lite-agent 0.1.0__py3-none-any.whl → 0.3.0__py3-none-any.whl

lite_agent/chunk_handler.py

@@ -1,166 +0,0 @@
-from collections.abc import AsyncGenerator
-from typing import Literal, TypedDict
-
-import litellm
-from funcall import Funcall
-
-from open_agents.loggers import logger
-from open_agents.processors import StreamChunkProcessor
-from open_agents.processors.stream_chunk_processor import AssistantMessage
-
-
-class LiteLLMRawChunk(TypedDict):
-    """
-    Define the type of chunk
-    """
-
-    type: Literal["litellm_raw"]
-    raw: litellm.ModelResponseStream
-
-
-class UsageChunk(TypedDict):
-    """
-    Define the type of usage info chunk
-    """
-
-    type: Literal["usage"]
-    usage: litellm.Usage
-
-
-class FinalMessageChunk(TypedDict):
-    """
-    Define the type of final message chunk
-    """
-
-    type: Literal["final_message"]
-    message: AssistantMessage
-    finish_reason: Literal["stop", "tool_calls"]
-
-
-class ToolCallChunk(TypedDict):
-    """
-    Define the type of tool call chunk
-    """
-
-    type: Literal["tool_call"]
-    name: str
-    arguments: str
-
-
-class ToolCallResultChunk(TypedDict):
-    """
-    Define the type of tool call result chunk
-    """
-
-    type: Literal["tool_call_result"]
-    tool_call_id: str
-    name: str
-    content: str
-
-
-class ContentDeltaChunk(TypedDict):
-    """
-    Define the type of message chunk
-    """
-
-    type: Literal["content_delta"]
-    delta: str
-
-
-class ToolCallDeltaChunk(TypedDict):
-    """
-    Define the type of tool call delta chunk
-    """
-
-    type: Literal["tool_call_delta"]
-    tool_call_id: str
-    name: str
-    arguments_delta: str
-
-
-AgentChunk = LiteLLMRawChunk | UsageChunk | FinalMessageChunk | ToolCallChunk | ToolCallResultChunk | ContentDeltaChunk
-
-
-async def chunk_handler(
-    resp: litellm.CustomStreamWrapper,
-    fc: Funcall,
-) -> AsyncGenerator[AgentChunk, None]:
-    """
-    Optimized chunk handler
-
-    Args:
-        resp: LiteLLM streaming response wrapper
-        fc: function call handler
-
-    Yields:
-        litellm.ModelResponseStream: processed response chunk
-
-    Raises:
-        Exception: various exceptions during processing
-    """
-    processor = StreamChunkProcessor(fc)
-    async for chunk in resp:
-        if not isinstance(chunk, litellm.ModelResponseStream):
-            logger.debug("unexpected chunk type: %s", type(chunk))
-            logger.debug("chunk content: %s", chunk)
-            continue
-
-        # Handle usage info
-        if usage := processor.handle_usage_info(chunk):
-            yield UsageChunk(type="usage", usage=usage)
-            continue
-
-        # Get choice and delta data
-        if not chunk.choices:
-            yield LiteLLMRawChunk(type="litellm_raw", raw=chunk)
-            continue
-
-        choice = chunk.choices[0]
-        delta = choice.delta
-        if not processor.current_message:
-            processor.initialize_message(chunk, choice)
-        if delta.content:
-            yield ContentDeltaChunk(type="content_delta", delta=delta.content)
-        processor.update_content(delta.content)
-        processor.update_tool_calls(delta.tool_calls)
-        if delta.tool_calls:
-            for tool_call in delta.tool_calls:
-                if tool_call.function.arguments:
-                    yield ToolCallDeltaChunk(
-                        type="tool_call_delta",
-                        tool_call_id=processor.current_message.tool_calls[-1].id,
-                        name=processor.current_message.tool_calls[-1].function.name,
-                        arguments_delta=tool_call.function.arguments,
-                    )
-        # Check if finished
-        if choice.finish_reason and processor.current_message:
-            current_message = processor.finalize_message()
-            yield FinalMessageChunk(type="final_message", message=current_message, finish_reason=choice.finish_reason)
-            # New: check tool_calls and handle
-            tool_calls = current_message.tool_calls
-            if tool_calls:
-                # Execute each tool_call and yield result
-                for tool_call in tool_calls:
-                    try:
-                        yield ToolCallChunk(
-                            type="tool_call",
-                            name=tool_call.function.name,
-                            arguments=tool_call.function.arguments,
-                        )
-                        content = await fc.call_function_async(tool_call.function.name, tool_call.function.arguments)
-                        yield ToolCallResultChunk(
-                            type="tool_call_result",
-                            tool_call_id=tool_call.id,
-                            name=tool_call.function.name,
-                            content=str(content),
-                        )
-                    except Exception as e:  # noqa: PERF203
-                        logger.exception("Tool call %s failed", tool_call.id)
-                        yield ToolCallResultChunk(
-                            type="tool_call_result",
-                            tool_call_id=tool_call.id,
-                            name=tool_call.function.name,
-                            content=str(e),
-                        )
-            continue
-        yield LiteLLMRawChunk(type="litellm_raw", raw=chunk)

lite_agent/types.py

@@ -1,152 +0,0 @@
-from typing import Literal, TypedDict
-
-import litellm
-from pydantic import BaseModel
-
-
-class ToolCallFunction(BaseModel):
-    name: str
-    arguments: str | None = None
-
-
-class ToolCall(BaseModel):
-    type: Literal["function"]
-    function: ToolCallFunction
-    id: str
-
-
-class AssistantMessage(BaseModel):
-    id: str
-    role: Literal["assistant"] = "assistant"
-    content: str = ""
-    tool_calls: list[ToolCall] | None = None
-
-
-class Message(TypedDict):
-    role: str
-    content: str
-
-
-class UserMessageContentItemText(TypedDict):
-    type: Literal["text"]
-    text: str
-
-
-class UserMessageContentItemImageURLImageURL(TypedDict):
-    url: str
-
-
-class UserMessageContentItemImageURL(TypedDict):
-    type: Literal["image_url"]
-    image_url: UserMessageContentItemImageURLImageURL
-
-
-class AgentUserMessage(TypedDict):
-    role: Literal["user"] = "user"
-    content: str | list[UserMessageContentItemText | UserMessageContentItemImageURL]
-
-
-class AssistantMessageToolCallFunction(TypedDict):
-    name: str
-    arguments: str
-
-
-class AssistantMessageToolCall(TypedDict):
-    id: str
-    type: Literal["function"]
-    function: AssistantMessageToolCallFunction
-    tool_call_id: str
-
-
-class AgentAssistantMessage(TypedDict):
-    role: Literal["assistant"] = "assistant"
-    content: str
-    tool_calls: list[AssistantMessageToolCall] | None
-
-
-class AgentSystemMessage(TypedDict):
-    role: Literal["system"] = "system"
-    content: str
-
-
-class AgentToolCallMessage(TypedDict):
-    role: Literal["tool"] = "tool"
-    tool_call_id: str
-    content: str
-
-
-RunnerMessage = AgentUserMessage | AgentAssistantMessage | AgentToolCallMessage
-AgentMessage = RunnerMessage | AgentSystemMessage
-RunnerMessages = list[RunnerMessage]
-
-
-class LiteLLMRawChunk(TypedDict):
-    """
-    Define the type of chunk
-    """
-
-    type: Literal["litellm_raw"]
-    raw: litellm.ModelResponseStream
-
-
-class UsageChunk(TypedDict):
-    """
-    Define the type of usage info chunk
-    """
-
-    type: Literal["usage"]
-    usage: litellm.Usage
-
-
-class FinalMessageChunk(TypedDict):
-    """
-    Define the type of final message chunk
-    """
-
-    type: Literal["final_message"]
-    message: AssistantMessage
-    finish_reason: Literal["stop", "tool_calls"]
-
-
-class ToolCallChunk(TypedDict):
-    """
-    Define the type of tool call chunk
-    """
-
-    type: Literal["tool_call"]
-    name: str
-    arguments: str
-
-
-class ToolCallResultChunk(TypedDict):
-    """
-    Define the type of tool call result chunk
-    """
-
-    type: Literal["tool_call_result"]
-    tool_call_id: str
-    name: str
-    content: str
-
-
-class ContentDeltaChunk(TypedDict):
-    """
-    Define the type of message chunk
-    """
-
-    type: Literal["content_delta"]
-    delta: str
-
-
-class ToolCallDeltaChunk(TypedDict):
-    """
-    Define the type of tool call delta chunk
-    """
-
-    type: Literal["tool_call_delta"]
-    tool_call_id: str
-    name: str
-    arguments_delta: str
-
-
-AgentChunk = LiteLLMRawChunk | UsageChunk | FinalMessageChunk | ToolCallChunk | ToolCallResultChunk | ContentDeltaChunk | ToolCallDeltaChunk

lite_agent-0.1.0.dist-info/METADATA

@@ -1,22 +0,0 @@
-Metadata-Version: 2.4
-Name: lite-agent
-Version: 0.1.0
-Summary: A lightweight, extensible framework for building AI agent.
-Author-email: Jianqi Pan <jannchie@gmail.com>
-License: MIT
-Keywords: AI,agent framework,assistant,chatbot,function call,openai,pydantic,rich,tooling
-Classifier: Intended Audience :: Developers
-Classifier: Intended Audience :: Science/Research
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Operating System :: OS Independent
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
-Classifier: Topic :: Communications :: Chat
-Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
-Classifier: Topic :: Software Development :: Libraries :: Python Modules
-Requires-Python: >=3.10
-Requires-Dist: funcall>=0.6.0
-Requires-Dist: prompt-toolkit>=3.0.51
-Requires-Dist: rich>=14.0.0

lite_agent-0.1.0.dist-info/RECORD

@@ -1,13 +0,0 @@
-lite_agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-lite_agent/__main__.py,sha256=D4Yg5a6hDYMq6aCEGg3MjILF4pVWD6mBKjri-qs7Rzk,3766
-lite_agent/agent.py,sha256=LmC5aqMVWsq5AkE18Od0FEVUdiZN1GTs5OlO6mrvxUM,1187
-lite_agent/chunk_handler.py,sha256=MDk9rW_axZ412VOPnITbQGVr-mCGAe2jy3khUXkYbvM,5135
-lite_agent/loggers.py,sha256=0-yb_1Ec8TodfTI_nZWeCJSpSZZM9R53UUOE5--L2iw,57
-lite_agent/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-lite_agent/runner.py,sha256=TaBm6pCFwdJtygNnmVZDlFbkK3-Z1s52YD9XviwrA6o,2257
-lite_agent/types.py,sha256=NUpoSDaaiWWYezWluDdWBb-DsYFvZFoNd4tLWqBzCZ4,3058
-lite_agent/processors/__init__.py,sha256=AIxxS9uLcNws1Erf-b8z8zZKaKA5NQ7ooZ7HaCIJTq4,115
-lite_agent/processors/stream_chunk_processor.py,sha256=WlVcfc4kdiCz0Fc1Ybe5GapN2stK0C_0DIymbiGpZfI,3420
-lite_agent-0.1.0.dist-info/METADATA,sha256=oFRlVtYwsr1P9RPOP2JNu309-Jpqc95f-TYrExmrFBc,972
-lite_agent-0.1.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-lite_agent-0.1.0.dist-info/RECORD,,