pydantic-ai-slim 0.7.5__py3-none-any.whl → 0.8.0__py3-none-any.whl

pydantic_ai/providers/mistral.py

@@ -3,7 +3,7 @@ from __future__ import annotations as _annotations
 import os
 from typing import overload
 
-from httpx import AsyncClient as AsyncHTTPClient
+import httpx
 
 from pydantic_ai.exceptions import UserError
 from pydantic_ai.models import cached_async_http_client
@@ -42,7 +42,7 @@ class MistralProvider(Provider[Mistral]):
     def __init__(self, *, mistral_client: Mistral | None = None) -> None: ...
 
     @overload
-    def __init__(self, *, api_key: str | None = None, http_client: AsyncHTTPClient | None = None) -> None: ...
+    def __init__(self, *, api_key: str | None = None, http_client: httpx.AsyncClient | None = None) -> None: ...
 
     def __init__(
         self,
@@ -50,7 +50,7 @@ class MistralProvider(Provider[Mistral]):
         api_key: str | None = None,
         mistral_client: Mistral | None = None,
         base_url: str | None = None,
-        http_client: AsyncHTTPClient | None = None,
+        http_client: httpx.AsyncClient | None = None,
     ) -> None:
         """Create a new Mistral provider.
 

pydantic_ai/providers/moonshotai.py

@@ -3,7 +3,7 @@ from __future__ import annotations as _annotations
 import os
 from typing import Literal, overload
 
-from httpx import AsyncClient as AsyncHTTPClient
+import httpx
 from openai import AsyncOpenAI
 
 from pydantic_ai.exceptions import UserError
@@ -59,9 +59,6 @@ class MoonshotAIProvider(Provider[AsyncOpenAI]):
             supports_json_object_output=True,
         ).update(profile)
 
-    # ---------------------------------------------------------------------
-    # Construction helpers
-    # ---------------------------------------------------------------------
     @overload
     def __init__(self) -> None: ...
 
@@ -69,7 +66,7 @@ class MoonshotAIProvider(Provider[AsyncOpenAI]):
     def __init__(self, *, api_key: str) -> None: ...
 
     @overload
-    def __init__(self, *, api_key: str, http_client: AsyncHTTPClient) -> None: ...
+    def __init__(self, *, api_key: str, http_client: httpx.AsyncClient) -> None: ...
 
     @overload
     def __init__(self, *, openai_client: AsyncOpenAI | None = None) -> None: ...
@@ -79,7 +76,7 @@ class MoonshotAIProvider(Provider[AsyncOpenAI]):
         *,
         api_key: str | None = None,
         openai_client: AsyncOpenAI | None = None,
-        http_client: AsyncHTTPClient | None = None,
+        http_client: httpx.AsyncClient | None = None,
     ) -> None:
         api_key = api_key or os.getenv('MOONSHOTAI_API_KEY')
         if not api_key and openai_client is None:

pydantic_ai/providers/ollama.py

@@ -58,7 +58,7 @@ class OllamaProvider(Provider[AsyncOpenAI]):
             if model_name.startswith(prefix):
                 profile = profile_func(model_name)
 
-        # As OllamaProvider is always used with OpenAIModel, which used to unconditionally use OpenAIJsonSchemaTransformer,
+        # As OllamaProvider is always used with OpenAIChatModel, which used to unconditionally use OpenAIJsonSchemaTransformer,
         # we need to maintain that behavior unless json_schema_transformer is set explicitly
         return OpenAIModelProfile(json_schema_transformer=OpenAIJsonSchemaTransformer).update(profile)
 

pydantic_ai/providers/openrouter.py

@@ -3,7 +3,7 @@ from __future__ import annotations as _annotations
 import os
 from typing import overload
 
-from httpx import AsyncClient as AsyncHTTPClient
+import httpx
 from openai import AsyncOpenAI
 
 from pydantic_ai.exceptions import UserError
@@ -68,7 +68,7 @@ class OpenRouterProvider(Provider[AsyncOpenAI]):
             model_name, *_ = model_name.split(':', 1)  # drop tags
             profile = provider_to_profile[provider](model_name)
 
-        # As OpenRouterProvider is always used with OpenAIModel, which used to unconditionally use OpenAIJsonSchemaTransformer,
+        # As OpenRouterProvider is always used with OpenAIChatModel, which used to unconditionally use OpenAIJsonSchemaTransformer,
         # we need to maintain that behavior unless json_schema_transformer is set explicitly
         return OpenAIModelProfile(json_schema_transformer=OpenAIJsonSchemaTransformer).update(profile)
 
@@ -79,7 +79,7 @@ class OpenRouterProvider(Provider[AsyncOpenAI]):
     def __init__(self, *, api_key: str) -> None: ...
 
     @overload
-    def __init__(self, *, api_key: str, http_client: AsyncHTTPClient) -> None: ...
+    def __init__(self, *, api_key: str, http_client: httpx.AsyncClient) -> None: ...
 
     @overload
     def __init__(self, *, openai_client: AsyncOpenAI | None = None) -> None: ...
@@ -89,7 +89,7 @@ class OpenRouterProvider(Provider[AsyncOpenAI]):
         *,
         api_key: str | None = None,
         openai_client: AsyncOpenAI | None = None,
-        http_client: AsyncHTTPClient | None = None,
+        http_client: httpx.AsyncClient | None = None,
     ) -> None:
         api_key = api_key or os.getenv('OPENROUTER_API_KEY')
         if not api_key and openai_client is None:

pydantic_ai/providers/together.py

@@ -3,7 +3,7 @@ from __future__ import annotations as _annotations
 import os
 from typing import overload
 
-from httpx import AsyncClient as AsyncHTTPClient
+import httpx
 from openai import AsyncOpenAI
 
 from pydantic_ai.exceptions import UserError
@@ -68,7 +68,7 @@ class TogetherProvider(Provider[AsyncOpenAI]):
     def __init__(self, *, api_key: str) -> None: ...
 
     @overload
-    def __init__(self, *, api_key: str, http_client: AsyncHTTPClient) -> None: ...
+    def __init__(self, *, api_key: str, http_client: httpx.AsyncClient) -> None: ...
 
     @overload
     def __init__(self, *, openai_client: AsyncOpenAI | None = None) -> None: ...
@@ -78,7 +78,7 @@ class TogetherProvider(Provider[AsyncOpenAI]):
         *,
         api_key: str | None = None,
         openai_client: AsyncOpenAI | None = None,
-        http_client: AsyncHTTPClient | None = None,
+        http_client: httpx.AsyncClient | None = None,
     ) -> None:
         api_key = api_key or os.getenv('TOGETHER_API_KEY')
         if not api_key and openai_client is None:

pydantic_ai/providers/vercel.py

@@ -3,7 +3,7 @@ from __future__ import annotations as _annotations
 import os
 from typing import overload
 
-from httpx import AsyncClient as AsyncHTTPClient
+import httpx
 
 from pydantic_ai.exceptions import UserError
 from pydantic_ai.models import cached_async_http_client
@@ -64,7 +64,7 @@ class VercelProvider(Provider[AsyncOpenAI]):
         if provider in provider_to_profile:
             profile = provider_to_profile[provider](model_name)
 
-        # As VercelProvider is always used with OpenAIModel, which used to unconditionally use OpenAIJsonSchemaTransformer,
+        # As VercelProvider is always used with OpenAIChatModel, which used to unconditionally use OpenAIJsonSchemaTransformer,
         # we need to maintain that behavior unless json_schema_transformer is set explicitly
         return OpenAIModelProfile(
             json_schema_transformer=OpenAIJsonSchemaTransformer,
@@ -77,7 +77,7 @@ class VercelProvider(Provider[AsyncOpenAI]):
     def __init__(self, *, api_key: str) -> None: ...
 
     @overload
-    def __init__(self, *, api_key: str, http_client: AsyncHTTPClient) -> None: ...
+    def __init__(self, *, api_key: str, http_client: httpx.AsyncClient) -> None: ...
 
     @overload
     def __init__(self, *, openai_client: AsyncOpenAI | None = None) -> None: ...
@@ -87,7 +87,7 @@ class VercelProvider(Provider[AsyncOpenAI]):
         *,
         api_key: str | None = None,
         openai_client: AsyncOpenAI | None = None,
-        http_client: AsyncHTTPClient | None = None,
+        http_client: httpx.AsyncClient | None = None,
     ) -> None:
         # Support Vercel AI Gateway's standard environment variables
         api_key = api_key or os.getenv('VERCEL_AI_GATEWAY_API_KEY') or os.getenv('VERCEL_OIDC_TOKEN')

pydantic_ai/result.py

@@ -22,7 +22,7 @@ from ._output import (
     ToolOutputSchema,
 )
 from ._run_context import AgentDepsT, RunContext
-from .messages import AgentStreamEvent
+from .messages import ModelResponseStreamEvent
 from .output import (
     OutputDataT,
     ToolOutput,
@@ -51,7 +51,7 @@ class AgentStream(Generic[AgentDepsT, OutputDataT]):
     _usage_limits: UsageLimits | None
     _tool_manager: ToolManager[AgentDepsT]
 
-    _agent_stream_iterator: AsyncIterator[AgentStreamEvent] | None = field(default=None, init=False)
+    _agent_stream_iterator: AsyncIterator[ModelResponseStreamEvent] | None = field(default=None, init=False)
     _initial_run_ctx_usage: RunUsage = field(init=False)
 
     def __post_init__(self):
@@ -221,8 +221,8 @@ class AgentStream(Generic[AgentDepsT, OutputDataT]):
                 deltas.append(text)
                 yield ''.join(deltas)
 
-    def __aiter__(self) -> AsyncIterator[AgentStreamEvent]:
-        """Stream [`AgentStreamEvent`][pydantic_ai.messages.AgentStreamEvent]s."""
+    def __aiter__(self) -> AsyncIterator[ModelResponseStreamEvent]:
+        """Stream [`ModelResponseStreamEvent`][pydantic_ai.messages.ModelResponseStreamEvent]s."""
         if self._agent_stream_iterator is None:
             self._agent_stream_iterator = _get_usage_checking_stream_response(
                 self._raw_stream_response, self._usage_limits, self.usage
@@ -426,7 +426,7 @@ def _get_usage_checking_stream_response(
     stream_response: models.StreamedResponse,
     limits: UsageLimits | None,
     get_usage: Callable[[], RunUsage],
-) -> AsyncIterator[AgentStreamEvent]:
+) -> AsyncIterator[ModelResponseStreamEvent]:
     if limits is not None and limits.has_token_limits():
 
         async def _usage_checking_iterator():

pydantic_ai/retries.py

@@ -13,25 +13,103 @@ The module includes:
 
 from __future__ import annotations
 
-from httpx import AsyncBaseTransport, AsyncHTTPTransport, BaseTransport, HTTPTransport, Request, Response
+from httpx import (
+    AsyncBaseTransport,
+    AsyncHTTPTransport,
+    BaseTransport,
+    HTTPStatusError,
+    HTTPTransport,
+    Request,
+    Response,
+)
 
 try:
-    from tenacity import AsyncRetrying, Retrying
+    from tenacity import AsyncRetrying, RetryCallState, RetryError, Retrying, retry, wait_exponential
 except ImportError as _import_error:
     raise ImportError(
         'Please install `tenacity` to use the retries utilities, '
         'you can use the `retries` optional group — `pip install "pydantic-ai-slim[retries]"`'
     ) from _import_error
 
-
-__all__ = ['TenacityTransport', 'AsyncTenacityTransport', 'wait_retry_after']
-
+from collections.abc import Awaitable
 from datetime import datetime, timezone
 from email.utils import parsedate_to_datetime
-from typing import Callable, cast
+from typing import TYPE_CHECKING, Any, Callable, NoReturn, cast
+
+from typing_extensions import TypedDict
+
+if TYPE_CHECKING:
+    from tenacity.asyncio.retry import RetryBaseT
+    from tenacity.retry import RetryBaseT as SyncRetryBaseT
+    from tenacity.stop import StopBaseT
+    from tenacity.wait import WaitBaseT
+
+__all__ = ['RetryConfig', 'TenacityTransport', 'AsyncTenacityTransport', 'wait_retry_after']
+
+
+class RetryConfig(TypedDict, total=False):
+    """The configuration for tenacity-based retrying.
+
+    These are precisely the arguments to the tenacity `retry` decorator, and they are generally
+    used internally by passing them to that decorator via `@retry(**config)` or similar.
+
+    All fields are optional, and if not provided, the default values from the `tenacity.retry` decorator will be used.
+    """
+
+    sleep: Callable[[int | float], None | Awaitable[None]]
+    """A sleep strategy to use for sleeping between retries.
+
+    Tenacity's default for this argument is `tenacity.nap.sleep`."""
+
+    stop: StopBaseT
+    """
+    A stop strategy to determine when to stop retrying.
+
+    Tenacity's default for this argument is `tenacity.stop.stop_never`."""
+
+    wait: WaitBaseT
+    """
+    A wait strategy to determine how long to wait between retries.
+
+    Tenacity's default for this argument is `tenacity.wait.wait_none`."""
+
+    retry: SyncRetryBaseT | RetryBaseT
+    """A retry strategy to determine which exceptions should trigger a retry.
+
+    Tenacity's default for this argument is `tenacity.retry.retry_if_exception_type()`."""
+
+    before: Callable[[RetryCallState], None | Awaitable[None]]
+    """
+    A callable that is called before each retry attempt.
 
-from httpx import HTTPStatusError
-from tenacity import RetryCallState, wait_exponential
+    Tenacity's default for this argument is `tenacity.before.before_nothing`."""
+
+    after: Callable[[RetryCallState], None | Awaitable[None]]
+    """
+    A callable that is called after each retry attempt.
+
+    Tenacity's default for this argument is `tenacity.after.after_nothing`."""
+
+    before_sleep: Callable[[RetryCallState], None | Awaitable[None]] | None
+    """
+    An optional callable that is called before sleeping between retries.
+
+    Tenacity's default for this argument is `None`."""
+
+    reraise: bool
+    """Whether to reraise the last exception if the retry attempts are exhausted, or raise a RetryError instead.
+
+    Tenacity's default for this argument is `False`."""
+
+    retry_error_cls: type[RetryError]
+    """The exception class to raise when the retry attempts are exhausted and `reraise` is False.
+
+    Tenacity's default for this argument is `tenacity.RetryError`."""
+
+    retry_error_callback: Callable[[RetryCallState], Any | Awaitable[Any]] | None
+    """An optional callable that is called when the retry attempts are exhausted and `reraise` is False.
+
+    Tenacity's default for this argument is `None`."""
 
 
 class TenacityTransport(BaseTransport):
@@ -47,8 +125,8 @@ class TenacityTransport(BaseTransport):
 
     Args:
         wrapped: The underlying transport to wrap and add retry functionality to.
-        controller: The tenacity Retrying instance that defines the retry behavior
-                   (retry conditions, wait strategy, stop conditions, etc.).
+        config: The arguments to use for the tenacity `retry` decorator, including retry conditions,
+            wait strategy, stop conditions, etc. See the tenacity docs for more info.
         validate_response: Optional callable that takes a Response and can raise an exception
             to be handled by the controller if the response should trigger a retry.
             Common use case is to raise exceptions for certain HTTP status codes.
@@ -57,17 +135,17 @@ class TenacityTransport(BaseTransport):
     Example:
         ```python
         from httpx import Client, HTTPTransport, HTTPStatusError
-        from tenacity import Retrying, stop_after_attempt, retry_if_exception_type
-        from pydantic_ai.retries import TenacityTransport, wait_retry_after
+        from tenacity import stop_after_attempt, retry_if_exception_type
+        from pydantic_ai.retries import RetryConfig, TenacityTransport, wait_retry_after
 
         transport = TenacityTransport(
-            HTTPTransport(),
-            Retrying(
+            RetryConfig(
                 retry=retry_if_exception_type(HTTPStatusError),
                 wait=wait_retry_after(max_wait=300),
                 stop=stop_after_attempt(5),
                 reraise=True
             ),
+            HTTPTransport(),
             validate_response=lambda r: r.raise_for_status()
         )
         client = Client(transport=transport)
@@ -76,11 +154,22 @@ class TenacityTransport(BaseTransport):
 
     def __init__(
         self,
-        controller: Retrying,
+        config: RetryConfig,
         wrapped: BaseTransport | None = None,
-        validate_response: Callable[[Response], None] | None = None,
+        validate_response: Callable[[Response], Any] | None = None,
+        **kwargs: NoReturn,
     ):
-        self.controller = controller
+        # TODO: Remove the following checks (and **kwargs) during v1 release
+        if 'controller' in kwargs:  # pragma: no cover
+            raise TypeError('The `controller` argument has been renamed to `config`, and now requires a `RetryConfig`.')
+        if kwargs:  # pragma: no cover
+            raise TypeError(f'Unexpected keyword arguments: {", ".join(kwargs)}')
+        if isinstance(config, Retrying):  # pragma: no cover
+            raise ValueError(
+                'Passing a Retrying instance is no longer supported; the `config` argument must be a `pydantic_ai.retries.RetryConfig`.'
+            )
+
+        self.config = config
         self.wrapped = wrapped or HTTPTransport()
         self.validate_response = validate_response
 
@@ -97,13 +186,19 @@ class TenacityTransport(BaseTransport):
             RuntimeError: If the retry controller did not make any attempts.
             Exception: Any exception raised by the wrapped transport or validation function.
         """
-        for attempt in self.controller:
-            with attempt:
-                response = self.wrapped.handle_request(request)
-                if self.validate_response:
-                    self.validate_response(response)
-                return response
-        raise RuntimeError('The retry controller did not make any attempts')  # pragma: no cover
+
+        @retry(**self.config)
+        def handle_request(req: Request) -> Response:
+            response = self.wrapped.handle_request(req)
+
+            # this is normally set by httpx _after_ calling this function, but we want the request in the validator:
+            response.request = req
+
+            if self.validate_response:
+                self.validate_response(response)
+            return response
+
+        return handle_request(request)
 
 
 class AsyncTenacityTransport(AsyncBaseTransport):
@@ -119,8 +214,8 @@ class AsyncTenacityTransport(AsyncBaseTransport):
 
     Args:
         wrapped: The underlying async transport to wrap and add retry functionality to.
-        controller: The tenacity AsyncRetrying instance that defines the retry behavior
-                   (retry conditions, wait strategy, stop conditions, etc.).
+        config: The arguments to use for the tenacity `retry` decorator, including retry conditions,
+            wait strategy, stop conditions, etc. See the tenacity docs for more info.
         validate_response: Optional callable that takes a Response and can raise an exception
             to be handled by the controller if the response should trigger a retry.
             Common use case is to raise exceptions for certain HTTP status codes.
@@ -129,11 +224,11 @@ class AsyncTenacityTransport(AsyncBaseTransport):
     Example:
         ```python
         from httpx import AsyncClient, HTTPStatusError
-        from tenacity import AsyncRetrying, stop_after_attempt, retry_if_exception_type
-        from pydantic_ai.retries import AsyncTenacityTransport, wait_retry_after
+        from tenacity import stop_after_attempt, retry_if_exception_type
+        from pydantic_ai.retries import AsyncTenacityTransport, RetryConfig, wait_retry_after
 
         transport = AsyncTenacityTransport(
-            AsyncRetrying(
+            RetryConfig(
                 retry=retry_if_exception_type(HTTPStatusError),
                 wait=wait_retry_after(max_wait=300),
                 stop=stop_after_attempt(5),
@@ -147,11 +242,22 @@ class AsyncTenacityTransport(AsyncBaseTransport):
 
     def __init__(
         self,
-        controller: AsyncRetrying,
+        config: RetryConfig,
         wrapped: AsyncBaseTransport | None = None,
-        validate_response: Callable[[Response], None] | None = None,
+        validate_response: Callable[[Response], Any] | None = None,
+        **kwargs: NoReturn,
     ):
-        self.controller = controller
+        # TODO: Remove the following checks (and **kwargs) during v1 release
+        if 'controller' in kwargs:  # pragma: no cover
+            raise TypeError('The `controller` argument has been renamed to `config`, and now requires a `RetryConfig`.')
+        if kwargs:  # pragma: no cover
+            raise TypeError(f'Unexpected keyword arguments: {", ".join(kwargs)}')
+        if isinstance(config, AsyncRetrying):  # pragma: no cover
+            raise ValueError(
+                'Passing an AsyncRetrying instance is no longer supported; the `config` argument must be a `pydantic_ai.retries.RetryConfig`.'
+            )
+
+        self.config = config
         self.wrapped = wrapped or AsyncHTTPTransport()
         self.validate_response = validate_response
 
@@ -168,13 +274,19 @@ class AsyncTenacityTransport(AsyncBaseTransport):
             RuntimeError: If the retry controller did not make any attempts.
             Exception: Any exception raised by the wrapped transport or validation function.
         """
-        async for attempt in self.controller:
-            with attempt:
-                response = await self.wrapped.handle_async_request(request)
-                if self.validate_response:
-                    self.validate_response(response)
-                return response
-        raise RuntimeError('The retry controller did not make any attempts')  # pragma: no cover
+
+        @retry(**self.config)
+        async def handle_async_request(req: Request) -> Response:
+            response = await self.wrapped.handle_async_request(req)
+
+            # this is normally set by httpx _after_ calling this function, but we want the request in the validator:
+            response.request = req
+
+            if self.validate_response:
+                self.validate_response(response)
+            return response
+
+        return await handle_async_request(request)
 
 
 def wait_retry_after(
@@ -202,11 +314,11 @@ def wait_retry_after(
     Example:
         ```python
         from httpx import AsyncClient, HTTPStatusError
-        from tenacity import AsyncRetrying, stop_after_attempt, retry_if_exception_type
-        from pydantic_ai.retries import AsyncTenacityTransport, wait_retry_after
+        from tenacity import stop_after_attempt, retry_if_exception_type
+        from pydantic_ai.retries import AsyncTenacityTransport, RetryConfig, wait_retry_after
 
         transport = AsyncTenacityTransport(
-            AsyncRetrying(
+            RetryConfig(
                 retry=retry_if_exception_type(HTTPStatusError),
                 wait=wait_retry_after(max_wait=120),
                 stop=stop_after_attempt(5),

{pydantic_ai_slim-0.7.5.dist-info → pydantic_ai_slim-0.8.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pydantic-ai-slim
-Version: 0.7.5
+Version: 0.8.0
 Summary: Agent Framework / shim to use Pydantic with LLMs, slim package
 Project-URL: Homepage, https://github.com/pydantic/pydantic-ai/tree/main/pydantic_ai_slim
 Project-URL: Source, https://github.com/pydantic/pydantic-ai/tree/main/pydantic_ai_slim
@@ -35,7 +35,7 @@ Requires-Dist: genai-prices>=0.0.22
 Requires-Dist: griffe>=1.3.2
 Requires-Dist: httpx>=0.27
 Requires-Dist: opentelemetry-api>=1.28.0
-Requires-Dist: pydantic-graph==0.7.5
+Requires-Dist: pydantic-graph==0.8.0
 Requires-Dist: pydantic>=2.10
 Requires-Dist: typing-inspection>=0.4.0
 Provides-Extra: a2a
@@ -57,7 +57,7 @@ Requires-Dist: cohere>=5.16.0; (platform_system != 'Emscripten') and extra == 'c
 Provides-Extra: duckduckgo
 Requires-Dist: ddgs>=9.0.0; extra == 'duckduckgo'
 Provides-Extra: evals
-Requires-Dist: pydantic-evals==0.7.5; extra == 'evals'
+Requires-Dist: pydantic-evals==0.8.0; extra == 'evals'
 Provides-Extra: google
 Requires-Dist: google-genai>=1.31.0; extra == 'google'
 Provides-Extra: groq
@@ -67,7 +67,7 @@ Requires-Dist: huggingface-hub[inference]>=0.33.5; extra == 'huggingface'
 Provides-Extra: logfire
 Requires-Dist: logfire>=3.14.1; extra == 'logfire'
 Provides-Extra: mcp
-Requires-Dist: mcp>=1.10.0; (python_version >= '3.10') and extra == 'mcp'
+Requires-Dist: mcp>=1.12.3; (python_version >= '3.10') and extra == 'mcp'
 Provides-Extra: mistral
 Requires-Dist: mistralai>=1.9.2; extra == 'mistral'
 Provides-Extra: openai