yahoo financeから株価データを入手してチャートを描画する

import os
import pandas as pd
import sys
from yahoo_finance_api2 import share
from yahoo_finance_api2.exceptions import YahooFinanceError

def get_data(name, year_num):
    my_share = share.Share(name)
    symbol_data = None
    
    try:
        symbol_data = my_share.get_historical(
            share.PERIOD_TYPE_YEAR,
            year_num,
            share.FREQUENCY_TYPE_DAY,
            1)
    except YahooFinanceError as e:
        print(e.message)
        sys.exit(1)

    # DataFrameに変換
    df_symbol_data = pd.DataFrame(symbol_data)
 
    # UNIX時間をUTC時間に変換
    df_symbol_data.timestamp = pd.to_datetime(df_symbol_data.timestamp, unit='ms').dt.strftime('%Y-%m-%d')
    df_symbol_data.timestamp = pd.to_datetime(df_symbol_data.timestamp)
    df_symbol_data = df_symbol_data.set_index('timestamp')

    return df_symbol_data

df_d = get_data(name='%5EDJI', year_num=1)
df_d

df_n = get_data(name='%5EN225', year_num=1)
df_n

import mplfinance as mpf

# 30日分を描画
mpf.plot(df_n.iloc[0:30,:],type='candle',style='charles')

# 移動平均線も描画
mpf.plot(df_n.iloc[0:30,:],type='candle',mav=(2,4,6),style='charles')

import matplotlib.pyplot as plt

fig, axes = plt.subplots(2, 1, figsize=(8,8))

# 日経平均株価を30日分描画
mpf.plot(df_n.iloc[0:30,:],ax=axes[0],type='candle',style='charles')
# ダウ平均株価を30日分描画
mpf.plot(df_d.iloc[0:30,:],ax=axes[1],type='candle',style='charles')

fig, axes = plt.subplots(2, 1, figsize=(5.12,5.12))

mpf.plot(df_n.iloc[0:30,:],ax=axes[0],type='candle',style='charles')
mpf.plot(df_d.iloc[0:30,:],ax=axes[1],type='candle',style='charles')

axes[0].axis("off")
axes[1].axis("off")

os.makedirs('./img', exist_ok=True)

def plot_chart(i, df_n, df_d):
    fig, axes = plt.subplots(2, 1, figsize=(5.12,5.12))
    mpf.plot(df_n.iloc[i-30:i,:],ax=axes[0],type='candle',style='charles')
    mpf.plot(df_d.iloc[i-30:i,:],ax=axes[1],type='candle',style='charles')

    axes[0].axis("off")
    axes[1].axis("off")

    name = str(df_n.index[i]).replace(' 00:00:00','')

    fig.savefig(f'./img/{name}.png')

# ダミー変数を作成
df_n['dummy_n']=0
df_d['dummy_d']=0

# お互いの欠損日を把握するため、結合する
df_n = pd.concat([df_n,df_d['dummy_d']], axis=1).sort_index()
df_d = pd.concat([df_d,df_n['dummy_n']], axis=1).sort_index()

# 始値と終値の差を計算
df_n['delta']=df_n['close'] - df_n['open']

# ラベルを作成
df_n['label']=0
df_n.loc[df_n['delta']>100,'label']=1
df_n.loc[df_n['delta']<-100,'label']=2

# データセット化
import datasets
import json

data_li = []

for i in range(len(df_n))[30:]:
    if df_n.iloc[i].isnull().any():
        continue

    plot_chart(i,df_n,df_d)

    name = str(df_n.index[i]).replace(' 00:00:00','')
    delta = df_n.iloc[i]['delta']
    label = df_n.iloc[i]['label']

    tmp_dic = {
        'name':f'./img/{name}.png',
        'delta':delta,
        'label':label,
    }

    data_li.append(tmp_dic)

dic_json = {
    'version': '0.1.0',
    'data':data_li,
}

with open('dataset.json', 'w') as f:
    json.dump(dic_json, f, indent=4)

from datasets import load_dataset
dataset = load_dataset("json", data_files="dataset.json")