gllm-inference-binary 0.5.2__cp313-cp313-macosx_13_0_x86_64.whl → 0.5.4__cp313-cp313-macosx_13_0_x86_64.whl

gllm_inference/constants.pyi

@@ -3,6 +3,6 @@ from _typeshed import Incomplete
 DEFAULT_AZURE_OPENAI_API_VERSION: str
 DOCUMENT_MIME_TYPES: Incomplete
 GOOGLE_SCOPES: Incomplete
-INVOKER_DEFAULT_RETRY_CONFIG: Incomplete
+INVOKER_DEFAULT_TIMEOUT: float
 HEX_REPR_LENGTH: int
 HTTP_STATUS_CODE_PATTERNS: Incomplete

gllm_inference/em_invoker/azure_openai_em_invoker.pyi

@@ -1,6 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import DEFAULT_AZURE_OPENAI_API_VERSION as DEFAULT_AZURE_OPENAI_API_VERSION, INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
+from gllm_inference.constants import DEFAULT_AZURE_OPENAI_API_VERSION as DEFAULT_AZURE_OPENAI_API_VERSION
 from gllm_inference.em_invoker.openai_em_invoker import OpenAIEMInvoker as OpenAIEMInvoker
 from gllm_inference.schema import ModelId as ModelId, ModelProvider as ModelProvider
 from typing import Any
@@ -67,7 +67,7 @@ class AzureOpenAIEMInvoker(OpenAIEMInvoker):
         ```
     '''
     client: Incomplete
-    def __init__(self, azure_endpoint: str, azure_deployment: str, api_key: str | None = None, api_version: str = ..., model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, azure_endpoint: str, azure_deployment: str, api_key: str | None = None, api_version: str = ..., model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig | None = None) -> None:
         """Initializes a new instance of the AzureOpenAIEMInvoker class.
 
         Args:
@@ -80,6 +80,6 @@ class AzureOpenAIEMInvoker(OpenAIEMInvoker):
             model_kwargs (dict[str, Any] | None, optional): Additional model parameters. Defaults to None.
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         """

gllm_inference/em_invoker/em_invoker.pyi

@@ -1,8 +1,8 @@
 import abc
 from _typeshed import Incomplete
 from abc import ABC
-from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import DOCUMENT_MIME_TYPES as DOCUMENT_MIME_TYPES, INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
+from gllm_core.utils.retry import RetryConfig
+from gllm_inference.constants import DOCUMENT_MIME_TYPES as DOCUMENT_MIME_TYPES, INVOKER_DEFAULT_TIMEOUT as INVOKER_DEFAULT_TIMEOUT
 from gllm_inference.exceptions import parse_error_message as parse_error_message
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, EMContent as EMContent, ModelId as ModelId, Vector as Vector
 from typing import Any
@@ -21,7 +21,7 @@ class BaseEMInvoker(ABC, metaclass=abc.ABCMeta):
     """
     default_hyperparameters: Incomplete
     retry_config: Incomplete
-    def __init__(self, model_id: ModelId, default_hyperparameters: dict[str, Any] | None = None, supported_attachments: set[str] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_id: ModelId, default_hyperparameters: dict[str, Any] | None = None, supported_attachments: set[str] | None = None, retry_config: RetryConfig | None = None) -> None:
         """Initializes a new instance of the BaseEMInvoker class.
 
         Args:
@@ -30,8 +30,8 @@ class BaseEMInvoker(ABC, metaclass=abc.ABCMeta):
                 embedding model. Defaults to None, in which case an empty dictionary is used.
             supported_attachments (set[str] | None, optional): A set of supported attachment types. Defaults to None,
                 in which case an empty set is used (indicating that no attachments are supported).
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         """
     @property
     def model_id(self) -> str:

gllm_inference/em_invoker/google_em_invoker.pyi

@@ -1,6 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import GOOGLE_SCOPES as GOOGLE_SCOPES, INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
+from gllm_inference.constants import GOOGLE_SCOPES as GOOGLE_SCOPES
 from gllm_inference.em_invoker.em_invoker import BaseEMInvoker as BaseEMInvoker
 from gllm_inference.schema import ModelId as ModelId, ModelProvider as ModelProvider, Vector as Vector
 from typing import Any
@@ -97,7 +97,7 @@ class GoogleEMInvoker(BaseEMInvoker):
         ```
     '''
     client_params: Incomplete
-    def __init__(self, model_name: str, api_key: str | None = None, credentials_path: str | None = None, project_id: str | None = None, location: str = 'us-central1', model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_name: str, api_key: str | None = None, credentials_path: str | None = None, project_id: str | None = None, location: str = 'us-central1', model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig | None = None) -> None:
         '''Initializes a new instance of the GoogleEMInvoker class.
 
         Args:
@@ -114,8 +114,8 @@ class GoogleEMInvoker(BaseEMInvoker):
                 Defaults to None.
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
 
         Note:
             If neither `api_key` nor `credentials_path` is provided, Google Gen AI will be used by default.

gllm_inference/em_invoker/langchain_em_invoker.pyi

@@ -1,6 +1,5 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.em_invoker.em_invoker import BaseEMInvoker as BaseEMInvoker
 from gllm_inference.schema import ModelId as ModelId, ModelProvider as ModelProvider, Vector as Vector
 from gllm_inference.utils import load_langchain_model as load_langchain_model, parse_model_data as parse_model_data
@@ -20,7 +19,7 @@ class LangChainEMInvoker(BaseEMInvoker):
         retry_config (RetryConfig): The retry configuration for the embedding model.
     """
     model: Incomplete
-    def __init__(self, model: Embeddings | None = None, model_class_path: str | None = None, model_name: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model: Embeddings | None = None, model_class_path: str | None = None, model_name: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig | None = None) -> None:
         '''Initializes a new instance of the LangChainEMInvoker class.
 
         Args:
@@ -35,6 +34,6 @@ class LangChainEMInvoker(BaseEMInvoker):
                 `model_class_path` is provided. Defaults to None.
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         '''

gllm_inference/em_invoker/openai_compatible_em_invoker.pyi

@@ -1,6 +1,5 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.em_invoker.openai_em_invoker import OpenAIEMInvoker as OpenAIEMInvoker
 from gllm_inference.schema import ModelId as ModelId, ModelProvider as ModelProvider
 from typing import Any
@@ -75,7 +74,7 @@ class OpenAICompatibleEMInvoker(OpenAIEMInvoker):
         ```
     '''
     client: Incomplete
-    def __init__(self, model_name: str, base_url: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_name: str, base_url: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig | None = None) -> None:
         """Initializes a new instance of the OpenAICompatibleEMInvoker class.
 
         Args:
@@ -86,6 +85,6 @@ class OpenAICompatibleEMInvoker(OpenAIEMInvoker):
             model_kwargs (dict[str, Any] | None, optional): Additional model parameters. Defaults to None.
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         """

gllm_inference/em_invoker/openai_em_invoker.pyi

@@ -1,6 +1,5 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.em_invoker.em_invoker import BaseEMInvoker as BaseEMInvoker
 from gllm_inference.schema import ModelId as ModelId, ModelProvider as ModelProvider, Vector as Vector
 from typing import Any
@@ -69,7 +68,7 @@ class OpenAIEMInvoker(BaseEMInvoker):
         ```
     '''
     client: Incomplete
-    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig | None = None) -> None:
         """Initializes a new instance of the OpenAIEMInvoker class.
 
         Args:
@@ -80,6 +79,6 @@ class OpenAIEMInvoker(BaseEMInvoker):
                 Defaults to None.
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         """

gllm_inference/em_invoker/twelevelabs_em_invoker.pyi

@@ -1,6 +1,5 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.em_invoker.em_invoker import BaseEMInvoker as BaseEMInvoker
 from gllm_inference.em_invoker.schema.twelvelabs import InputType as InputType, Key as Key, OutputType as OutputType
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, EMContent as EMContent, ModelId as ModelId, ModelProvider as ModelProvider, Vector as Vector
@@ -83,7 +82,7 @@ class TwelveLabsEMInvoker(BaseEMInvoker):
         ```
     '''
     client: Incomplete
-    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig | None = None) -> None:
         """Initializes a new instance of the TwelveLabsEMInvoker class.
 
         Args:
@@ -94,6 +93,6 @@ class TwelveLabsEMInvoker(BaseEMInvoker):
                 Defaults to None.
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         """

gllm_inference/em_invoker/voyage_em_invoker.pyi

@@ -1,6 +1,5 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.em_invoker.em_invoker import BaseEMInvoker as BaseEMInvoker
 from gllm_inference.em_invoker.schema.voyage import InputType as InputType, Key as Key
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, EMContent as EMContent, ModelId as ModelId, ModelProvider as ModelProvider, Vector as Vector
@@ -85,7 +84,7 @@ class VoyageEMInvoker(BaseEMInvoker):
         ```
     '''
     client: Incomplete
-    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, retry_config: RetryConfig | None = None) -> None:
         """Initializes a new instance of the VoyageEMInvoker class.
 
         Args:
@@ -96,6 +95,6 @@ class VoyageEMInvoker(BaseEMInvoker):
                 Defaults to None.
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
-            retry_config (RetryConfig, optional): The retry configuration for the embedding model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the embedding model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         """

gllm_inference/lm_invoker/anthropic_lm_invoker.pyi

@@ -1,7 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
 from gllm_inference.lm_invoker.schema.anthropic import InputType as InputType, Key as Key, OutputType as OutputType
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, EmitDataType as EmitDataType, LMOutput as LMOutput, Message as Message, ModelId as ModelId, ModelProvider as ModelProvider, Reasoning as Reasoning, ResponseSchema as ResponseSchema, TokenUsage as TokenUsage, ToolCall as ToolCall, ToolResult as ToolResult
@@ -222,7 +221,7 @@ class AnthropicLMInvoker(BaseLMInvoker):
     client: Incomplete
     thinking: Incomplete
     thinking_budget: Incomplete
-    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ..., thinking: bool = False, thinking_budget: int = ...) -> None:
+    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None, thinking: bool = False, thinking_budget: int = ...) -> None:
         """Initializes the AnthropicLmInvoker instance.
 
         Args:
@@ -237,8 +236,8 @@ class AnthropicLMInvoker(BaseLMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
             thinking (bool, optional): Whether to enable thinking. Only allowed for thinking models. Defaults to False.
             thinking_budget (int, optional): The tokens allocated for the thinking process. Must be greater than or
                 equal to 1024. Only allowed for thinking models. Defaults to DEFAULT_THINKING_BUDGET.

gllm_inference/lm_invoker/azure_openai_lm_invoker.pyi

@@ -1,6 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import DEFAULT_AZURE_OPENAI_API_VERSION as DEFAULT_AZURE_OPENAI_API_VERSION, INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
+from gllm_inference.constants import DEFAULT_AZURE_OPENAI_API_VERSION as DEFAULT_AZURE_OPENAI_API_VERSION
 from gllm_inference.lm_invoker.openai_lm_invoker import OpenAILMInvoker as OpenAILMInvoker, ReasoningEffort as ReasoningEffort, ReasoningSummary as ReasoningSummary
 from gllm_inference.schema import ModelId as ModelId, ModelProvider as ModelProvider, ResponseSchema as ResponseSchema
 from langchain_core.tools import Tool as Tool
@@ -218,7 +218,7 @@ class AzureOpenAILMInvoker(OpenAILMInvoker):
                 Defaults to an empty list.
     '''
     client: Incomplete
-    def __init__(self, azure_endpoint: str, azure_deployment: str, api_key: str | None = None, api_version: str = ..., model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ..., reasoning_effort: ReasoningEffort | None = None, reasoning_summary: ReasoningSummary | None = None) -> None:
+    def __init__(self, azure_endpoint: str, azure_deployment: str, api_key: str | None = None, api_version: str = ..., model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None, reasoning_effort: ReasoningEffort | None = None, reasoning_summary: ReasoningSummary | None = None) -> None:
         """Initializes a new instance of the AzureOpenAILMInvoker class.
 
         Args:
@@ -237,8 +237,8 @@ class AzureOpenAILMInvoker(OpenAILMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
             reasoning_effort (ReasoningEffort | None, optional): The reasoning effort for reasoning models. Not allowed
                 for non-reasoning models. If None, the model will perform medium reasoning effort. Defaults to None.
             reasoning_summary (ReasoningSummary | None, optional): The reasoning summary level for reasoning models.

gllm_inference/lm_invoker/bedrock_lm_invoker.pyi

@@ -1,7 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
 from gllm_inference.lm_invoker.schema.bedrock import InputType as InputType, Key as Key, OutputType as OutputType
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, LMOutput as LMOutput, Message as Message, ModelId as ModelId, ModelProvider as ModelProvider, ResponseSchema as ResponseSchema, TokenUsage as TokenUsage, ToolCall as ToolCall, ToolResult as ToolResult
@@ -180,7 +179,7 @@ class BedrockLMInvoker(BaseLMInvoker):
     '''
     session: Incomplete
     client_kwargs: Incomplete
-    def __init__(self, model_name: str, access_key_id: str | None = None, secret_access_key: str | None = None, region_name: str = 'us-east-1', model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_name: str, access_key_id: str | None = None, secret_access_key: str | None = None, region_name: str = 'us-east-1', model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None) -> None:
         '''Initializes the BedrockLMInvoker instance.
 
         Args:
@@ -198,8 +197,8 @@ class BedrockLMInvoker(BaseLMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
 
         Raises:
             ValueError: If `response_schema` is provided, but `tools` are also provided.

gllm_inference/lm_invoker/datasaur_lm_invoker.pyi

@@ -1,7 +1,7 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import DOCUMENT_MIME_TYPES as DOCUMENT_MIME_TYPES, INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
+from gllm_inference.constants import DOCUMENT_MIME_TYPES as DOCUMENT_MIME_TYPES
 from gllm_inference.lm_invoker.openai_compatible_lm_invoker import OpenAICompatibleLMInvoker as OpenAICompatibleLMInvoker
 from gllm_inference.lm_invoker.schema.datasaur import InputType as InputType, Key as Key
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, LMOutput as LMOutput, Message as Message, ModelId as ModelId, ModelProvider as ModelProvider, ResponseSchema as ResponseSchema, ToolCall as ToolCall, ToolResult as ToolResult
@@ -121,7 +121,7 @@ class DatasaurLMInvoker(OpenAICompatibleLMInvoker):
     '''
     client: Incomplete
     citations: Incomplete
-    def __init__(self, base_url: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, output_analytics: bool = False, retry_config: RetryConfig = ..., citations: bool = False) -> None:
+    def __init__(self, base_url: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None, citations: bool = False) -> None:
         """Initializes a new instance of the DatasaurLMInvoker class.
 
         Args:
@@ -132,8 +132,8 @@ class DatasaurLMInvoker(OpenAICompatibleLMInvoker):
             default_hyperparameters (dict[str, Any] | None, optional): Default hyperparameters for invoking the model.
                 Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
             citations (bool, optional): Whether to output the citations. Defaults to False.
 
         Raises:

gllm_inference/lm_invoker/google_lm_invoker.pyi

@@ -1,7 +1,7 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import GOOGLE_SCOPES as GOOGLE_SCOPES, INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
+from gllm_inference.constants import GOOGLE_SCOPES as GOOGLE_SCOPES
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
 from gllm_inference.lm_invoker.schema.google import InputType as InputType, Key as Key
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, EmitDataType as EmitDataType, LMOutput as LMOutput, Message as Message, MessageRole as MessageRole, ModelId as ModelId, ModelProvider as ModelProvider, Reasoning as Reasoning, ResponseSchema as ResponseSchema, TokenUsage as TokenUsage, ToolCall as ToolCall, ToolResult as ToolResult
@@ -250,7 +250,7 @@ class GoogleLMInvoker(BaseLMInvoker):
     client_params: Incomplete
     thinking: Incomplete
     thinking_budget: Incomplete
-    def __init__(self, model_name: str, api_key: str | None = None, credentials_path: str | None = None, project_id: str | None = None, location: str = 'us-central1', model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ..., thinking: bool | None = None, thinking_budget: int = ...) -> None:
+    def __init__(self, model_name: str, api_key: str | None = None, credentials_path: str | None = None, project_id: str | None = None, location: str = 'us-central1', model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None, thinking: bool | None = None, thinking_budget: int = ...) -> None:
         '''Initializes a new instance of the GoogleLMInvoker class.
 
         Args:
@@ -273,8 +273,8 @@ class GoogleLMInvoker(BaseLMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
             thinking (bool | None, optional): Whether to enable thinking. Only allowed for thinking models.
                 Defaults to True for Gemini 2.5 Pro models and False for other models.
             thinking_budget (int, optional): The tokens allowed for thinking process. Only allowed for thinking models.

gllm_inference/lm_invoker/langchain_lm_invoker.pyi

@@ -1,7 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
 from gllm_inference.lm_invoker.schema.langchain import InputType as InputType, Key as Key
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, LMOutput as LMOutput, Message as Message, MessageRole as MessageRole, ModelId as ModelId, ModelProvider as ModelProvider, ResponseSchema as ResponseSchema, TokenUsage as TokenUsage, ToolCall as ToolCall, ToolResult as ToolResult
@@ -204,7 +203,7 @@ class LangChainLMInvoker(BaseLMInvoker):
                 Defaults to an empty list.
     '''
     model: Incomplete
-    def __init__(self, model: BaseChatModel | None = None, model_class_path: str | None = None, model_name: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model: BaseChatModel | None = None, model_class_path: str | None = None, model_name: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None) -> None:
         '''Initializes a new instance of the LangChainLMInvoker class.
 
         Args:
@@ -224,8 +223,8 @@ class LangChainLMInvoker(BaseLMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
 
         Raises:
             ValueError: If `response_schema` is provided, but `tools` are also provided.

gllm_inference/lm_invoker/litellm_lm_invoker.pyi

@@ -1,7 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.lm_invoker.openai_compatible_lm_invoker import OpenAICompatibleLMInvoker as OpenAICompatibleLMInvoker
 from gllm_inference.lm_invoker.openai_lm_invoker import ReasoningEffort as ReasoningEffort
 from gllm_inference.schema import AttachmentType as AttachmentType, LMOutput as LMOutput, ModelId as ModelId, ModelProvider as ModelProvider, ResponseSchema as ResponseSchema
@@ -228,7 +227,7 @@ class LiteLLMLMInvoker(OpenAICompatibleLMInvoker):
                 Defaults to an empty list.
     '''
     completion: Incomplete
-    def __init__(self, model_id: str, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ..., reasoning_effort: ReasoningEffort | None = None) -> None:
+    def __init__(self, model_id: str, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None, reasoning_effort: ReasoningEffort | None = None) -> None:
         """Initializes a new instance of the LiteLLMLMInvoker class.
 
         Args:
@@ -240,8 +239,8 @@ class LiteLLMLMInvoker(OpenAICompatibleLMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
             reasoning_effort (ReasoningEffort | None, optional): The reasoning effort for reasoning models.
                 Defaults to None.
         """

gllm_inference/lm_invoker/lm_invoker.pyi

@@ -3,7 +3,7 @@ from _typeshed import Incomplete
 from abc import ABC
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig
-from gllm_inference.constants import DOCUMENT_MIME_TYPES as DOCUMENT_MIME_TYPES, INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
+from gllm_inference.constants import DOCUMENT_MIME_TYPES as DOCUMENT_MIME_TYPES, INVOKER_DEFAULT_TIMEOUT as INVOKER_DEFAULT_TIMEOUT
 from gllm_inference.exceptions import parse_error_message as parse_error_message
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, EmitDataType as EmitDataType, LMOutput as LMOutput, Message as Message, MessageContent as MessageContent, MessageRole as MessageRole, ModelId as ModelId, Reasoning as Reasoning, ResponseSchema as ResponseSchema, ToolCall as ToolCall, ToolResult as ToolResult
 from langchain_core.tools import Tool as Tool
@@ -46,7 +46,7 @@ class BaseLMInvoker(ABC, metaclass=abc.ABCMeta):
     response_schema: Incomplete
     output_analytics: Incomplete
     retry_config: Incomplete
-    def __init__(self, model_id: ModelId, default_hyperparameters: dict[str, Any] | None = None, supported_attachments: set[str] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ...) -> None:
+    def __init__(self, model_id: ModelId, default_hyperparameters: dict[str, Any] | None = None, supported_attachments: set[str] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None) -> None:
         """Initializes a new instance of the BaseLMInvoker class.
 
         Args:
@@ -61,8 +61,8 @@ class BaseLMInvoker(ABC, metaclass=abc.ABCMeta):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
         """
     @property
     def model_id(self) -> str:

gllm_inference/lm_invoker/openai_compatible_lm_invoker.pyi

@@ -1,7 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
 from gllm_inference.lm_invoker.schema.openai_compatible import InputType as InputType, Key as Key, ReasoningEffort as ReasoningEffort
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, EmitDataType as EmitDataType, LMOutput as LMOutput, Message as Message, MessageRole as MessageRole, ModelId as ModelId, ModelProvider as ModelProvider, Reasoning as Reasoning, ResponseSchema as ResponseSchema, TokenUsage as TokenUsage, ToolCall as ToolCall, ToolResult as ToolResult
@@ -231,7 +230,7 @@ class OpenAICompatibleLMInvoker(BaseLMInvoker):
                 Defaults to an empty list.
     '''
     client: Incomplete
-    def __init__(self, model_name: str, base_url: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ..., reasoning_effort: ReasoningEffort | None = None) -> None:
+    def __init__(self, model_name: str, base_url: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None, reasoning_effort: ReasoningEffort | None = None) -> None:
         '''Initializes a new instance of the OpenAICompatibleLMInvoker class.
 
         Args:
@@ -249,8 +248,8 @@ class OpenAICompatibleLMInvoker(BaseLMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
             reasoning_effort (str | None, optional): The reasoning effort for the language model. Defaults to None.
         '''
     def set_response_schema(self, response_schema: ResponseSchema | None) -> None:

gllm_inference/lm_invoker/openai_lm_invoker.pyi

@@ -1,7 +1,6 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_core.utils.retry import RetryConfig as RetryConfig
-from gllm_inference.constants import INVOKER_DEFAULT_RETRY_CONFIG as INVOKER_DEFAULT_RETRY_CONFIG
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
 from gllm_inference.lm_invoker.schema.openai import InputType as InputType, Key as Key, OutputType as OutputType, ReasoningEffort as ReasoningEffort, ReasoningSummary as ReasoningSummary
 from gllm_inference.schema import Attachment as Attachment, AttachmentType as AttachmentType, CodeExecResult as CodeExecResult, EmitDataType as EmitDataType, LMOutput as LMOutput, Message as Message, MessageRole as MessageRole, ModelId as ModelId, ModelProvider as ModelProvider, Reasoning as Reasoning, ResponseSchema as ResponseSchema, TokenUsage as TokenUsage, ToolCall as ToolCall, ToolResult as ToolResult
@@ -319,7 +318,7 @@ class OpenAILMInvoker(BaseLMInvoker):
                 enabled and the language model decides to execute any codes. Defaults to an empty list.
     '''
     client: Incomplete
-    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig = ..., reasoning_effort: ReasoningEffort | None = None, reasoning_summary: ReasoningSummary | None = None, code_interpreter: bool = False, web_search: bool = False) -> None:
+    def __init__(self, model_name: str, api_key: str | None = None, model_kwargs: dict[str, Any] | None = None, default_hyperparameters: dict[str, Any] | None = None, tools: list[Tool] | None = None, response_schema: ResponseSchema | None = None, output_analytics: bool = False, retry_config: RetryConfig | None = None, reasoning_effort: ReasoningEffort | None = None, reasoning_summary: ReasoningSummary | None = None, code_interpreter: bool = False, web_search: bool = False) -> None:
         """Initializes a new instance of the OpenAILMInvoker class.
 
         Args:
@@ -335,8 +334,8 @@ class OpenAILMInvoker(BaseLMInvoker):
                 output a structured response as defined by the schema. Supports both Pydantic BaseModel and JSON schema
                 dictionary. Defaults to None.
             output_analytics (bool, optional): Whether to output the invocation analytics. Defaults to False.
-            retry_config (RetryConfig, optional): The retry configuration for the language model.
-                Defaults to a default config with no retry and 30.0 seconds timeout.
+            retry_config (RetryConfig | None, optional): The retry configuration for the language model.
+                Defaults to None, in which case a default config with no retry and 30.0 seconds timeout will be used.
             reasoning_effort (ReasoningEffort | None, optional): The reasoning effort for reasoning models. Not allowed
                 for non-reasoning models. If None, the model will perform medium reasoning effort. Defaults to None.
             reasoning_summary (ReasoningSummary | None, optional): The reasoning summary level for reasoning models.

gllm_inference/schema/attachment.pyi

@@ -32,6 +32,18 @@ class Attachment(BaseModel):
             Attachment: The instantiated Attachment.
         """
     @classmethod
+    def from_base64(cls, base64_data: str, filename: str | None = None) -> Attachment:
+        """Creates an Attachment from a base64 string.
+
+        Args:
+            base64_data (str): The base64 string of the file.
+            filename (str | None, optional): The filename of the file. Defaults to None,
+                in which case the filename will be derived from the mime type.
+
+        Returns:
+            Attachment: The instantiated Attachment.
+        """
+    @classmethod
     def from_data_url(cls, data_url: str, filename: str | None = None) -> Attachment:
         """Creates an Attachment from a data URL (data:[mime/type];base64,[bytes]).
 

{gllm_inference_binary-0.5.2.dist-info → gllm_inference_binary-0.5.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: gllm-inference-binary
-Version: 0.5.2
+Version: 0.5.4
 Summary: A library containing components related to model inferences in Gen AI applications.
 Author: Henry Wicaksono
 Author-email: henry.wicaksono@gdplabs.id

{gllm_inference_binary-0.5.2.dist-info → gllm_inference_binary-0.5.4.dist-info}/RECORD

@@ -8,35 +8,35 @@ gllm_inference/catalog/__init__.pyi,sha256=JBkPGTyiiZ30GECzJBW-mW8LekWyY2qyzal3e
 gllm_inference/catalog/catalog.pyi,sha256=a4RNG1lKv51GxQpOqh47tz-PAROMPaeP2o5XNLBSZaU,4790
 gllm_inference/catalog/lm_request_processor_catalog.pyi,sha256=wjzufPEqey-byBU3hPWwEawT9c182WwjzSWOJ2bnqIs,4599
 gllm_inference/catalog/prompt_builder_catalog.pyi,sha256=o4JSzISjlStBataofZ2MB2_t3wnGTkkFJ3Dm_NSm5qo,3159
-gllm_inference/constants.pyi,sha256=mK_YuuN1MU0Kk4aD-4kDhRZnwPiPFy2l34Mg2ijoir8,230
+gllm_inference/constants.pyi,sha256=A16iMdS6QLnDx7ToiVuu1rSxvEwcr0OMrghPUGQL0L4,220
 gllm_inference/em_invoker/__init__.pyi,sha256=XESsrYo1PZeeHe7AMRyuzKoV7XDD5oN89ZTH01zRf4k,873
-gllm_inference/em_invoker/azure_openai_em_invoker.pyi,sha256=TDE5g8PM0dJUTH9yBQ8uIM2Wd9NKKwIoVkbRwJhV-HE,4487
-gllm_inference/em_invoker/em_invoker.pyi,sha256=ffEwrFpdfV6vlwjuop4iPQm_dgcojLIMHXAhHacRlWM,4318
-gllm_inference/em_invoker/google_em_invoker.pyi,sha256=dPuc68qk1hdjswCCAoPNJ6ZoeW3oX1VYRW2LhWAp8IY,6081
+gllm_inference/em_invoker/azure_openai_em_invoker.pyi,sha256=1HgCMcw7Hqv2ah4v9ma1Ioa-PpI-v2g7MfuKxxk2ZPU,4473
+gllm_inference/em_invoker/em_invoker.pyi,sha256=hiH8FB5R-KxhI8Ds2htF3cjRcIcH92yHPcOdpgc4FDo,4341
+gllm_inference/em_invoker/google_em_invoker.pyi,sha256=LQDUdsnOrB6ihBh0VdoOZHzFrY5dE3wWgBsl_slr1SI,6067
 gllm_inference/em_invoker/langchain/__init__.pyi,sha256=VYGKE5OgU0my1RlhgzkU_A7-GLGnUDDnNFuctuRwILE,148
 gllm_inference/em_invoker/langchain/em_invoker_embeddings.pyi,sha256=6nASLqi0FXCpqyYPl7kM3g7hAW-xS5ZwsS3GFudns98,2347
-gllm_inference/em_invoker/langchain_em_invoker.pyi,sha256=bRrywoGXIyZYIye85WhSbt0Pz5Qft1FWcp8hog2ikfk,2627
-gllm_inference/em_invoker/openai_compatible_em_invoker.pyi,sha256=w1vhe5PCHgOs16qKhdomSPnD513WbIpbtzv_4yEVJTw,4852
-gllm_inference/em_invoker/openai_em_invoker.pyi,sha256=JXpVxNY26jrx3sraI1bGUQqut4zoRNd_ZmKzsT8VxQA,4120
+gllm_inference/em_invoker/langchain_em_invoker.pyi,sha256=HSusS_ef5VZ20HPBFe153olueYOGQ57hmKY4406LUcE,2577
+gllm_inference/em_invoker/openai_compatible_em_invoker.pyi,sha256=GkoL-Z_Rl5v_NwcbvhUE6quLr6F9KnUB6Sb9Q712hNM,4802
+gllm_inference/em_invoker/openai_em_invoker.pyi,sha256=Gdyh93f8Mi3JYqnV1sBnYk1LAZpcsP47JeLjWxqyins,4070
 gllm_inference/em_invoker/schema/__init__.pyi,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 gllm_inference/em_invoker/schema/twelvelabs.pyi,sha256=y9nfqvuzs0o1jlZ5bwNmJkMSmvHAwBElWNMU0wJNc64,334
 gllm_inference/em_invoker/schema/voyage.pyi,sha256=7d9hv8uCJ4NboRaqK77g5ekb0X78pPW__sz2EYeKJKw,219
-gllm_inference/em_invoker/twelevelabs_em_invoker.pyi,sha256=MtC-JC0QHxs00WQyU4M0buft3JalcuW68ZzE9JuEBR0,5014
-gllm_inference/em_invoker/voyage_em_invoker.pyi,sha256=tW79qfg-O_8HuH1OQUzIuWTQ70l9rJQE34mDQzSpDcg,5082
+gllm_inference/em_invoker/twelevelabs_em_invoker.pyi,sha256=9QctCbM6c9q5eHKIYEwVlhDmzcCysQupiT-4e8hllPE,4964
+gllm_inference/em_invoker/voyage_em_invoker.pyi,sha256=17NBfuii3FyBbNtNRT2y6FUMTMsDemanXnY1j_0dmZg,5032
 gllm_inference/exceptions/__init__.pyi,sha256=v9uxjW5DssIn7n_bKqT7L83CeqFET2Z45GFOvi78UuE,977
 gllm_inference/exceptions/error_parser.pyi,sha256=4RkVfS2Fl9kjz_h2bK9eoAeI-Y-VkHcUqXWj68BsYig,2393
 gllm_inference/exceptions/exceptions.pyi,sha256=5YRackwVNvyOJjOtiVszqu8q87s8ioXTa-XwaYmeiC4,4643
 gllm_inference/lm_invoker/__init__.pyi,sha256=8oUFgavpItvjRUqsLqLSHqQVIPyTKCgNQ_euf58H2zY,1104
-gllm_inference/lm_invoker/anthropic_lm_invoker.pyi,sha256=J_Y63fX1oQlDFcruD7QZuADovDj21lv-6uiWlRTaoto,14713
-gllm_inference/lm_invoker/azure_openai_lm_invoker.pyi,sha256=AUgRctJrMfh9QxnAygDplajcd8F7zavz44Qq6QurHi0,14421
-gllm_inference/lm_invoker/bedrock_lm_invoker.pyi,sha256=YZYK_zLRlmx97M14h0C8P1icuvWi718tgt-s0chXnaw,12350
-gllm_inference/lm_invoker/datasaur_lm_invoker.pyi,sha256=a3jbg8cAxx7hnV1iRYX13LkH8zU4HImnUjsQhYHHMWE,9063
-gllm_inference/lm_invoker/google_lm_invoker.pyi,sha256=-0cKbJaJ1egEnP5KcGemW8SAQUTeeYqv1tHDpbCQfc0,16407
-gllm_inference/lm_invoker/langchain_lm_invoker.pyi,sha256=35vw9SDgE1ucgG3ucePPALQpKhdWYP1jQ1T6_mIfv08,13068
-gllm_inference/lm_invoker/litellm_lm_invoker.pyi,sha256=RrCBEzxb3aQVYf3iQnj1POna4dT-3-042la0iKOUSYA,13087
-gllm_inference/lm_invoker/lm_invoker.pyi,sha256=Vj-FkS4wcAHfAdGTVSwIeNvW9j-SvUnRTXx7bhT1a7c,7572
-gllm_inference/lm_invoker/openai_compatible_lm_invoker.pyi,sha256=fllMyWvAebhua2j1MRU3-ie4B8TVUJZ3pPFKtYE5R4k,14809
-gllm_inference/lm_invoker/openai_lm_invoker.pyi,sha256=E2gIVn4k_6WvI-7VWvc6BkKItCY7A7HpQhIo2XPQMKw,19375
+gllm_inference/lm_invoker/anthropic_lm_invoker.pyi,sha256=Ihq8U4w_efN74gf-nDFUY_zXO4Hf4KYZ2YhWcu9MqNA,14663
+gllm_inference/lm_invoker/azure_openai_lm_invoker.pyi,sha256=98ecMvAsEKEYNS1HehL3xTKT3MLdI9AfmJVOi9jeHXM,14407
+gllm_inference/lm_invoker/bedrock_lm_invoker.pyi,sha256=0ZeBJ_PWuFY2dOlON11VsI4BENhFJhK4md-8DwNpcyU,12300
+gllm_inference/lm_invoker/datasaur_lm_invoker.pyi,sha256=EzovSc7FXXZLoRrOuA2rFHmBAcBEohoBaUkAFk7hQig,9049
+gllm_inference/lm_invoker/google_lm_invoker.pyi,sha256=dRK3_--UQEBEMkLItav9MeTXp5p-iM-pqzt2JnwocPU,16393
+gllm_inference/lm_invoker/langchain_lm_invoker.pyi,sha256=2p-JM7XojIi20wxKvu1n6eJWVeQz6P069M3Quoe5ViE,13018
+gllm_inference/lm_invoker/litellm_lm_invoker.pyi,sha256=ad_tVyOjATiFyYF-f_1vX2Sl4BXy-fDXEW_D8fD1Db8,13037
+gllm_inference/lm_invoker/lm_invoker.pyi,sha256=XvcR2AAH_vnIdZ3cgDer59ZT724mW22_OSF8mpeo_kk,7610
+gllm_inference/lm_invoker/openai_compatible_lm_invoker.pyi,sha256=gHYFz7S8kTq1qkh-kqXOxkBAydUWxXJyoyNkL90Kvig,14759
+gllm_inference/lm_invoker/openai_lm_invoker.pyi,sha256=EsWW-XPmpmQ_27jFezJgT5lbpQf1Hj4O6MCvN63M3v0,19325
 gllm_inference/lm_invoker/schema/__init__.pyi,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 gllm_inference/lm_invoker/schema/anthropic.pyi,sha256=Y7NAjB7H0Wmiwv3YGZuG83K1rOCjSDB6hoZ6LjndWss,971
 gllm_inference/lm_invoker/schema/bedrock.pyi,sha256=FOUMZkBi6KRa__mYoy3FNJ5sP0EC6rgLuhoijnwelIg,927
@@ -71,7 +71,7 @@ gllm_inference/request_processor/__init__.pyi,sha256=hVnfdNZnkTBJHnmLtN3Na4ANP0y
 gllm_inference/request_processor/lm_request_processor.pyi,sha256=7pVNb2GwITb1jTflZP498qZ321G15b16jayZuuhuO1o,5424
 gllm_inference/request_processor/uses_lm_mixin.pyi,sha256=He-ytjwv2H5Hn312WFBAlBK96ALKTtDO3AT_80hCGTg,2321
 gllm_inference/schema/__init__.pyi,sha256=bJeO_4fCFPmYrVZ-4RQw0eQn62X228pB0w2tYDM-oek,1238
-gllm_inference/schema/attachment.pyi,sha256=FgVab4bU7KpGovexsOOQF4iuCrFbmV8gJDK3VP45hUU,2810
+gllm_inference/schema/attachment.pyi,sha256=jApuzjOHJDCz4lr4MlHzBgIndh559nbWu2Xp1fk3hso,3297
 gllm_inference/schema/code_exec_result.pyi,sha256=ZTHh6JtRrPIdQ059P1UAiD2L-tAO1_S5YcMsAXfJ5A0,559
 gllm_inference/schema/enums.pyi,sha256=Nvc_Qsd8yyiH_tCtymN39O6EZ0DT_wxYHmlKLllxC00,605
 gllm_inference/schema/lm_output.pyi,sha256=GafJV0KeD-VSwWkwG1oz-uruXrQ7KDZTuoojPCBRpg8,1956
@@ -86,8 +86,8 @@ gllm_inference/utils/__init__.pyi,sha256=npmBmmlBv7cPHMg1hdL3S2_RelD6vk_LhCsGELh
 gllm_inference/utils/langchain.pyi,sha256=VluQiHkGigDdqLUbhB6vnXiISCP5hHqV0qokYY6dC1A,1164
 gllm_inference/utils/validation.pyi,sha256=toxBtRp-VItC_X7sNi-GDd7sjibBdWMrR0q01OI2D7k,385
 gllm_inference.build/.gitignore,sha256=aEiIwOuxfzdCmLZe4oB1JsBmCUxwG8x-u-HBCV9JT8E,1
-gllm_inference.cpython-313-darwin.so,sha256=f2ej4tXy35HQvzaHJJeEqtSbz16v8fXL_KwEkScIJic,3694400
+gllm_inference.cpython-313-darwin.so,sha256=CU5l3wmqrwrKCtgPQEJwBVcu_QpFBKU3sk8Na7gfsMU,3672136
 gllm_inference.pyi,sha256=xOoh8lTQxXc6A4XYKBobWn8RJNszAlinAmbHPJyqi30,3315
-gllm_inference_binary-0.5.2.dist-info/METADATA,sha256=g7UfIlC7oMewzWaJ8cI7SKlNYdEKL11XARSYJGpyEbM,4531
-gllm_inference_binary-0.5.2.dist-info/WHEEL,sha256=PCOZcL_jcbAVhuFR5ylE4Mr-7HPGHAcfJk9OpuMh7RQ,107
-gllm_inference_binary-0.5.2.dist-info/RECORD,,
+gllm_inference_binary-0.5.4.dist-info/METADATA,sha256=pDdTHJqMvgTxEDuOVk85fsZsvKVBbUaLd-QB7J-jRVM,4531
+gllm_inference_binary-0.5.4.dist-info/WHEEL,sha256=PCOZcL_jcbAVhuFR5ylE4Mr-7HPGHAcfJk9OpuMh7RQ,107
+gllm_inference_binary-0.5.4.dist-info/RECORD,,