fix(types): accept SDK NotGiven in LLM Settings fields used for passthrough

Three LLM services initialize certain Settings fields with the SDK's
NOT_GIVEN (openai.NOT_GIVEN or anthropic.NOT_GIVEN) so the value
flows unmodified into SDK API calls. The inherited field types from
LLMSettings only admit pipecat's _NotGiven, so pyright flagged each
constructor call as a flavor mismatch.

Widen the field types in each service-specific Settings subclass so
they accept both pipecat's _NotGiven (for delta-mode defaults) and
the corresponding SDK NotGiven (for store-mode passthrough):

- OpenAILLMSettings: frequency_penalty, presence_penalty, seed,
  temperature, top_p, max_tokens, max_completion_tokens.
- OpenAIResponsesLLMSettings: temperature, top_p,
  max_completion_tokens.
- AnthropicLLMSettings: temperature, top_k, top_p, thinking.

Every overridden field is genuinely read from self._settings and
passed directly to the SDK, so none of the overrides are vestigial.

Clears 21 pyright errors and restores test_service_settings_complete
parity with the pre-NOT_GIVEN-swap state.

src/pipecat/services/anthropic/llm.py

@@ -44,6 +44,7 @@ from pipecat.utils.tracing.service_decorators import traced_llm
 
 try:
     from anthropic import NOT_GIVEN, APITimeoutError, AsyncAnthropic
+    from anthropic import NotGiven as AnthropicNotGiven
 except ModuleNotFoundError as e:
     logger.error(f"Exception: {e}")
     logger.error("In order to use Anthropic, you need to `pip install pipecat-ai[anthropic]`.")
@@ -79,7 +80,15 @@ class AnthropicLLMSettings(LLMSettings):
     """
 
     enable_prompt_caching: bool | _NotGiven = field(default_factory=lambda: _NOT_GIVEN)
-    thinking: Union["AnthropicLLMService.ThinkingConfig", _NotGiven] = field(
+    # Override inherited LLMSettings fields to also accept anthropic's NotGiven
+    # sentinel. The service stores anthropic's NOT_GIVEN in these fields so
+    # they can be passed through unchanged to the AsyncAnthropic client.
+    temperature: float | None | _NotGiven | AnthropicNotGiven = field(
+        default_factory=lambda: _NOT_GIVEN
+    )
+    top_k: int | None | _NotGiven | AnthropicNotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    top_p: float | None | _NotGiven | AnthropicNotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    thinking: Union["AnthropicLLMService.ThinkingConfig", _NotGiven, AnthropicNotGiven] = field(
         default_factory=lambda: _NOT_GIVEN
     )
 

src/pipecat/services/openai/base_llm.py

@@ -22,6 +22,7 @@ from openai import (
     AsyncStream,
     DefaultAsyncHttpxClient,
 )
+from openai._types import NotGiven as OpenAINotGiven
 from openai.types.chat import ChatCompletionChunk
 from pydantic import BaseModel, Field
 
@@ -50,7 +51,24 @@ class OpenAILLMSettings(LLMSettings):
         max_completion_tokens: Maximum completion tokens to generate.
     """
 
-    max_completion_tokens: int | _NotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    # Override inherited LLMSettings fields to also accept openai's NotGiven
+    # sentinel. The service stores openai's NOT_GIVEN in these fields so they
+    # can be passed through unchanged to the AsyncOpenAI client.
+    frequency_penalty: float | None | _NotGiven | OpenAINotGiven = field(
+        default_factory=lambda: _NOT_GIVEN
+    )
+    presence_penalty: float | None | _NotGiven | OpenAINotGiven = field(
+        default_factory=lambda: _NOT_GIVEN
+    )
+    seed: int | None | _NotGiven | OpenAINotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    temperature: float | None | _NotGiven | OpenAINotGiven = field(
+        default_factory=lambda: _NOT_GIVEN
+    )
+    top_p: float | None | _NotGiven | OpenAINotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    max_tokens: int | None | _NotGiven | OpenAINotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    max_completion_tokens: int | _NotGiven | OpenAINotGiven = field(
+        default_factory=lambda: _NOT_GIVEN
+    )
 
 
 class BaseOpenAILLMService(LLMService):

src/pipecat/services/openai/responses/llm.py

@@ -18,6 +18,7 @@ from typing import Any
 import httpx
 from loguru import logger
 from openai import NOT_GIVEN, AsyncOpenAI, AsyncStream, DefaultAsyncHttpxClient
+from openai._types import NotGiven as OpenAINotGiven
 from openai.types.responses import (
     ResponseCompletedEvent,
     ResponseFunctionCallArgumentsDeltaEvent,
@@ -97,7 +98,16 @@ class OpenAIResponsesLLMSettings(LLMSettings):
         max_completion_tokens: Maximum completion tokens to generate.
     """
 
-    max_completion_tokens: int | _NotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    # Override inherited LLMSettings fields to also accept openai's NotGiven
+    # sentinel. The service stores openai's NOT_GIVEN in these fields so they
+    # can be passed through unchanged to the AsyncOpenAI client.
+    temperature: float | None | _NotGiven | OpenAINotGiven = field(
+        default_factory=lambda: _NOT_GIVEN
+    )
+    top_p: float | None | _NotGiven | OpenAINotGiven = field(default_factory=lambda: _NOT_GIVEN)
+    max_completion_tokens: int | _NotGiven | OpenAINotGiven = field(
+        default_factory=lambda: _NOT_GIVEN
+    )
 
 
 # ---------------------------------------------------------------------------