気象庁から名古屋と福岡の日時の気圧と気温をスクレイピングするスクリプト

scrape.py

from bs4 import BeautifulSoup
from pathlib import Path
import requests
import numpy as np
import pandas as pd
import time


class JMADataDaily:
    def __init__(self, prec_no=51, block_no=47636, year=2025, month=12, suffix='_nagoya'):
        cache_path = Path((
            '~/cache/jma-daily_{}_{}_{}-{}.html'
        ).format(prec_no, block_no, year, str(month).zfill(2))).expanduser()
        if not cache_path.is_file():
            print('Cache not found. Attempting to download...')
            time.sleep(1)
            url = (
                'https://www.data.jma.go.jp/stats/etrn/view/daily_s1.php'
                '?prec_no={}&block_no={}&year={}&month={}&day=&view='
            ).format(prec_no, block_no, year, month)
            html = requests.get(url).text
            cache_path.write_text(html, newline='\n', encoding='utf8')
            print(f'Saved to {cache_path}.')
        else:
            html = cache_path.read_text(encoding='utf8')
        date_prefix = str(year) + '-' + str(month).zfill(2) + '-'
        soup = BeautifulSoup(html, 'html.parser')
        table = soup.find('table', id='tablefix1')
        rows = table.find_all('tr')
        data = []
        for row in rows:
            cols = row.find_all('td')
            if len(cols) == 0:
                continue
            data.append([
                date_prefix + cols[0].string.zfill(2),
                cols[1].string.strip().split()[0],
                cols[2].string.strip().split()[0],
                cols[6].string.strip().split()[0],
                cols[7].string.strip().split()[0],
                cols[8].string.strip().split()[0],
            ])
        self.df = pd.DataFrame(np.array(data), columns=[
            'date',
            'atmo_local' + suffix,
            'atmo_sea' + suffix,
            'temp_avg' + suffix,
            'temp_high' + suffix,
            'temp_low' + suffix,
        ])

    @classmethod
    def get_months(cls, prec_no=51, block_no=47636, year=2025, months=[10, 11, 12], suffix='_nagoya'):
        df = cls(prec_no, block_no, year, month=months[0], suffix=suffix).df
        for month in months[1:]:
            df = pd.concat([df, cls(prec_no, block_no, year, month=month, suffix=suffix).df])
        return df

if __name__ == '__main__':
    df = JMADataDaily.get_months(
        prec_no=51, block_no=47636, year=2025,
        months=list(range(1, 13)), suffix='_nagoya',
    )
    df_ = JMADataDaily.get_months(
        prec_no=82, block_no=47807, year=2025,
        months=list(range(1, 13)), suffix='_fukuoka',
    )
    df = pd.concat([df, df_.iloc[:, 1:]], axis=1)
    df.to_csv('workspace/nazuna/tests/data/jma-daily_2025.csv', index=False, lineterminator='\n')