gllm-inference-binary 0.4.62__cp311-cp311-win_amd64.whl → 0.5.0__cp311-cp311-win_amd64.whl

gllm_inference/prompt_builder/prompt_builder.pyi

@@ -1,13 +1,8 @@
-import abc
 from _typeshed import Incomplete
-from abc import ABC
-from gllm_inference.constants import MESSAGE_TUPLE_LENGTH as MESSAGE_TUPLE_LENGTH
-from gllm_inference.schema import Attachment as Attachment, MultimodalContent as MultimodalContent, MultimodalPrompt as MultimodalPrompt, PromptRole as PromptRole
-from gllm_inference.utils import get_prompt_keys as get_prompt_keys, validate_prompt_builder_kwargs as validate_prompt_builder_kwargs
+from gllm_inference.schema import Message as Message, MessageContent as MessageContent, MessageRole as MessageRole
 from typing import Any
 
-__ALL__: str
-DEPRECATION_MESSAGE: str
+KEY_EXTRACTOR_REGEX: Incomplete
 
 class PromptBuilder:
     """A prompt builder class used in Gen AI applications.
@@ -36,111 +31,23 @@ class PromptBuilder:
         Raises:
             ValueError: If both `system_template` and `user_template` are empty.
         """
-    def format(self, history: MultimodalPrompt | None = None, extra_contents: list[MultimodalContent] | None = None, attachments: list[Attachment] | None = None, **kwargs: Any) -> MultimodalPrompt:
-        """Formats the prompt templates into a `MultimodalPrompt`.
+    def format(self, history: list[Message] | None = None, extra_contents: list[MessageContent] | None = None, **kwargs: Any) -> list[Message]:
+        """Formats the prompt templates into a list of messages.
 
         This method processes each prompt template, replacing the placeholders in the template content with the
         corresponding values from `kwargs`. If any required key is missing from `kwargs`, it raises a `ValueError`.
-        It also handles the provided history and extra contents. It formats the prompt as a `MultimodalPrompt`.
+        It also handles the provided history and extra contents. It formats the prompt as a list of messages.
 
         Args:
-            history (MultimodalPrompt | None, optional): The optional history to be included in the prompt.
-                Defaults to None.
-            extra_contents (list[MultimodalContent] | None, optional): The optional extra contents to be included in
-                the user message. Defaults to None.
-            attachments (list[Attachment] | None, optional): Deprecated parameter to handle attachments.
-                Will be removed in v0.5.0. Defaults to None.
+            history (list[Message] | None, optional): The history to be included in the prompt. Defaults to None.
+            extra_contents (list[MessageContent] | None, optional): The extra contents to be included in the user
+                message. Defaults to None.
             **kwargs (Any): A dictionary of placeholder values to be injected into the prompt templates.
                 Values must be either a string or an object that can be serialized to a string.
 
         Returns:
-            MultimodalPrompt: A multimodal prompt.
+            list[Message]: A formatted list of messages.
 
         Raises:
             ValueError: If a required key for the prompt template is missing from `kwargs`.
         """
-
-class BasePromptBuilder(ABC, metaclass=abc.ABCMeta):
-    """A base class for prompt builders used in Gen AI applications.
-
-    Attributes:
-        system_template (str): The system prompt template. May contain placeholders enclosed in curly braces `{}`.
-        user_template (str): The user prompt template. May contain placeholders enclosed in curly braces `{}`.
-        prompt_key_set (set[str]): A set of expected keys that must be present in the prompt templates.
-        ignore_extra_keys (bool): Whether to ignore extra keys when formatting the prompt.
-    """
-    system_template: Incomplete
-    user_template: Incomplete
-    prompt_key_set: Incomplete
-    ignore_extra_keys: Incomplete
-    logger: Incomplete
-    def __init__(self, system_template: str = '', user_template: str = '', ignore_extra_keys: bool = False) -> None:
-        """Initializes a new instance of the BasePromptBuilder class.
-
-        Args:
-            system_template (str, optional): The system prompt template. May contain placeholders enclosed in curly
-                braces `{}`. Defaults to an empty string.
-            user_template (str, optional): The user prompt template. May contain placeholders enclosed in curly
-                braces `{}`. Defaults to an empty string.
-            ignore_extra_keys (bool, optional): Whether to ignore extra keys when formatting the prompt.
-                Defaults to False.
-
-        Raises:
-            ValueError: If both `system_template` and `user_template` are empty.
-        """
-    def format_as_message_list(self, history: list[tuple[PromptRole, list[Any] | str]] | None = None, attachments: list[Attachment] | None = None, system_multimodal_contents: list[Any] | None = None, user_multimodal_contents: list[Any] | None = None, is_multimodal: bool | None = None, **kwargs: Any) -> list[tuple[PromptRole, list[Any] | str]]:
-        """Formats the prompt templates as a list of message tuples (role, formatted content).
-
-        This method processes each prompt template, replacing the placeholders in the template content with the
-        corresponding values from `kwargs`. If a required key is missing from `kwargs`, it raises a `ValueError`. It
-        returns a list of tuples, where each tuple consists of a role and the corresponding formatted prompt content.
-
-        Args:
-            history (list[tuple[PromptRole, list[Any] | str]] | None, optional): The optional chat history to be
-                included in the prompt. Defaults to None.
-            attachments (list[Attachment] | None, optional): The optional attachments to be included in the prompt.
-                Defaults to None.
-            system_multimodal_contents (list[Any] | None, optional): Deprecated parameter to handle attachments.
-                Will be removed in v0.5.0. Defaults to None.
-            user_multimodal_contents (list[Any] | None, optional): Deprecated parameter to handle attachments.
-                Will be removed in v0.5.0. Defaults to None.
-            is_multimodal (bool | None, optional): Whether the prompt supports multimodal inputs. Will be deprecated in
-                v0.5.0, in which multimodality will always be True. Defaults to None.
-            **kwargs (Any): A dictionary of placeholder values to be injected into the prompt templates.
-
-        Returns:
-            list[tuple[PromptRole, list[Any] | str]]: A list of tuples, each containing a role and the corresponding
-                formatted prompt content.
-
-        Raises:
-            ValueError: If a required key for the prompt template is missing from `kwargs`.
-            ValueError: If multimodal contents are provided when `is_multimodal` is False.
-        """
-    def format_as_string(self, history: list[tuple[PromptRole, str]] | None = None, **kwargs: Any) -> str:
-        """Formats the prompt as a string.
-
-        This method formats the prompt as a string by first converting the prompt templates to a list of messages and
-        then formatting the message list as a string.
-
-        Args:
-            history (list[tuple[PromptRole, str]] | None, optional): The optional chat history to be included in the
-                prompt. Defaults to None.
-            **kwargs (Any): A dictionary of placeholder values to be injected into the prompt templates.
-
-        Returns:
-            str: The formatted prompt with the placeholders replaced by the provided values.
-
-        Raises:
-            ValueError: If any required key is missing or there are extra keys in the kwargs.
-        """
-    @property
-    def compatible_model_list(self) -> list[str]:
-        '''Returns the list of compatible models for the prompt builder.
-
-        This property returns the set of models that the prompt builder is compatible with. If the builder is
-        model-specific, it returns the list of models in `_compatible_model_list`. Otherwise, it returns a list
-        containing `"All"` to indicate compatibility with all models.
-
-        Returns:
-            list[str]: A list of compatible model names, or `["All"]` if the prompt builder is not model-specific.
-        '''

gllm_inference/prompt_formatter/agnostic_prompt_formatter.pyi

@@ -1,6 +1,6 @@
 from _typeshed import Incomplete
 from gllm_inference.prompt_formatter.prompt_formatter import BasePromptFormatter as BasePromptFormatter
-from gllm_inference.schema import PromptRole as PromptRole
+from gllm_inference.schema import MessageRole as MessageRole
 
 class AgnosticPromptFormatter(BasePromptFormatter):
     '''A prompt formatter that formats prompt without any specific model formatting.
@@ -19,9 +19,9 @@ class AgnosticPromptFormatter(BasePromptFormatter):
         Usage example:
         ```python
         prompt = [
-            (PromptRole.USER, ["Hello", "how are you?"]),
-            (PromptRole.ASSISTANT, ["I\'m fine", "thank you!"]),
-            (PromptRole.USER, ["What is the capital of France?"]),
+            (MessageRole.USER, ["Hello", "how are you?"]),
+            (MessageRole.ASSISTANT, ["I\'m fine", "thank you!"]),
+            (MessageRole.USER, ["What is the capital of France?"]),
         ]
         prompt_formatter = AgnosticPromptFormatter(
             message_separator="\\n###\\n",

gllm_inference/prompt_formatter/huggingface_prompt_formatter.pyi

@@ -1,5 +1,5 @@
 from gllm_inference.prompt_formatter.prompt_formatter import BasePromptFormatter as BasePromptFormatter
-from gllm_inference.schema import PromptRole as PromptRole
+from gllm_inference.schema import MessageRole as MessageRole
 
 TOKENIZER_LOAD_ERROR_MESSAGE: str
 
@@ -22,9 +22,9 @@ class HuggingFacePromptFormatter(BasePromptFormatter):
         Usage example:
         ```python
         prompt = [
-            (PromptRole.USER, ["Hello", "how are you?"]),
-            (PromptRole.ASSISTANT, ["I\'m fine", "thank you!"]),
-            (PromptRole.USER, ["What is the capital of France?"]),
+            (MessageRole.USER, ["Hello", "how are you?"]),
+            (MessageRole.ASSISTANT, ["I\'m fine", "thank you!"]),
+            (MessageRole.USER, ["What is the capital of France?"]),
         ]
         prompt_formatter = HuggingFacePromptFormatter(
             model_name_or_path="mistralai/Mistral-7B-Instruct-v0.1",

gllm_inference/prompt_formatter/llama_prompt_formatter.pyi

@@ -19,9 +19,9 @@ class LlamaPromptFormatter(HuggingFacePromptFormatter):
         Usage example:
         ```python
         prompt = [
-            (PromptRole.USER, ["Hello", "how are you?"]),
-            (PromptRole.ASSISTANT, ["I\'m fine", "thank you!"]),
-            (PromptRole.USER, ["What is the capital of France?"]),
+            (MessageRole.USER, ["Hello", "how are you?"]),
+            (MessageRole.ASSISTANT, ["I\'m fine", "thank you!"]),
+            (MessageRole.USER, ["What is the capital of France?"]),
         ]
         prompt_formatter = LlamaPromptFormatter(
             model_name_or_path="meta-llama/Meta-Llama-3.1-8B-Instruct",

gllm_inference/prompt_formatter/mistral_prompt_formatter.pyi

@@ -19,9 +19,9 @@ class MistralPromptFormatter(HuggingFacePromptFormatter):
         Usage example:
         ```python
         prompt = [
-            (PromptRole.USER, ["Hello", "how are you?"]),
-            (PromptRole.ASSISTANT, ["I\'m fine", "thank you!"]),
-            (PromptRole.USER, ["What is the capital of France?"]),
+            (MessageRole.USER, ["Hello", "how are you?"]),
+            (MessageRole.ASSISTANT, ["I\'m fine", "thank you!"]),
+            (MessageRole.USER, ["What is the capital of France?"]),
         ]
         prompt_formatter = MistralPromptFormatter(
             model_name_or_path="mistralai/Mistral-7B-Instruct-v0.1",

gllm_inference/prompt_formatter/openai_prompt_formatter.pyi

@@ -1,5 +1,5 @@
 from gllm_inference.prompt_formatter.prompt_formatter import BasePromptFormatter as BasePromptFormatter
-from gllm_inference.schema import PromptRole as PromptRole
+from gllm_inference.schema import MessageRole as MessageRole
 
 class OpenAIPromptFormatter(BasePromptFormatter):
     '''A prompt formatter that formats prompt with OpenAI\'s specific formatting.
@@ -16,9 +16,9 @@ class OpenAIPromptFormatter(BasePromptFormatter):
         Usage example:
         ```python
         prompt = [
-            (PromptRole.USER, ["Hello", "how are you?"]),
-            (PromptRole.ASSISTANT, ["I\'m fine", "thank you!"]),
-            (PromptRole.USER, ["What is the capital of France?"]),
+            (MessageRole.USER, ["Hello", "how are you?"]),
+            (MessageRole.ASSISTANT, ["I\'m fine", "thank you!"]),
+            (MessageRole.USER, ["What is the capital of France?"]),
         ]
         prompt_formatter = OpenAIPromptFormatter(
             content_separator="---"

gllm_inference/prompt_formatter/prompt_formatter.pyi

@@ -1,7 +1,7 @@
 import abc
 from _typeshed import Incomplete
 from abc import ABC
-from gllm_inference.schema import MultimodalContent as MultimodalContent, MultimodalPrompt as MultimodalPrompt, PromptRole as PromptRole
+from gllm_inference.schema import Message as Message, MessageContent as MessageContent, MessageRole as MessageRole
 
 class BasePromptFormatter(ABC, metaclass=abc.ABCMeta):
     """A base class for prompt formatters used in Gen AI applications.
@@ -19,12 +19,12 @@ class BasePromptFormatter(ABC, metaclass=abc.ABCMeta):
             content_separator (str, optional): The separator to be used between the string in a single message.
                 Defaults to "\\n".
         '''
-    def format(self, prompt: MultimodalPrompt) -> str:
+    def format(self, messages: list[Message]) -> str:
         """Formats the prompt as a string.
 
         Args:
-            prompt (MultimodalPrompt): The prompt to be formatted as a string.
+            messages (list[Message]): The messages to be formatted as a string.
 
         Returns:
-            str: The formatted prompt as a string.
+            str: The formatted messages as a string.
         """

gllm_inference/request_processor/lm_request_processor.pyi

@@ -1,11 +1,9 @@
 from _typeshed import Incomplete
 from gllm_core.event import EventEmitter as EventEmitter
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
-from gllm_inference.multimodal_lm_invoker.multimodal_lm_invoker import BaseMultimodalLMInvoker as BaseMultimodalLMInvoker
-from gllm_inference.multimodal_prompt_builder import MultimodalPromptBuilder as MultimodalPromptBuilder
 from gllm_inference.output_parser.output_parser import BaseOutputParser as BaseOutputParser
-from gllm_inference.prompt_builder.prompt_builder import BasePromptBuilder as BasePromptBuilder, PromptBuilder as PromptBuilder
-from gllm_inference.schema import Attachment as Attachment, LMOutput as LMOutput, MultimodalContent as MultimodalContent, MultimodalOutput as MultimodalOutput, MultimodalPrompt as MultimodalPrompt, PromptRole as PromptRole, ResponseSchema as ResponseSchema, ToolCall as ToolCall, ToolResult as ToolResult
+from gllm_inference.prompt_builder.prompt_builder import PromptBuilder as PromptBuilder
+from gllm_inference.schema import LMOutput as LMOutput, Message as Message, MessageContent as MessageContent, ResponseSchema as ResponseSchema, ToolCall as ToolCall, ToolResult as ToolResult
 from langchain_core.tools import Tool as Tool
 from typing import Any
 
@@ -17,10 +15,8 @@ class LMRequestProcessor:
     the inference process in Gen AI applications.
 
     Attributes:
-        prompt_builder (PromptBuilder | BasePromptBuilder | MultimodalPromptBuilder): The prompt builder used to
-            format the prompt.
-        lm_invoker (BaseLMInvoker | BaseMultimodalLMInvoker): The language model invoker that handles the model
-            inference.
+        prompt_builder (PromptBuilder): The prompt builder used to format the prompt.
+        lm_invoker (BaseLMInvoker): The language model invoker that handles the model inference.
         output_parser (BaseOutputParser | None): The optional parser to process the model's output, if any.
         tool_dict (dict[str, Tool]): A dictionary of tools provided to the language model to enable tool calling,
             if any. The dictionary maps the tool name to the tools themselves.
@@ -29,18 +25,14 @@ class LMRequestProcessor:
     lm_invoker: Incomplete
     output_parser: Incomplete
     tool_dict: Incomplete
-    def __init__(self, prompt_builder: PromptBuilder | BasePromptBuilder | MultimodalPromptBuilder, lm_invoker: BaseLMInvoker | BaseMultimodalLMInvoker, output_parser: BaseOutputParser | None = None) -> None:
+    def __init__(self, prompt_builder: PromptBuilder, lm_invoker: BaseLMInvoker, output_parser: BaseOutputParser | None = None) -> None:
         """Initializes a new instance of the LMRequestProcessor class.
 
         Args:
-            prompt_builder (PromptBuilder | BasePromptBuilder | MultimodalPromptBuilder): The prompt builder used to
-                format the prompt.
-            lm_invoker (BaseLMInvoker | BaseMultimodalLMInvoker): The language model invoker that handles the model
-                inference.
+            prompt_builder (PromptBuilder): The prompt builder used to format the prompt.
+            lm_invoker (BaseLMInvoker): The language model invoker that handles the model inference.
             output_parser (BaseOutputParser, optional): An optional parser to process the model's output.
                 Defaults to None.
-
-        WARNING: Support for MultimodalPromptBuilder is deprecated and will be removed in version 0.5.0.
         """
     def set_tools(self, tools: list[Tool]) -> None:
         """Sets the tools for the LM invoker.
@@ -68,7 +60,7 @@ class LMRequestProcessor:
 
         This method clears the response schema for the LM invoker.
         """
-    async def process(self, prompt_kwargs: dict[str, Any], history: MultimodalPrompt | None = None, extra_contents: list[MultimodalContent] | None = None, hyperparameters: dict[str, Any] | None = None, event_emitter: EventEmitter | None = None, attachments: list[Attachment] | None = None, system_multimodal_contents: list[Any] | None = None, user_multimodal_contents: list[Any] | None = None, auto_execute_tools: bool = True, max_lm_calls: int = 5) -> Any:
+    async def process(self, prompt_kwargs: dict[str, Any] | None = None, history: list[Message] | None = None, extra_contents: list[MessageContent] | None = None, hyperparameters: dict[str, Any] | None = None, event_emitter: EventEmitter | None = None, auto_execute_tools: bool = True, max_lm_calls: int = 5) -> Any:
         """Processes a language model inference request.
 
         This method processes the language model inference request as follows:
@@ -80,21 +72,16 @@ class LMRequestProcessor:
            LMOutput object, the output parser will process the `response` attribute of the LMOutput object.
 
         Args:
-            prompt_kwargs (dict[str, Any]): A dictionary of arguments used to format the prompt.
-            history (MultimodalPrompt | None, optional): A list of conversation history to be included in the prompt.
+            prompt_kwargs (dict[str, Any], optional): A dictionary of arguments used to format the prompt.
+                Defaults to None, in which case no arguments will be passed to the prompt builder.
+            history (list[Message] | None, optional): A list of conversation history to be included in the prompt.
                 Defaults to None.
-            extra_contents (list[MultimodalContent] | None, optional): A list of extra contents to be included in the
+            extra_contents (list[MessageContent] | None, optional): A list of extra contents to be included in the
                 user message. Defaults to None.
             hyperparameters (dict[str, Any] | None, optional): A dictionary of hyperparameters for the model invocation.
                 Defaults to None.
             event_emitter (EventEmitter | None, optional): An event emitter for streaming model outputs.
                 Defaults to None.
-            attachments (list[Attachment] | None, optional): Deprecated parameter to handle attachments.
-                Will be removed in v0.5.0. Defaults to None.
-            system_multimodal_contents (list[Any] | None, optional): Deprecated parameter to handle attachments.
-                Will be removed in v0.5.0. Defaults to None.
-            user_multimodal_contents (list[Any] | None, optional): Deprecated parameter to handle attachments.
-                Will be removed in v0.5.0. Defaults to None.
             auto_execute_tools (bool, optional): Whether to automatically execute tools if the LM invokes output
                 tool calls. Defaults to True.
             max_lm_calls (int, optional): The maximum number of times the language model can be invoked

gllm_inference/request_processor/uses_lm_mixin.pyi

@@ -1,8 +1,6 @@
 from gllm_inference.lm_invoker.lm_invoker import BaseLMInvoker as BaseLMInvoker
-from gllm_inference.multimodal_lm_invoker.multimodal_lm_invoker import BaseMultimodalLMInvoker as BaseMultimodalLMInvoker
-from gllm_inference.multimodal_prompt_builder import MultimodalPromptBuilder as MultimodalPromptBuilder
 from gllm_inference.output_parser.output_parser import BaseOutputParser as BaseOutputParser
-from gllm_inference.prompt_builder.prompt_builder import BasePromptBuilder as BasePromptBuilder, PromptBuilder as PromptBuilder
+from gllm_inference.prompt_builder.prompt_builder import PromptBuilder as PromptBuilder
 from gllm_inference.request_processor.lm_request_processor import LMRequestProcessor as LMRequestProcessor
 from typing import Any
 
@@ -35,22 +33,18 @@ class UsesLM:
         first argument.
     '''
     @classmethod
-    def from_lm_components(cls, prompt_builder: PromptBuilder | BasePromptBuilder | MultimodalPromptBuilder, lm_invoker: BaseLMInvoker | BaseMultimodalLMInvoker, output_parser: BaseOutputParser | None = None, **kwargs: Any):
+    def from_lm_components(cls, prompt_builder: PromptBuilder, lm_invoker: BaseLMInvoker, output_parser: BaseOutputParser | None = None, **kwargs: Any):
         """Creates an instance by initializing LMRequestProcessor with given components.
 
         This method is a shortcut to initialize the class by providing the LMRequestProcessor components directly.
 
         Args:
-            prompt_builder (PromptBuilder | BasePromptBuilder | MultimodalPromptBuilder): The prompt builder used to
-                format the prompt.
-            lm_invoker (BaseLMInvoker | BaseMultimodalLMInvoker): The language model invoker that handles the model
-                inference.
+            prompt_builder (PromptBuilder): The prompt builder used to format the prompt.
+            lm_invoker (BaseLMInvoker): The language model invoker that handles the model inference.
             output_parser (BaseOutputParser, optional): An optional parser to process the model's output.
                 Defaults to None.
             **kwargs (Any): Additional keyword arguments to be passed to the class constructor.
 
         Returns:
             An instance of the class that mixes in this mixin.
-
-        WARNING: Support for MultimodalPromptBuilder is deprecated and will be removed in version 0.5.0.
         """

gllm_inference/schema/__init__.pyi

@@ -1,6 +1,13 @@
-from gllm_inference.schema.enums import AttachmentType as AttachmentType, EmitDataType as EmitDataType, PromptRole as PromptRole
+from gllm_inference.schema.attachment import Attachment as Attachment
+from gllm_inference.schema.code_exec_result import CodeExecResult as CodeExecResult
+from gllm_inference.schema.enums import AttachmentType as AttachmentType, EmitDataType as EmitDataType, MessageRole as MessageRole
+from gllm_inference.schema.lm_output import LMOutput as LMOutput
+from gllm_inference.schema.message import Message as Message
 from gllm_inference.schema.model_id import ModelId as ModelId, ModelProvider as ModelProvider
-from gllm_inference.schema.model_io import Attachment as Attachment, CodeExecResult as CodeExecResult, ContentPlaceholder as ContentPlaceholder, LMOutput as LMOutput, Reasoning as Reasoning, TokenUsage as TokenUsage, ToolCall as ToolCall, ToolResult as ToolResult
-from gllm_inference.schema.type_alias import EMContent as EMContent, ErrorResponse as ErrorResponse, MultimodalContent as MultimodalContent, MultimodalOutput as MultimodalOutput, MultimodalPrompt as MultimodalPrompt, ResponseSchema as ResponseSchema, UnimodalContent as UnimodalContent, UnimodalPrompt as UnimodalPrompt, Vector as Vector
+from gllm_inference.schema.reasoning import Reasoning as Reasoning
+from gllm_inference.schema.token_usage import TokenUsage as TokenUsage
+from gllm_inference.schema.tool_call import ToolCall as ToolCall
+from gllm_inference.schema.tool_result import ToolResult as ToolResult
+from gllm_inference.schema.type_alias import EMContent as EMContent, ErrorResponse as ErrorResponse, MessageContent as MessageContent, ResponseSchema as ResponseSchema, Vector as Vector
 
-__all__ = ['Attachment', 'AttachmentType', 'CodeExecResult', 'ContentPlaceholder', 'EMContent', 'EmitDataType', 'ErrorResponse', 'LMOutput', 'ModelId', 'ModelProvider', 'MultimodalContent', 'MultimodalOutput', 'MultimodalPrompt', 'PromptRole', 'Reasoning', 'ResponseSchema', 'TokenUsage', 'ToolCall', 'ToolResult', 'UnimodalContent', 'UnimodalPrompt', 'Vector']
+__all__ = ['Attachment', 'AttachmentType', 'CodeExecResult', 'EMContent', 'EmitDataType', 'ErrorResponse', 'MessageContent', 'LMOutput', 'ModelId', 'ModelProvider', 'Message', 'MessageRole', 'Reasoning', 'ResponseSchema', 'TokenUsage', 'ToolCall', 'ToolResult', 'Vector']

gllm_inference/schema/attachment.pyi

@@ -0,0 +1,76 @@
+from _typeshed import Incomplete
+from gllm_inference.constants import HEX_REPR_LENGTH as HEX_REPR_LENGTH
+from pydantic import BaseModel
+
+logger: Incomplete
+
+class Attachment(BaseModel):
+    """Defines a file attachment schema.
+
+    Attributes:
+        data (bytes): The content data of the file attachment.
+        filename (str): The filename of the file attachment.
+        mime_type (str): The mime type of the file attachment.
+        extension (str): The extension of the file attachment.
+        url (str | None): The URL of the file attachment. Defaults to None.
+    """
+    data: bytes
+    filename: str
+    mime_type: str
+    extension: str
+    url: str | None
+    @classmethod
+    def from_bytes(cls, bytes: bytes, filename: str | None = None) -> Attachment:
+        """Creates an Attachment from bytes.
+
+        Args:
+            bytes (bytes): The bytes 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 extension.
+
+        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]).
+
+        Args:
+            data_url (str): The data URL 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_url(cls, url: str, filename: str | None = None) -> Attachment:
+        """Creates an Attachment from a URL.
+
+        Args:
+            url (str): The URL 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 URL.
+
+        Returns:
+            Attachment: The instantiated Attachment.
+        """
+    @classmethod
+    def from_path(cls, path: str, filename: str | None = None) -> Attachment:
+        """Creates an Attachment from a path.
+
+        Args:
+            path (str): The path to the file.
+            filename (str | None, optional): The filename of the file. Defaults to None,
+                in which case the filename will be derived from the path.
+
+        Returns:
+            Attachment: The instantiated Attachment.
+        """
+    def write_to_file(self, path: str | None = None) -> None:
+        """Writes the Attachment to a file.
+
+        Args:
+            path (str | None, optional): The path to the file. Defaults to None,
+                in which case the filename will be used as the path.
+        """

gllm_inference/schema/code_exec_result.pyi

@@ -0,0 +1,14 @@
+from gllm_inference.schema.attachment import Attachment as Attachment
+from pydantic import BaseModel
+
+class CodeExecResult(BaseModel):
+    """Defines a code execution result when a language model is configured to execute code.
+
+    Attributes:
+        id (str): The ID of the code execution. Defaults to an empty string.
+        code (str): The executed code. Defaults to an empty string.
+        output (list[str | Attachment]): The output of the executed code. Defaults to an empty list.
+    """
+    id: str
+    code: str
+    output: list[str | Attachment]

gllm_inference/schema/enums.pyi

@@ -1,5 +1,12 @@
 from enum import StrEnum
 
+class AttachmentType(StrEnum):
+    """Defines valid attachment types."""
+    AUDIO = 'audio'
+    DOCUMENT = 'document'
+    IMAGE = 'image'
+    VIDEO = 'video'
+
 class EmitDataType(StrEnum):
     """Defines valid data types for emitting events."""
     ACTIVITY = 'activity'
@@ -10,15 +17,8 @@ class EmitDataType(StrEnum):
     THINKING_START = 'thinking_start'
     THINKING_END = 'thinking_end'
 
-class PromptRole(StrEnum):
-    """Defines valid prompt roles."""
+class MessageRole(StrEnum):
+    """Defines valid message roles."""
     SYSTEM = 'system'
     USER = 'user'
     ASSISTANT = 'assistant'
-
-class AttachmentType(StrEnum):
-    """Defines valid attachment types."""
-    AUDIO = 'audio'
-    DOCUMENT = 'document'
-    IMAGE = 'image'
-    VIDEO = 'video'

gllm_inference/schema/lm_output.pyi

@@ -0,0 +1,36 @@
+from gllm_core.schema import Chunk as Chunk
+from gllm_inference.schema.code_exec_result import CodeExecResult as CodeExecResult
+from gllm_inference.schema.reasoning import Reasoning as Reasoning
+from gllm_inference.schema.token_usage import TokenUsage as TokenUsage
+from gllm_inference.schema.tool_call import ToolCall as ToolCall
+from pydantic import BaseModel
+from typing import Any
+
+class LMOutput(BaseModel):
+    """Defines the output of a language model.
+
+    Attributes:
+        response (str): The text response. Defaults to an empty string.
+        tool_calls (list[ToolCall]): The tool calls, if the language model decides to invoke tools.
+            Defaults to an empty list.
+        structured_output (dict[str, Any] | BaseModel | None): The structured output, if a response schema is defined
+            for the language model. Defaults to None.
+        token_usage (TokenUsage | None): The token usage analytics, if requested. Defaults to None.
+        duration (float | None): The duration of the invocation in seconds, if requested. Defaults to None.
+        finish_details (dict[str, Any]): The details about how the generation finished, if requested.
+            Defaults to an empty dictionary.
+        reasoning (list[Reasoning]): The reasoning, if the language model is configured to output reasoning.
+            Defaults to an empty list.
+        citations (list[Chunk]): The citations, if the language model outputs citations. Defaults to an empty list.
+        code_exec_results (list[CodeExecResult]): The code execution results, if the language model decides to
+            execute code. Defaults to an empty list.
+    """
+    response: str
+    tool_calls: list[ToolCall]
+    structured_output: dict[str, Any] | BaseModel | None
+    token_usage: TokenUsage | None
+    duration: float | None
+    finish_details: dict[str, Any]
+    reasoning: list[Reasoning]
+    citations: list[Chunk]
+    code_exec_results: list[CodeExecResult]

gllm_inference/schema/message.pyi

@@ -0,0 +1,52 @@
+from gllm_inference.schema.enums import MessageRole as MessageRole
+from gllm_inference.schema.type_alias import MessageContent as MessageContent
+from pydantic import BaseModel
+from typing import Any
+
+class Message(BaseModel):
+    """Defines a message schema to be used as inputs for a language model.
+
+    Attributes:
+        role (MessageRole): The role of the message.
+        contents (list[MessageContent]): The contents of the message.
+        metadata (dict[str, Any]): The metadata of the message.
+    """
+    role: MessageRole
+    contents: list[MessageContent]
+    metadata: dict[str, Any]
+    @classmethod
+    def system(cls, contents: MessageContent | list[MessageContent], metadata: dict[str, Any] | None = None) -> Message:
+        """Create a system message.
+
+        Args:
+            contents (MessageContent | list[MessageContent]): The message contents.
+                If a single content is provided, it will be wrapped in a list.
+            metadata (dict[str, Any], optional): Additional metadata for the message. Defaults to None.
+
+        Returns:
+            Message: A new message with SYSTEM role.
+        """
+    @classmethod
+    def user(cls, contents: MessageContent | list[MessageContent], metadata: dict[str, Any] | None = None) -> Message:
+        """Create a user message.
+
+        Args:
+            contents (MessageContent | list[MessageContent]): The message contents.
+                If a single content is provided, it will be wrapped in a list.
+            metadata (dict[str, Any], optional): Additional metadata for the message. Defaults to None.
+
+        Returns:
+            Message: A new message with USER role.
+        """
+    @classmethod
+    def assistant(cls, contents: MessageContent | list[MessageContent], metadata: dict[str, Any] | None = None) -> Message:
+        """Create an assistant message.
+
+        Args:
+            contents (MessageContent | list[MessageContent]): The message contents.
+                If a single content is provided, it will be wrapped in a list.
+            metadata (dict[str, Any], optional): Additional metadata for the message. Defaults to None.
+
+        Returns:
+            Message: A new message with ASSISTANT role.
+        """

gllm_inference/schema/model_id.pyi

@@ -1,5 +1,5 @@
 from enum import StrEnum
-from gllm_inference.utils.utils import validate_string_enum as validate_string_enum
+from gllm_inference.utils import validate_string_enum as validate_string_enum
 from pydantic import BaseModel
 
 PROVIDER_SEPARATOR: str

gllm_inference/schema/reasoning.pyi

@@ -0,0 +1,15 @@
+from pydantic import BaseModel
+
+class Reasoning(BaseModel):
+    """Defines a reasoning output when a language model is configured to use reasoning.
+
+    Attributes:
+        id (str): The ID of the reasoning output. Defaults to an empty string.
+        reasoning (str): The reasoning text. Defaults to an empty string.
+        type (str): The type of the reasoning output. Defaults to an empty string.
+        data (str): The additional data of the reasoning output. Defaults to an empty string.
+    """
+    id: str
+    reasoning: str
+    type: str
+    data: str