app.py

import gradio as gr
import yfinance as yf
from pypfopt.discrete_allocation import DiscreteAllocation, get_latest_prices
from pypfopt import EfficientFrontier
from pypfopt import risk_models
from pypfopt import expected_returns
from pypfopt import plotting
import copy
import numpy as np
import pandas as pd
import plotly.express as px
import matplotlib.pyplot as plt
from datetime import datetime

def plot_cum_returns(data, title, initial_capital=1000):
    # 일일 누적 수익률 계산 및 시각화
    daily_cum_returns = (1 + data.dropna().pct_change()).cumprod() * initial_capital
    fig = px.line(daily_cum_returns, title=title)
    return fig

def plot_efficient_frontier_and_max_sharpe(mu, S):
    # 최대 샤프 비율로 포트폴리오 최적화 및 효율적 투자선 그리기
    ef = EfficientFrontier(mu, S)
    fig, ax = plt.subplots(figsize=(6, 4))
    ef_max_sharpe = copy.deepcopy(ef)
    plotting.plot_efficient_frontier(ef, ax=ax, show_assets=False)
    ef_max_sharpe.max_sharpe(risk_free_rate=0.02)
    ret_tangent, std_tangent, _ = ef_max_sharpe.portfolio_performance()
    ax.scatter(std_tangent, ret_tangent, marker="*", s=100, c="r", label="MAX Sharpe")
    ax.legend()
    return fig

def plot_weights(weights):
    # 포트폴리오 최적 투자 비율 그래프 출력
    labels = weights.keys()
    sizes = weights.values()
    fig, ax = plt.subplots()
    ax.pie(sizes, labels=labels, autopct='%1.1f%%')
    ax.axis('equal')
    return fig

def output_results(start_date, end_date, tickers_string):
    tickers = tickers_string.split(',')
    stocks_df = yf.download(tickers, start=start_date, end=end_date)['Adj Close']
    fig_indiv_prices = px.line(stocks_df, title='개별 주식 가격')
    fig_cum_returns = plot_cum_returns(stocks_df, '개별 주식의 누적 수익률 ($1,000 시작)')
    corr_df = stocks_df.corr().round(2)
    fig_corr = px.imshow(corr_df, text_auto=True, title='주식 간 상관 관계')
    mu = expected_returns.mean_historical_return(stocks_df)
    S = risk_models.sample_cov(stocks_df)
    ef = EfficientFrontier(mu, S)
    weights = ef.max_sharpe(risk_free_rate=0.02)
    cleaned_weights = ef.clean_weights()
    fig_weights = plot_weights(cleaned_weights)
    expected_annual_return, annual_volatility, sharpe_ratio = ef.portfolio_performance()
    fig_efficient_frontier = plot_efficient_frontier_and_max_sharpe(mu, S)
    
    return fig_cum_returns, fig_efficient_frontier, fig_corr, fig_indiv_prices, fig_weights, \
           f"{expected_annual_return*100:.2f}%", f"{annual_volatility*100:.2f}%", f"{sharpe_ratio:.2f}"

css = """
footer { 
    visibility: hidden; 
}
"""

# Gradio 인터페이스 구성
with gr.Blocks(css=css) as app:
    gr.Markdown("<h1>글로벌 자산(주식,지수,BTC,상품 등) 포트폴리오 최적화 도구</h1>")
    with gr.Row():
        start_date = gr.Textbox("2013-01-01", label="시작 일자")
        end_date = gr.Textbox(datetime.now().date(), label="종료 일자")
        tickers_string = gr.Textbox("NVDA,^GSPC,GC=F,MSFT,BTC-USD", label="주식 티커를 쉼표로 구분하여 입력하세요")
        btn = gr.Button("포트폴리오 최적화 결과 보기")
    
    with gr.Row():
        expected_annual_return = gr.Text(label="예상 연간 수익률")
        annual_volatility = gr.Text(label="연간 변동성")
        sharpe_ratio = gr.Text(label="샤프 비율")
    
    with gr.Column():
        fig_cum_returns = gr.Plot(label="최적화된 포트폴리오의 누적 수익률 (시작 가격 $1,000)")
        fig_efficient_frontier = gr.Plot(label="효율적 투자선")
        fig_corr = gr.Plot(label="주식 간 상관 관계")
        fig_indiv_prices = gr.Plot(label="개별 주식 가격")
        fig_weights = gr.Plot(label="포트폴리오 최적 투자 비율")
    
    btn.click(fn=output_results, inputs=[start_date, end_date, tickers_string],
              outputs=[fig_cum_returns, fig_efficient_frontier, fig_corr, fig_indiv_prices, fig_weights, expected_annual_return, annual_volatility, sharpe_ratio])

app.launch()