from langgraph.prebuilt import create_react_agent
from langchain_ollama import ChatOllama
from typing import Annotated
from langgraph.prebuilt import ToolNode, tools_condition

from typing_extensions import TypedDict

from langchain_core.messages import HumanMessage, SystemMessage
from langgraph.graph import StateGraph, START, END
from langgraph.graph.message import add_messages
import logging
import prompt_toolkit
import prompt_toolkit.history
import prompt_toolkit.auto_suggest

logger = logging.getLogger(__name__)

from . import tools

cli_history = prompt_toolkit.history.FileHistory('output/cli_history.txt')

#MODEL = "gemma3:27b"
#MODEL = "qwen3:latest"
MODEL = 'hf.co/unsloth/Qwen3-30B-A3B-GGUF:Q4_K_M'

def create_model():
    available_tools = tools.get_tools()
    logger.info("Available tools:")
    for tool in available_tools:
        logger.info("- %s", tool.name)

    llm = ChatOllama(model=MODEL)
    llm.bind_tools(tools=available_tools )
    return create_react_agent(llm, tools=available_tools )

SYSTEM_MESSAGE = '''
You are a useful assistant with access to built in system tools.
Format responses as markdown.
Provide links when available.
'''

def main():
    logging.basicConfig(level='INFO')
    messages = [SystemMessage(SYSTEM_MESSAGE)]
    llm = create_model()
    prev_idx = 0
    while True:
        user_input  = prompt_toolkit.prompt("Human: ",
                                            history=cli_history,
                                            auto_suggest=prompt_toolkit.auto_suggest.AutoSuggestFromHistory(),
                                            )
        messages.append(HumanMessage(user_input))

        result = llm.invoke({
            'messages': messages,
        })
        messages = result['messages']
        for msg in messages[prev_idx:]:
            print(msg.pretty_repr())
            del msg
        prev_idx = len(messages)


if __name__ == '__main__':
    main()