.project_id}'
+            f'/locations/{self.region}'
+            f'/publishers/{self.model_publisher}/models/'
+        )
+
+    @property
+    def client(self) -> httpx.AsyncClient:
+        return self._client
+
+    def __init__(
+        self,
+        service_account_file: Path | str | None = None,
+        project_id: str | None = None,
+        region: VertexAiRegion = 'us-central1',
+        model_publisher: str = 'google',
+        http_client: httpx.AsyncClient | None = None,
+    ) -> None:
+        """Create a new Vertex AI provider.
+
+        Args:
+            service_account_file: Path to a service account file.
+                If not provided, the default environment credentials will be used.
+            project_id: The project ID to use, if not provided it will be taken from the credentials.
+            region: The region to make requests to.
+            model_publisher: The model publisher to use, I couldn't find a good list of available publishers,
+                and from trial and error it seems non-google models don't work with the `generateContent` and
+                `streamGenerateContent` functions, hence only `google` is currently supported.
+                Please create an issue or PR if you know how to use other publishers.
+            http_client: An existing `httpx.AsyncClient` to use for making HTTP requests.
+        """
+        self._client = http_client or cached_async_http_client()
+        self.service_account_file = service_account_file
+        self.project_id = project_id
+        self.region = region
+        self.model_publisher = model_publisher
+
+        self._client.auth = _VertexAIAuth(service_account_file, project_id, region)
+        self._client.base_url = self.base_url
+
+
+class _VertexAIAuth(httpx.Auth):
+    """Auth class for Vertex AI API."""
+
+    credentials: BaseCredentials | ServiceAccountCredentials | None
+
+    def __init__(
+        self,
+        service_account_file: Path | str | None = None,
+        project_id: str | None = None,
+        region: VertexAiRegion = 'us-central1',
+    ) -> None:
+        self.service_account_file = service_account_file
+        self.project_id = project_id
+        self.region = region
+
+        self.credentials = None
+        self.token_created: datetime | None = None
+
+    async def async_auth_flow(self, request: httpx.Request) -> AsyncGenerator[httpx.Request, httpx.Response]:
+        if self.credentials is None:
+            self.credentials = await self._get_credentials()
+        if self.credentials.token is None or self._token_expired():  # type: ignore[reportUnknownMemberType]
+            await anyio.to_thread.run_sync(self._refresh_token)
+            self.token_created = datetime.now()
+        request.headers['Authorization'] = f'Bearer {self.credentials.token}'  # type: ignore[reportUnknownMemberType]
+
+        # NOTE: This workaround is in place because we might get the project_id from the credentials.
+        request.url = httpx.URL(str(request.url).replace('projects/None', f'projects/{self.project_id}'))
+        yield request
+
+    async def _get_credentials(self) -> BaseCredentials | ServiceAccountCredentials:
+        if self.service_account_file is not None:
+            creds = await _creds_from_file(self.service_account_file)
+            assert creds.project_id is None or isinstance(creds.project_id, str)  # type: ignore[reportUnknownMemberType]
+            creds_project_id: str | None = creds.project_id
+            creds_source = 'service account file'
+        else:
+            creds, creds_project_id = await _async_google_auth()
+            creds_source = '`google.auth.default()`'
+
+        if self.project_id is None:
+            if creds_project_id is None:
+                raise UserError(f'No project_id provided and none found in {creds_source}')
+            self.project_id = creds_project_id
+        return creds
+
+    def _token_expired(self) -> bool:
+        if self.token_created is None:
+            return True
+        else:
+            return (datetime.now() - self.token_created) > MAX_TOKEN_AGE
+
+    def _refresh_token(self) -> str:  # pragma: no cover
+        assert self.credentials is not None
+        self.credentials.refresh(Request())  # type: ignore[reportUnknownMemberType]
+        assert isinstance(self.credentials.token, str), f'Expected token to be a string, got {self.credentials.token}'  # type: ignore[reportUnknownMemberType]
+        return self.credentials.token
+
+
+async def _async_google_auth() -> tuple[BaseCredentials, str | None]:
+    return await anyio.to_thread.run_sync(google.auth.default, ['https://www.googleapis.com/auth/cloud-platform'])  # type: ignore
+
+
+async def _creds_from_file(service_account_file: str | Path) -> ServiceAccountCredentials:
+    service_account_credentials_from_file = functools.partial(
+        ServiceAccountCredentials.from_service_account_file,  # type: ignore[reportUnknownMemberType]
+        scopes=['https://www.googleapis.com/auth/cloud-platform'],
+    )
+    return await anyio.to_thread.run_sync(service_account_credentials_from_file, str(service_account_file))
+
+
+VertexAiRegion = Literal[
+    'us-central1',
+    'us-east1',
+    'us-east4',
+    'us-south1',
+    'us-west1',
+    'us-west2',
+    'us-west3',
+    'us-west4',
+    'us-east5',
+    'europe-central2',
+    'europe-north1',
+    'europe-southwest1',
+    'europe-west1',
+    'europe-west2',
+    'europe-west3',
+    'europe-west4',
+    'europe-west6',
+    'europe-west8',
+    'europe-west9',
+    'europe-west12',
+    'africa-south1',
+    'asia-east1',
+    'asia-east2',
+    'asia-northeast1',
+    'asia-northeast2',
+    'asia-northeast3',
+    'asia-south1',
+    'asia-southeast1',
+    'asia-southeast2',
+    'australia-southeast1',
+    'australia-southeast2',
+    'me-central1',
+    'me-central2',
+    'me-west1',
+    'northamerica-northeast1',
+    'northamerica-northeast2',
+    'southamerica-east1',
+    'southamerica-west1',
+]
+"""Regions available for Vertex AI.
+
+More details [here](https://cloud.google.com/vertex-ai/docs/reference/rest#rest_endpoints).
+"""

pydantic_ai/providers/openai.py

@@ -0,0 +1,72 @@
+from __future__ import annotations as _annotations
+
+import os
+from typing import TypeVar
+
+import httpx
+
+from pydantic_ai.models import cached_async_http_client
+
+try:
+    from openai import AsyncOpenAI
+except ImportError as _import_error:  # pragma: no cover
+    raise ImportError(
+        'Please install `openai` to use the OpenAI provider, '
+        "you can use the `openai` optional group — `pip install 'pydantic-ai-slim[openai]'`"
+    ) from _import_error
+
+
+from . import Provider
+
+InterfaceClient = TypeVar('InterfaceClient')
+
+
+class OpenAIProvider(Provider[AsyncOpenAI]):
+    """Provider for OpenAI API."""
+
+    @property
+    def name(self) -> str:
+        return 'openai'  # pragma: no cover
+
+    @property
+    def base_url(self) -> str:
+        return self._base_url
+
+    @property
+    def client(self) -> AsyncOpenAI:
+        return self._client
+
+    def __init__(
+        self,
+        base_url: str | None = None,
+        api_key: str | None = None,
+        openai_client: AsyncOpenAI | None = None,
+        http_client: httpx.AsyncClient | None = None,
+    ) -> None:
+        """Create a new OpenAI provider.
+
+        Args:
+            base_url: The base url for the OpenAI requests. If not provided, the `OPENAI_BASE_URL` environment variable
+                will be used if available. Otherwise, defaults to OpenAI's base url.
+            api_key: The API key to use for authentication, if not provided, the `OPENAI_API_KEY` environment variable
+                will be used if available.
+            openai_client: An existing
+                [`AsyncOpenAI`](https://github.com/openai/openai-python?tab=readme-ov-file#async-usage)
+                client to use. If provided, `base_url`, `api_key`, and `http_client` must be `None`.
+            http_client: An existing `httpx.AsyncClient` to use for making HTTP requests.
+        """
+        self._base_url = base_url or 'https://api.openai.com/v1'
+        # This is a workaround for the OpenAI client requiring an API key, whilst locally served,
+        # openai compatible models do not always need an API key, but a placeholder (non-empty) key is required.
+        if api_key is None and 'OPENAI_API_KEY' not in os.environ and openai_client is None:
+            api_key = 'api-key-not-set'
+
+        if openai_client is not None:
+            assert base_url is None, 'Cannot provide both `openai_client` and `base_url`'
+            assert http_client is None, 'Cannot provide both `openai_client` and `http_client`'
+            assert api_key is None, 'Cannot provide both `openai_client` and `api_key`'
+            self._client = openai_client
+        elif http_client is not None:
+            self._client = AsyncOpenAI(base_url=self.base_url, api_key=api_key, http_client=http_client)
+        else:
+            self._client = AsyncOpenAI(base_url=self.base_url, api_key=api_key, http_client=cached_async_http_client())