Automatically download all handouts from NCTU e3.

.gitignore

.idea/
secret

main.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

import os
import re
import shutil
from time import sleep
from urllib.parse import unquote

from selenium import webdriver

course_pg = 0
course_no = 0
p = re.compile(r'id="ctl00_ContentPlaceHolder1_dgCourseHandout_.+?_lnkFile"')

with open("secret") as file:
    usr = file.readline().strip()
    pwd = file.readline().strip()
    src_dir = file.readline().strip()
    dest_dir = file.readline().strip()

driver = webdriver.Chrome()
driver.implicitly_wait(10)
driver.get("https://dcpc.nctu.edu.tw")
print(driver.title)

driver.find_element_by_id("txtAccount").send_keys(usr)
driver.find_element_by_id("txtPwd").send_keys(pwd)
driver.find_element_by_id("btnLoginIn").click()
sleep(5)
print(driver.find_element_by_id("ctl00_lbWelcomeT").text)

while True:

    driver.find_element_by_link_text("歷年課程").click()
    sleep(5)
    for _ in range(course_pg):
        next_button = driver.find_element_by_id("ctl00_ContentPlaceHolder1_DataNavigator1_ctl03")
        assert "not-allowed" not in next_button.get_attribute("style")
        next_button.click()
        sleep(5)
    course_links = driver.find_elements_by_link_text("進入課程")
    if course_no == len(course_links):
        print("page", course_pg + 1, ", number", course_no)
        course_no = 0
        course_pg += 1
        next_button = driver.find_element_by_id("ctl00_ContentPlaceHolder1_DataNavigator1_ctl03")
        if "not-allowed" in next_button.get_attribute("style"):
            driver.quit()
            exit(0)
        else:
            next_button.click()
            sleep(5)
            course_links = driver.find_elements_by_link_text("進入課程")
    course_links[course_no].click()
    sleep(5)
    course_name = driver.find_element_by_id("ctl00_lbCurrentCourseName").text
    print(course_name)
    driver.find_element_by_link_text("教材列表").click()
    sleep(5)

    while True:
        html = driver.page_source
        iterator = p.finditer(html)
        for match in iterator:
            handout_id = match.group()
            driver.find_element_by_id(handout_id[4:-1]).click()
            sleep(5)
            driver.switch_to.window(driver.window_handles[-1])
            driver.switch_to.frame("frmMain")
            driver.find_element_by_id("ctl00_ContentPlaceHolder1_btnUrl").click()
            sleep(5)
            driver.close()
            driver.switch_to.window(driver.window_handles[0])
        next_button = driver.find_element_by_id("ctl00_ContentPlaceHolder1_DataNavigator1_ctl03")
        if "not-allowed" in next_button.get_attribute("style"):
            break
        else:
            next_button.click()
            sleep(5)

    driver.find_element_by_id("ctl00_btnBackIndex").click()
    sleep(5)
    course_no += 1

    os.makedirs(os.path.join(dest_dir, course_name), exist_ok=True)
    for filename in os.listdir(src_dir):
        shutil.move(
            os.path.join(src_dir, filename),
            os.path.join(dest_dir, course_name, unquote(filename))
        )