mirror of
https://github.com/microsoft/semantic-kernel.git
synced 2026-10-12 11:19:56 +00:00
5e50e190c4
## Summary Adds reasoning capabilities to Python OpenAI ResponsesAgent, bringing parity with the C# implementation. This enables fine-grained control over reasoning effort for O-series models (o1, o3-mini, o4-mini) and gpt-5 that support the reasoning parameter. Fixes #12843 ## Changes ### Core Implementation - **`openai_responses_agent.py`**: Added reasoning configuration support - Constructor-level reasoning effort setting - Per-invocation reasoning effort override capability - Proper parameter validation and model compatibility checks - **`responses_agent_thread_actions.py`**: Extended thread actions to support reasoning parameters - Reasoning effort propagation through thread operations - Metadata preservation for reasoning tokens and summaries ### Sample and Tests - **`responses_agent_reasoning.py`**: Comprehensive demonstration sample - Constructor vs per-invocation reasoning configuration - Function calling integration with reasoning - Reasoning comparison scenarios (low/medium/high effort) - Error handling and troubleshooting guidance - **`test_openai_responses_agent_reasoning.py`**: Full unit test coverage - Parameter validation tests - Integration scenarios with function calling - Edge cases and error conditions ## Features ✅ **Constructor-Level Reasoning**: Set default reasoning effort when creating agents ✅ **Per-Invocation Override**: Override reasoning effort per request ✅ **Priority Hierarchy**: per-invocation > constructor > model default ✅ **Function Calling Compatible**: Works seamlessly with existing plugin system ✅ **Azure OpenAI & OpenAI Support**: Compatible with both service providers ✅ **Model Validation**: Automatic compatibility checks for O-series models nad GÜT-5 ✅ **Metadata Access**: Reasoning tokens and summaries available in response metadata ## Usage Example ```python # Constructor-level reasoning # Constructor-level reasoning configuration agent = AzureResponsesAgent( ai_model_id="gpt-5", client=client, reasoning={"effort": "low"} # Default reasoning for all requests ) # Per-invocation override response = await agent.invoke( "Solve this complex problem step by step", reasoning={"effort": "high"} ) # Invoke with reasoning callback to capture intermediate thoughts response = await agent.invoke( "Analyze this data step by step", reasoning={"effort": "high", "summary": "detailed"}, on_intermediate_message=handle_reasoning_message ) # Streaming with reasoning async for response in agent.invoke_stream( "Explain quantum computing in detail", reasoning={"effort": "high", "summary": "detailed"}, on_intermediate_message=handle_reasoning_message ): print(response.content, end="", flush=True) --------- Co-authored-by: Evan Mattson <[email protected]> Co-authored-by: Evan Mattson <[email protected]>
60 lines
2.3 KiB
Python
60 lines
2.3 KiB
Python
# Copyright (c) Microsoft. All rights reserved.
|
|
|
|
"""Reasoning content model."""
|
|
|
|
from html import unescape
|
|
from typing import ClassVar, Literal, TypeVar
|
|
from xml.etree.ElementTree import Element # nosec
|
|
|
|
from pydantic import Field
|
|
|
|
from semantic_kernel.contents.const import REASONING_CONTENT_TAG, ContentTypes
|
|
from semantic_kernel.contents.kernel_content import KernelContent
|
|
from semantic_kernel.exceptions.content_exceptions import ContentInitializationError
|
|
|
|
_T = TypeVar("_T", bound="ReasoningContent")
|
|
|
|
|
|
class ReasoningContent(KernelContent):
|
|
"""Represents reasoning content.
|
|
|
|
Exposes a human-readable reasoning ``text``. Any provider-specific fields (for example: ids, encrypted blobs,
|
|
statuses, token info) must be carried in ``metadata`` on the base ``KernelContent``.
|
|
|
|
Attributes:
|
|
content_type: Literal identifying this instance as reasoning content.
|
|
tag: XML tag name used when serializing to/from XML.
|
|
text: The reasoning text to surface to callers.
|
|
|
|
Methods:
|
|
__str__: Return the reasoning text.
|
|
to_element: Serialize to an XML Element using ``tag`` and ``text``.
|
|
from_element: Deserialize from an XML Element into a ReasoningContent.
|
|
to_dict: Serialize to a dict suitable for message payloads.
|
|
"""
|
|
|
|
content_type: Literal[ContentTypes.REASONING_CONTENT] = Field(ContentTypes.REASONING_CONTENT, init=False)
|
|
tag: ClassVar[str] = REASONING_CONTENT_TAG
|
|
text: str | None = None
|
|
|
|
def __str__(self) -> str:
|
|
"""Return the text of the reasoning content."""
|
|
return self.text or ""
|
|
|
|
def to_element(self) -> Element:
|
|
"""Convert the instance to an XML Element."""
|
|
element = Element(self.tag)
|
|
element.text = self.text
|
|
return element
|
|
|
|
@classmethod
|
|
def from_element(cls: type[_T], element: Element) -> _T:
|
|
"""Create an instance from an XML Element."""
|
|
if element.tag != cls.tag:
|
|
raise ContentInitializationError(f"Element tag is not {cls.tag}") # pragma: no cover
|
|
return cls(text=unescape(element.text) if element.text else "")
|
|
|
|
def to_dict(self) -> dict[str, str]:
|
|
"""Convert the instance to a dictionary suitable for message serialization."""
|
|
return {"type": "reasoning", "text": self.text or ""}
|