PythonでWebスクレイピング ~REDMINEからの情報抽出とエクセル出力~

import time
from selenium import webdriver
from webdriver_manager.chrome import ChromeDriverManager
from selenium.webdriver.common.by import By
import pandas as pd
import datetime
import pprint

chrome_options = webdriver.ChromeOptions()
chrome_options.add_experimental_option("excludeSwitches", ["enable-automation",'enable-logging'])


# スクレイピングメイン処理
def main():
    driver = webdriver.Chrome(ChromeDriverManager().install(), options=chrome_options)
    driver.get("http://127.0.0.1/redmine/projects/scraping/issues/gantt")
    # ログイン用にwait
    print("ログインを完了し、ガントチャートを開いてください")
    for try_cnt in range(20):
        print(try_cnt)
        time.sleep(1)
    # チケット情報取得
    get_redmine_inf(driver)
    driver.quit()
    return


# チケット情報取得
def get_redmine_inf(driver):
    ticket_url_list = []
    ticket_info_dec = {"ID": "","チケット名": "","ステータス": "","優先度": "","担当者": "","開始日": "","期日": "","進捗率": "",}
    excel = make_excel(ticket_info_dec)
    # ガントチャートから各チケットのURLを取得
    gnt_all_ticket = driver.find_elements(By.CSS_SELECTOR,'.issue-subject.hascontextmenu')
    for gnt_ticket in gnt_all_ticket:
        ticket_url = gnt_ticket.find_element(By.TAG_NAME,'a').get_attribute('href')
        ticket_url_list.append(ticket_url)
    # チケットの情報取得
    for ticket_url in ticket_url_list:
        driver.get(ticket_url)
        time.sleep(1)
        # ID
        ticket_info_dec["ID"] = ticket_url.split("/")[-1]
        # チケット名
        ticket_info_dec["チケット名"] = driver.find_element(By.CSS_SELECTOR,'.subject').find_element(By.TAG_NAME,'h3').text
        # ステータス
        ticket_info_dec["ステータス"] = driver.find_element(By.CSS_SELECTOR,'.status.attribute').find_element(By.CSS_SELECTOR,'.value').text
        # 優先度
        ticket_info_dec["優先度"] = driver.find_element(By.CSS_SELECTOR,'.priority.attribute').find_element(By.CSS_SELECTOR,'.value').text
        # 担当者
        ticket_info_dec["担当者"] = driver.find_element(By.CSS_SELECTOR,'.assigned-to.attribute').find_element(By.CSS_SELECTOR,'.value').text
        # 開始日
        ticket_info_dec["開始日"] = driver.find_element(By.CSS_SELECTOR,'.start-date.attribute').find_element(By.CSS_SELECTOR,'.value').text
        # 期日
        ticket_info_dec["期日"] = driver.find_element(By.CSS_SELECTOR,'.due-date.attribute').find_element(By.CSS_SELECTOR,'.value').text
        # 進捗率
        ticket_info_dec["進捗率"] = driver.find_element(By.CSS_SELECTOR,'.progress.attribute').find_element(By.CSS_SELECTOR,'.value').text
        pprint.pprint(ticket_info_dec)
        excel.add_inf(ticket_info_dec)
    # エクセルで保存
    now = datetime.datetime.now()
    file_name = '{}.xlsx'.format(now.strftime('%Y%m%d_%H%M%S'))
    excel.save_file(file_name,"抽出結果")


class make_excel:
    # エクセルヘッダ記入
    def __init__( self , init_dic) :
        key_list = init_dic.keys()
        self.df = pd.DataFrame(columns=key_list)
    # 行追加
    def add_inf ( self , add_dict ) :
        self.df = self.df.append( add_dict , ignore_index=True)
        return
    # ファイル保存
    def save_file( self , file_name , title ):
        self.df.to_excel(file_name, sheet_name=title)
        return


if __name__ == "__main__":
    main()

    # ログイン用にwait
    print("ログインを完了し、ガントチャートを開いてください")
    for try_cnt in range(20):
        print(try_cnt)
        time.sleep(1)