import inspect
from tqdm import tqdm
path_hf=inspect.getfile(HuggingFacePipeline)
from subprocess import Popen, PIPE as P
from langchain_experimental.tools.python.tool import PythonREPLTool as PYT
from langchain.agents import load_tools, create_structured_chat_agent as Agent,AgentExecutor as Ex, AgentType as Type
from langchain.agents.agent_toolkits import create_retriever_tool as crt
from langchain_community.agent_toolkits import FileManagementToolkit as FMT
from langchain.tools import Tool,YouTubeSearchTool as YTS
from langchain.memory import ConversationBufferMemory as MEM,RedisChatMessageHistory as HIS
from langchain.schema import SystemMessage as SM,HumanMessage as HM, AIMessage as AM
from langchain import hub
import os
import torch
from langchain import hub
torch.set_flush_denormal(True)
import importlib.util
import logging
from typing import Any, Dict, Iterator, List, Mapping, Optional
with open(path_hf,"r") as f:
    s=f.read()
with open(path_hf,"w") as f:
    f.write(s.replace("        model = model_cls.from_pretrained(model_id, **_model_kwargs)","        model = torch.compile(model_cls.from_pretrained(model_id, **_model_kwargs),mode='max-autotune')"))
from langchain_core.prompts.chat import ChatPromptTemplate, MessagesPlaceholder
#system = '''Respond to the human as helpfully and accurately as possible. You have access to the following tools:

#{tools}

#Use a json blob to specify a tool by providing an action key (tool name) and an action_input key (tool input).

#Valid "action" values: "Final Answer" or {tool_names}

#Provide only ONE action per $JSON_BLOB, as shown:

#```
#{{
  #"action": $TOOL_NAME,
 # "action_input": $INPUT
#}}
#```
#'''
#Follow this format:

#Question: input question to answer
#Thought: consider previous and subsequent steps
#Action:
#```
#$JSON_BLOB
#```
#Observation: action result
#... (repeat Thought/Action/Observation N times)
#Thought: I know what to respond
#Action:
#```
#{
  #"action": "Final Answer",
  #"action_input": "Final response to human"
#}}

#Begin! Reminder to ALWAYS respond with a valid json blob of a single action. Use tools if necessary. Respond directly if appropriate. Format is Action:```$JSON_BLOB```then Observation'''
#'''


#(reminder to respond in a JSON blob no matter what)'''

prompt=hub.pull("hwchase17/structured-chat-agent")
from typing import Any, Dict, List, Optional
from langchain_core.language_models import BaseChatModel
from langchain_core.messages import AIMessage, BaseMessage, HumanMessage
from langchain_core.outputs import ChatResult, ChatGeneration
from langchain_core.callbacks.manager import CallbackManagerForLLMRun
from langchain_core.callbacks.manager import AsyncCallbackManagerForLLMRun
from langchain_core.runnables import run_in_executor
from transformers import AutoTokenizer, AutoModelForCausalLM
import requests
import os
if os.path.exists("./llama-3-open-ko-8b-instruct-preview-q5_k_m.gguf"):
    pass
else:
    req=requests.get("https://huggingface.co/peterpeter8585/Llama-3-Open-Ko-8B-Instruct-preview-Q5_K_M-GGUF/resolve/main/llama-3-open-ko-8b-instruct-preview-q5_k_m.gguf",stream=True)
    with open("./llama-3-open-ko-8b-instruct-preview-q5_k_m.gguf","wb") as f:
        for i in tqdm(req.iter_content(100000000000000000000)):
            f.write(i)

#from transformers import pipeline,AutoModelForCausalLM as M,AutoTokenizer as T
#m=M.from_pretrained("peterpeter8585/syai4.3")
#t=T.from_pretrained("peterpeter8585/syai4.3")
#pipe=pipeline(model=m,tokenizer=t,task="text-generation")
import multiprocessing

from langchain_community.chat_models import ChatLlamaCpp

llm = ChatLlamaCpp(
    temperature=0,
    model_path="./llama-3-open-ko-8b-instruct-preview-q5_k_m.gguf",
    n_ctx=10000,
    n_batch=300,  # Should be between 1 and n_ctx, consider the amount of VRAM in your GPU.
    n_threads=multiprocessing.cpu_count() - 1,
    repeat_penalty=1.5,
    top_p=0.5,
)
from langchain.retrievers import WikipediaRetriever as Wiki
import gradio as gr
chatbot = gr.Chatbot(
    label="SYAI4.1",
    show_copy_button=True,
    layout="panel"
)
def terminal(c):
    a=Popen(c,shell=True,stdin=P,stdout=P,stderr=P)
    return a.stdout.read()+a.stderr.read()
tools=FMT().get_tools()
tools.append(PYT())
tools.append(YTS())
tools.extend(load_tools(["requests_all"],allow_dangerous_tools=True))
tools.extend(load_tools(["llm-math","ddg-search"],llm=llm))
tools.append(Tool.from_function(func=terminal,name="terminal",description="터미널 명령어실행에 적합함"))
memory=MEM()
tools.append(crt(name="wiki",description="위키 백과를 검색하여 정보를 가져온다",retriever=Wiki(lang="ko",top_k_results=1)))
agent=Ex(agent=Agent(llm,tools,prompt),tools=tools,verbose=True,handle_parsing_errors=True,memory=memory)
def chat(message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p, chat_session):
    return agent.invoke({"input":message})
ai1=gr.ChatInterface(
    chat,
    chatbot=chatbot,
    additional_inputs=[
        gr.Textbox(value="You are a helpful assistant.", label="System message",  interactive=True),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.1, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.1,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
        gr.Textbox(label="chat_id(please enter the chat id!)")
    ],
    
)
ai1.launch()