app.py

import streamlit as st
import numpy as np
from langchain.chains import LLMChain
from langchain.prompts import PromptTemplate
from pydantic.v1 import BaseModel, Field
from typing import Any, Optional, Dict, List
from huggingface_hub import InferenceClient
from langchain.llms.base import LLM
from markup import app_intro
import os

HF_token = os.getenv("apiToken")

model_name = "mistralai/Mistral-7B-Instruct-v0.1"
hf_token = HF_token
kwargs = {"max_new_tokens":10, "temperature":0.1, "top_p":0.95, "repetition_penalty":1.0, "do_sample":True}

class KwArgsModel(BaseModel):
    kwargs: Dict[str, Any] = Field(default_factory=dict)

class CustomInferenceClient(LLM, KwArgsModel):
    model_name: str
    inference_client: InferenceClient

    def __init__(self, model_name: str, hf_token: str, kwargs: Optional[Dict[str, Any]] = None):
        inference_client = InferenceClient(model=model_name, token=hf_token)
        super().__init__(
            model_name=model_name,
            hf_token=hf_token,
            kwargs=kwargs,
            inference_client=inference_client
        )

    def _call(
        self,
        prompt: str,
        stop: Optional[List[str]] = None
    ) -> str:
        if stop is not None:
            raise ValueError("stop kwargs are not permitted.")
        response_gen = self.inference_client.text_generation(prompt, **self.kwargs, stream=True, return_full_text=False)
        response = ''.join(response_gen)  
        return response

    @property
    def _llm_type(self) -> str:
        return "custom"

    @property
    def _identifying_params(self) -> dict:
        return {"model_name": self.model_name}

def check_winner(board):
    for row in board:
        if len(set(row)) == 1 and row[0] != "":
            return row[0]
    for col in board.T:
        if len(set(col)) == 1 and col[0] != "":
            return col[0]
    if len(set(board.diagonal())) == 1 and board[0, 0] != "":
        return board[0, 0]
    if len(set(np.fliplr(board).diagonal())) == 1 and board[0, 2] != "":
        return board[0, 2]
    return None

def check_draw(board):
    return not np.any(board == "")

def main():
    st.set_page_config(page_title="Tic Tac Toe", page_icon=":memo:", layout="wide")

    col1, col2 = st.columns([1, 2])
    with col1:
        st.image("image.jpg", use_column_width=True)
    with col2:
        st.markdown(app_intro(), unsafe_allow_html=True) 

    st.markdown("____") 

    scores = st.session_state.get("scores", {"X": 0, "O": 0})
    board = np.array(st.session_state.get("board", [["" for _ in range(3)] for _ in range(3)]))
    current_player = st.session_state.get("current_player", "X")
    winner = check_winner(board)

    if winner is not None:
        scores[winner] += 1
        st.write(f"Player {winner} wins! Score: X - {scores['X']} | O - {scores['O']}")
    elif check_draw(board):
        st.write("Draw!")
    else:
        for row in range(3):
            cols = st.columns(3)
            for col in range(3):
                button_key = f"button_{row}_{col}"
                if board[row, col] == "" and current_player == "X":
                    if cols[col].button(" ", key=button_key):
                        board[row, col] = current_player
                        st.session_state.board = board
                        st.session_state.current_player = "O"
                        progress = st.session_state.get("progress", [])
                        progress.append(f"{current_player}: {chr(65 + row)}{col + 1}")
                        st.session_state.progress = progress
                        st.experimental_rerun()
                else:
                    cols[col].write(board[row, col])

    if current_player == "O" and winner is None:
        with st.spinner("Calculating AI Move..."):
            ai_progress = ", ".join(st.session_state.progress)
            ai_move = get_ai_move(ai_progress)
            ai_row, ai_col = ai_move.split(": ")[1]
            ai_row = ord(ai_row[0]) - 65
            ai_col = int(ai_col) - 1
            board[ai_row, ai_col] = "O"
            st.session_state.board = board

            progress = st.session_state.get("progress", [])
            progress.append(f"O: {chr(65 + ai_row)}{ai_col + 1}")
            st.session_state.progress = progress

            st.session_state.current_player = "X"
            st.experimental_rerun()

    st.markdown("____")
    if st.button("Reset game"):
        st.session_state.board = [["" for _ in range(3)] for _ in range(3)]
        st.session_state.current_player = "X"
        st.session_state.progress = []       
        st.experimental_rerun()

    if st.button("Reset Scores"):
        scores["X"] = 0
        scores["O"] = 0
        st.session_state.scores = scores
        st.experimental_rerun()

    
    progress = st.session_state.get("progress", [])
    st.write(", ".join(progress))

    st.write(f"Score: X - {scores['X']} | O - {scores['O']}")

def get_ai_move(progress):
    print("progress", progress)
    llm = CustomInferenceClient(model_name=model_name, hf_token=hf_token, kwargs=kwargs)

    template = """<s>[INST] Decide the next O move in tic tac toe game[/INST]
    Example output format = O: B2
    
    {progress}
    
    Next move:"""

    prompt = PromptTemplate(template=template, input_variables=["progress"])
    llm_chain = LLMChain(prompt=prompt, llm=llm)

    answer = llm_chain.run(progress)
    answer = answer.replace("</s>", "")
    
    print("ai move", answer)
    return answer


if __name__ == "__main__":
    main()