import os
import asyncio
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage
from langchain.tools import tool
from deepagents import create_deep_agent
from deepagents.backends import FilesystemBackend, LocalShellBackend, CompositeBackend

# LLM - OpenRouter
llm = ChatOpenAI(
    model="openai/gpt-oss-20b:free",
    base_url="https://openrouter.ai/api/v1",
    api_key=os.getenv("OPENAI_API_KEY"),
    temperature=0.0,
)

# Simple backend - filesystem + local shell
backend = CompositeBackend(
    [
        LocalShellBackend(workspace_dir="./workspace"),
        FilesystemBackend(),
    ]
)

# Example tool - echo (replace with real logic if needed)
@tool
def echo(query: str) -> str:
    """Return the received query unchanged."""
    return query

# Create the deep agent
agent = create_deep_agent(
    model=llm,
    tools=[echo],
    backend=backend,
    system_prompt="You are a helpful assistant that streams its answer token by token.",
)

def format_message(message) -> str:
    """Convert a LangChain message to a printable string."""
    if getattr(message, "content", None):
        return message.content
    # If the message is a tool call, show the call
    if getattr(message, "tool_calls", None):
        tc = message.tool_calls[0]
        return f"{tc['name']}({tc['args']})"
    return ""

def format_chunk_message(chunk):
    """Print a token chunk, adding a separator when the step changes."""
    message, meta = chunk
    global current_step
    step = meta.get("langgraph_step", 0)
    if step != current_step:
        current_step = step
        print("\n--- --- ---\n")
    if getattr(message, "content", None):
        print(message.content, end="", flush=True)

async def main():
    # Prepare the input
    user_input = "Расскажи, как приготовить борщ, используя инструмент echo для демонстрации."
    stream = agent.stream(
        {"messages": [HumanMessage(content=user_input)]},
        stream_mode=["messages", "updates"],
    )

    global current_step
    current_step = 0

    # Iterate over the stream
    for chunk_type, chunk_data in stream:
        if chunk_type == "messages":
            format_chunk_message(chunk_data)
        elif chunk_type == "updates":
            # When a model update arrives, print the last model message (tool call or final answer)
            model_info = chunk_data.get("model")
            if model_info and "messages" in model_info:
                last_msg = model_info["messages"][-1]
                print("\n" + format_message(last_msg))
    # Ensure the final newline
    print()

if __name__ == "__main__":
    asyncio.run(main())