agentbyte 0.4.7__tar.gz → 0.4.8__tar.gz

{agentbyte-0.4.7 → agentbyte-0.4.8}/CHANGELOG.md

@@ -4,6 +4,21 @@ All notable changes to Agentbyte are documented in this file.
 
 The format follows Keep a Changelog principles and semantic versioning.
 
+## [0.4.8] - 2026-04-06
+
+### Added
+- Added `RetryPolicy` as a public Pydantic v2 model in `agentbyte.llm`. Exported from the package root so callers can define a policy once and share it across multiple clients.
+- Added `RetryMixin` as a shared retry behaviour mixin used by all four provider clients. Provides a single implementation of `_retry_with_backoff` and `_retry_stream_with_backoff` replacing the six previously duplicated methods.
+- All four client constructors (`OpenAIChatCompletionClient`, `AzureOpenAIChatCompletionClient`, `OpenAIEmbeddingClient`, `AzureOpenAIEmbeddingClient`) now accept `retry_policy=` directly. All Azure auth factories (`from_api_key`, `from_default_credential`, `from_certificate`, `from_ad_token`) accept `retry_policy=` on both chat and embedding clients.
+- `RetryPolicy` is now the serialized form for all four provider config classes. `dump_component()` produces a nested `retry_policy` dict; `ComponentLoader.load_component()` restores it.
+
+### Changed
+- Provider config classes (`OpenAIChatCompletionClientConfig`, `AzureOpenAIChatCompletionClientConfig`, `OpenAIEmbeddingClientConfig`, `AzureOpenAIEmbeddingClientConfig`) replace the three flat scalar fields (`max_retries`, `initial_retry_delay`, `max_retry_delay`) with a nested `retry_policy: RetryPolicy` field.
+- `RetryPolicy` validation now rejects negative `max_retries`, negative delays, and configurations where `max_retry_delay < initial_retry_delay`.
+
+### Breaking
+- Component config serialization shape changed: `retry_policy` is now a nested object rather than three flat keys. Any stored component configs must be migrated to the new shape.
+
 ## [0.4.7] - 2026-04-06
 
 ### Added

{agentbyte-0.4.7 → agentbyte-0.4.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agentbyte
-Version: 0.4.7
+Version: 0.4.8
 Summary: A toolkit for designing multiagent systems
 Project-URL: Homepage, https://gitlab.com/pyninja/aiengineering/agentbyte
 Project-URL: Repository, https://gitlab.com/pyninja/aiengineering/agentbyte
@@ -36,17 +36,25 @@ Requires-Dist: uvicorn>=0.42.0; extra == 'webui'
 Description-Content-Type: text/markdown
 
 <p align="center">
-  <img src="https://gitlab.com/pyninja/aiengineering/agentbyte/-/raw/v0.4.7/logo/agent-byte-avatar-low.png" alt="Agentbyte" width="200"/>
+  <img src="https://gitlab.com/pyninja/aiengineering/agentbyte/-/raw/v0.4.8/logo/agent-byte-avatar-low.png" alt="Agentbyte" width="200"/>
 </p>
 
 # Agentbyte
 
 Agentbyte is an observability-first agentic AI framework for building and studying multiagent systems with a learning-first, implementation-oriented workflow.
 
-Current release: **0.4.7**
+Current release: **0.4.8**
 
 Repository: [gitlab.com/pyninja/aiengineering/agentbyte](https://gitlab.com/pyninja/aiengineering/agentbyte)
 
+## What's New in 0.4.8
+
+- Added `RetryPolicy` as a public Pydantic v2 model exported from `agentbyte.llm`. Define a policy once and share it across multiple clients.
+- Unified retry logic: `RetryMixin` replaces six duplicated retry method implementations across the four provider clients with a single shared implementation.
+- All four client constructors and all Azure auth factories now accept `retry_policy=` directly.
+- Provider config serialization now uses a nested `retry_policy` object instead of three flat scalar fields.
+- `RetryPolicy` validation rejects negative `max_retries`, negative delays, and inverted delay bounds.
+
 ## What's New in 0.4.7
 
 - Reorganised `agentbyte.llm` into provider packages: `agentbyte.llm.openai` and `agentbyte.llm.azure`. All existing flat `agentbyte.llm.*` imports continue to work via compatibility shims.

{agentbyte-0.4.7 → agentbyte-0.4.8}/README.md

@@ -1,15 +1,23 @@
 <p align="center">
-  <img src="https://gitlab.com/pyninja/aiengineering/agentbyte/-/raw/v0.4.7/logo/agent-byte-avatar-low.png" alt="Agentbyte" width="200"/>
+  <img src="https://gitlab.com/pyninja/aiengineering/agentbyte/-/raw/v0.4.8/logo/agent-byte-avatar-low.png" alt="Agentbyte" width="200"/>
 </p>
 
 # Agentbyte
 
 Agentbyte is an observability-first agentic AI framework for building and studying multiagent systems with a learning-first, implementation-oriented workflow.
 
-Current release: **0.4.7**
+Current release: **0.4.8**
 
 Repository: [gitlab.com/pyninja/aiengineering/agentbyte](https://gitlab.com/pyninja/aiengineering/agentbyte)
 
+## What's New in 0.4.8
+
+- Added `RetryPolicy` as a public Pydantic v2 model exported from `agentbyte.llm`. Define a policy once and share it across multiple clients.
+- Unified retry logic: `RetryMixin` replaces six duplicated retry method implementations across the four provider clients with a single shared implementation.
+- All four client constructors and all Azure auth factories now accept `retry_policy=` directly.
+- Provider config serialization now uses a nested `retry_policy` object instead of three flat scalar fields.
+- `RetryPolicy` validation rejects negative `max_retries`, negative delays, and inverted delay bounds.
+
 ## What's New in 0.4.7
 
 - Reorganised `agentbyte.llm` into provider packages: `agentbyte.llm.openai` and `agentbyte.llm.azure`. All existing flat `agentbyte.llm.*` imports continue to work via compatibility shims.

{agentbyte-0.4.7 → agentbyte-0.4.8}/src/agentbyte/__about__.py

@@ -1,2 +1,2 @@
-__version__ = "0.4.7"
+__version__ = "0.4.8"
 VERSION = __version__

{agentbyte-0.4.7 → agentbyte-0.4.8}/src/agentbyte/llm/__init__.py

@@ -31,6 +31,7 @@ from .base import (
     AuthenticationError,
     InvalidRequestError,
 )
+from .retry_policy import RetryPolicy
 from .embeddings_base import BaseEmbeddingClient, BaseEmbeddingClientConfig
 from .openai import OpenAIChatCompletionClient, OpenAIChatCompletionClientConfig
 from .openai import OpenAIEmbeddingClient, OpenAIEmbeddingClientConfig
@@ -93,6 +94,8 @@ __all__ = [
     "OpenAISettings",
     "AzureOpenAISettings",
     "AzureServicePrincipalSettings",
+    # Retry policy
+    "RetryPolicy",
     # Exception types
     "ModelClientError",
     "RateLimitError",

{agentbyte-0.4.7 → agentbyte-0.4.8}/src/agentbyte/llm/azure/chat.py

@@ -2,12 +2,11 @@
 Azure OpenAI chat completion client implementation.
 """
 
-import asyncio
 import json
 import logging
 import time
 from collections.abc import AsyncGenerator
-from typing import Any, Callable, Dict, List, Optional, TypeVar
+from typing import Any, Dict, List, Optional, TypeVar
 
 from openai import AsyncAzureOpenAI
 
@@ -15,7 +14,7 @@ from agentbyte.messages import AssistantMessage, Message, ToolCallRequest
 
 from agentbyte.component import Component
 
-from .._retry_observability import build_retry_record, build_retry_summary
+from .._retry_observability import build_retry_summary
 from ..base import (
     BaseChatCompletionClient,
     BaseChatCompletionClientConfig,
@@ -24,6 +23,7 @@ from ..base import (
     InvalidRequestError,
 )
 from ..pricing import PricingRegistry
+from ..retry_policy import RetryMixin, RetryPolicy
 from ..types import (
     ChatCompletionChunk,
     ChatCompletionResult,
@@ -40,9 +40,7 @@ T = TypeVar("T")
 class AzureOpenAIChatCompletionClientConfig(BaseChatCompletionClientConfig):
     """Configuration for AzureOpenAIChatCompletionClient serialization."""
 
-    max_retries: int = 3
-    initial_retry_delay: float = 1.0
-    max_retry_delay: float = 60.0
+    retry_policy: RetryPolicy = RetryPolicy()
     key_env_var: str = "AZURE_OPENAI_API_KEY"
     deployment_env_var: str = "AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"
     api_version_env_var: str = "AZURE_OPENAI_API_VERSION"
@@ -51,6 +49,7 @@ class AzureOpenAIChatCompletionClientConfig(BaseChatCompletionClientConfig):
 class AzureOpenAIChatCompletionClient(
     BaseChatCompletionClient,
     Component[AzureOpenAIChatCompletionClientConfig],
+    RetryMixin,
 ):
     """Azure OpenAI chat completion client."""
 
@@ -67,12 +66,30 @@ class AzureOpenAIChatCompletionClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ):
         super().__init__(model, client, config)
         self.model_pricing = model_pricing
-        self.max_retries = max_retries
-        self.initial_retry_delay = initial_retry_delay
-        self.max_retry_delay = max_retry_delay
+        if retry_policy is not None:
+            self.retry_policy = retry_policy
+        else:
+            self.retry_policy = RetryPolicy(
+                max_retries=max_retries,
+                initial_retry_delay=initial_retry_delay,
+                max_retry_delay=max_retry_delay,
+            )
+
+    @property
+    def max_retries(self) -> int:
+        return self.retry_policy.max_retries
+
+    @property
+    def initial_retry_delay(self) -> float:
+        return self.retry_policy.initial_retry_delay
+
+    @property
+    def max_retry_delay(self) -> float:
+        return self.retry_policy.max_retry_delay
 
     @classmethod
     def from_api_key(
@@ -86,6 +103,7 @@ class AzureOpenAIChatCompletionClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIChatCompletionClient":
         """Create a client using an Azure OpenAI API key."""
         from pydantic import ValidationError
@@ -134,6 +152,7 @@ class AzureOpenAIChatCompletionClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
     @classmethod
@@ -148,6 +167,7 @@ class AzureOpenAIChatCompletionClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIChatCompletionClient":
         """Create a client using DefaultAzureCredential."""
         try:
@@ -198,6 +218,7 @@ class AzureOpenAIChatCompletionClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
     @classmethod
@@ -213,6 +234,7 @@ class AzureOpenAIChatCompletionClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIChatCompletionClient":
         """Create a client using certificate-based Azure AD service principal auth."""
         from pydantic import ValidationError
@@ -271,6 +293,7 @@ class AzureOpenAIChatCompletionClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
     @classmethod
@@ -285,6 +308,7 @@ class AzureOpenAIChatCompletionClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIChatCompletionClient":
         """Create a client using a pre-fetched Azure AD bearer token."""
         from pydantic import ValidationError
@@ -326,83 +350,9 @@ class AzureOpenAIChatCompletionClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
-    async def _retry_with_backoff(
-        self,
-        func: Callable[..., Any],
-        *args: Any,
-        _retry_history: Optional[List] = None,
-        **kwargs: Any,
-    ) -> Any:
-        """Execute a function with exponential backoff retry logic."""
-        last_exception = None
-        current_delay = self.initial_retry_delay
-
-        for attempt in range(self.max_retries + 1):
-            try:
-                return await func(*args, **kwargs)
-            except (RateLimitError, ModelClientError) as e:
-                last_exception = e
-
-                if attempt >= self.max_retries:
-                    logger.warning(
-                        f"Max retries ({self.max_retries}) reached for {func.__name__}"
-                    )
-                    raise
-
-                if _retry_history is not None:
-                    _retry_history.append(
-                        build_retry_record(attempt + 1, current_delay, type(e), "retrying")
-                    )
-
-                logger.debug(
-                    f"Transient error in {func.__name__} "
-                    f"(attempt {attempt + 1}/{self.max_retries + 1}): {e}. "
-                    f"Retrying in {current_delay:.1f}s"
-                )
-                await asyncio.sleep(current_delay)
-                current_delay = min(current_delay * 2, self.max_retry_delay)
-
-            except (AuthenticationError, InvalidRequestError) as e:
-                logger.error(f"Non-transient error in {func.__name__}: {e}")
-                raise
-
-            except Exception as e:
-                logger.error(f"Unexpected error in {func.__name__}: {e}")
-                raise
-
-        raise last_exception or ModelClientError("Retry logic error")
-
-    async def _retry_stream_with_backoff(
-        self,
-        func: Callable[..., AsyncGenerator[ChatCompletionChunk, None]],
-        *args: Any,
-        **kwargs: Any,
-    ) -> AsyncGenerator[ChatCompletionChunk, None]:
-        """Execute a streaming function with exponential backoff retry logic."""
-        current_delay = self.initial_retry_delay
-
-        for attempt in range(self.max_retries + 1):
-            saw_chunk = False
-            try:
-                async for chunk in func(*args, **kwargs):
-                    saw_chunk = True
-                    yield chunk
-                return
-            except (RateLimitError, ModelClientError) as e:
-                if saw_chunk or attempt >= self.max_retries:
-                    raise
-                logger.debug(
-                    f"Transient stream error in {func.__name__} "
-                    f"(attempt {attempt + 1}/{self.max_retries + 1}): {e}. "
-                    f"Retrying in {current_delay:.1f}s"
-                )
-                await asyncio.sleep(current_delay)
-                current_delay = min(current_delay * 2, self.max_retry_delay)
-            except (AuthenticationError, InvalidRequestError):
-                raise
-
     async def create(
         self,
         messages: List[Message],
@@ -682,9 +632,7 @@ class AzureOpenAIChatCompletionClient(
         return AzureOpenAIChatCompletionClientConfig(
             model=self.model,
             config=self.config.copy(),
-            max_retries=self.max_retries,
-            initial_retry_delay=self.initial_retry_delay,
-            max_retry_delay=self.max_retry_delay,
+            retry_policy=self.retry_policy,
             key_env_var=getattr(self, "_key_env_var", "AZURE_OPENAI_API_KEY"),
             deployment_env_var=getattr(self, "_deployment_env_var", "AZURE_OPENAI_CHAT_DEPLOYMENT_NAME"),
             api_version_env_var=getattr(self, "_api_version_env_var", "AZURE_OPENAI_API_VERSION"),
@@ -707,9 +655,7 @@ class AzureOpenAIChatCompletionClient(
             model=config.model if config.model != "gpt-4.1-mini" else None,
             api_key=resolved_key,
             config=config.config or None,
-            max_retries=config.max_retries,
-            initial_retry_delay=config.initial_retry_delay,
-            max_retry_delay=config.max_retry_delay,
+            retry_policy=config.retry_policy,
         )
         instance._key_env_var = config.key_env_var
         instance._deployment_env_var = config.deployment_env_var

{agentbyte-0.4.7 → agentbyte-0.4.8}/src/agentbyte/llm/azure/embedding.py

@@ -2,7 +2,6 @@
 Azure OpenAI embeddings client implementation.
 """
 
-import asyncio
 import logging
 import time
 from typing import Any, Dict, List, Optional
@@ -12,10 +11,11 @@ from openai import AsyncAzureOpenAI
 from agentbyte.component import Component
 from agentbyte.middleware.base import BaseMiddleware
 
-from .._retry_observability import build_retry_record, build_retry_summary
+from .._retry_observability import build_retry_summary
 from ..base import AuthenticationError, InvalidRequestError, RateLimitError
 from ..embeddings_base import BaseEmbeddingClient, BaseEmbeddingClientConfig
 from ..pricing import PricingRegistry
+from ..retry_policy import RetryMixin, RetryPolicy
 from ..types import EmbeddingResult, ModelClientError, ModelPricing, Usage
 
 logger = logging.getLogger(__name__)
@@ -24,9 +24,7 @@ logger = logging.getLogger(__name__)
 class AzureOpenAIEmbeddingClientConfig(BaseEmbeddingClientConfig):
     """Configuration for AzureOpenAIEmbeddingClient serialization."""
 
-    max_retries: int = 3
-    initial_retry_delay: float = 1.0
-    max_retry_delay: float = 60.0
+    retry_policy: RetryPolicy = RetryPolicy()
     key_env_var: str = "AZURE_OPENAI_API_KEY"
     deployment_env_var: str = "AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME"
     api_version_env_var: str = "AZURE_OPENAI_API_VERSION"
@@ -35,6 +33,7 @@ class AzureOpenAIEmbeddingClientConfig(BaseEmbeddingClientConfig):
 class AzureOpenAIEmbeddingClient(
     BaseEmbeddingClient,
     Component[AzureOpenAIEmbeddingClientConfig],
+    RetryMixin,
 ):
     """Azure OpenAI embeddings client with single and batch methods."""
 
@@ -52,12 +51,30 @@ class AzureOpenAIEmbeddingClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ):
         super().__init__(model, client, config, middlewares=middlewares)
         self.model_pricing = model_pricing
-        self.max_retries = max_retries
-        self.initial_retry_delay = initial_retry_delay
-        self.max_retry_delay = max_retry_delay
+        if retry_policy is not None:
+            self.retry_policy = retry_policy
+        else:
+            self.retry_policy = RetryPolicy(
+                max_retries=max_retries,
+                initial_retry_delay=initial_retry_delay,
+                max_retry_delay=max_retry_delay,
+            )
+
+    @property
+    def max_retries(self) -> int:
+        return self.retry_policy.max_retries
+
+    @property
+    def initial_retry_delay(self) -> float:
+        return self.retry_policy.initial_retry_delay
+
+    @property
+    def max_retry_delay(self) -> float:
+        return self.retry_policy.max_retry_delay
 
     @classmethod
     def from_api_key(
@@ -72,6 +89,7 @@ class AzureOpenAIEmbeddingClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIEmbeddingClient":
         """Create client using Azure OpenAI API key authentication."""
         from pydantic import ValidationError
@@ -121,6 +139,7 @@ class AzureOpenAIEmbeddingClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
     @classmethod
@@ -136,6 +155,7 @@ class AzureOpenAIEmbeddingClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIEmbeddingClient":
         """Create client using DefaultAzureCredential token provider."""
         try:
@@ -187,6 +207,7 @@ class AzureOpenAIEmbeddingClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
     @classmethod
@@ -203,6 +224,7 @@ class AzureOpenAIEmbeddingClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIEmbeddingClient":
         """Create client using certificate-based Azure AD authentication."""
         from pydantic import ValidationError
@@ -262,6 +284,7 @@ class AzureOpenAIEmbeddingClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
     @classmethod
@@ -277,6 +300,7 @@ class AzureOpenAIEmbeddingClient(
         max_retries: int = 3,
         initial_retry_delay: float = 1.0,
         max_retry_delay: float = 60.0,
+        retry_policy: Optional[RetryPolicy] = None,
     ) -> "AzureOpenAIEmbeddingClient":
         """Create client using a pre-fetched Azure AD bearer token."""
         from pydantic import ValidationError
@@ -319,43 +343,9 @@ class AzureOpenAIEmbeddingClient(
             max_retries=max_retries,
             initial_retry_delay=initial_retry_delay,
             max_retry_delay=max_retry_delay,
+            retry_policy=retry_policy,
         )
 
-    async def _retry_with_backoff(
-        self,
-        func: Any,
-        *args: Any,
-        _retry_history: Optional[List] = None,
-        **kwargs: Any,
-    ) -> Any:
-        """Execute an async function with exponential backoff on transient failures."""
-        current_delay = self.initial_retry_delay
-
-        for attempt in range(self.max_retries + 1):
-            try:
-                return await func(*args, **kwargs)
-            except (RateLimitError, ModelClientError) as exc:
-                if attempt >= self.max_retries:
-                    logger.warning(
-                        f"Max retries ({self.max_retries}) reached for {func.__name__}"
-                    )
-                    raise
-                if _retry_history is not None:
-                    _retry_history.append(
-                        build_retry_record(attempt + 1, current_delay, type(exc), "retrying")
-                    )
-                logger.debug(
-                    f"Transient error in {func.__name__} "
-                    f"(attempt {attempt + 1}/{self.max_retries + 1}): {exc}. "
-                    f"Retrying in {current_delay:.1f}s"
-                )
-                await asyncio.sleep(current_delay)
-                current_delay = min(current_delay * 2, self.max_retry_delay)
-            except (AuthenticationError, InvalidRequestError):
-                raise
-
-        raise ModelClientError("Retry logic error")
-
     async def create(self, input_text: str, **kwargs: Any) -> EmbeddingResult:
         """Create one embedding result from a single text input."""
         return await self.create_batch([input_text], **kwargs)
@@ -474,9 +464,7 @@ class AzureOpenAIEmbeddingClient(
         return AzureOpenAIEmbeddingClientConfig(
             model=self.model,
             config=self.config.copy(),
-            max_retries=self.max_retries,
-            initial_retry_delay=self.initial_retry_delay,
-            max_retry_delay=self.max_retry_delay,
+            retry_policy=self.retry_policy,
             key_env_var=getattr(self, "_key_env_var", "AZURE_OPENAI_API_KEY"),
             deployment_env_var=getattr(
                 self, "_deployment_env_var", "AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME"
@@ -502,9 +490,7 @@ class AzureOpenAIEmbeddingClient(
             model=config.model,
             api_key=resolved_key,
             config=config.config or None,
-            max_retries=config.max_retries,
-            initial_retry_delay=config.initial_retry_delay,
-            max_retry_delay=config.max_retry_delay,
+            retry_policy=config.retry_policy,
         )
         instance._key_env_var = config.key_env_var
         instance._deployment_env_var = config.deployment_env_var

{agentbyte-0.4.7 → agentbyte-0.4.8}/src/agentbyte/llm/base.py

@@ -50,10 +50,12 @@ class BaseChatCompletionClientConfig(BaseModel):
 
     Example:
         ```python
+        from agentbyte.llm import RetryPolicy
+
         class OpenAIChatCompletionClientConfig(BaseChatCompletionClientConfig):
             model: str
             config: Dict[str, Any] = {}
-            max_retries: int = 3
+            retry_policy: RetryPolicy = RetryPolicy()
 
         # Serialize: client → config → JSON
         config = client._to_config()
@@ -428,7 +430,7 @@ class BaseChatCompletionClient(ABC, ComponentLoader):
                 config={"temperature": 0.7}
             )
             config = client._to_config()
-            # config now has: model, config, max_retries, etc.
+            # config now has: model, config, retry_policy (nested RetryPolicy), etc.
             ```
         """
         pass