1. from bs4 import BeautifulSoup
    2. import requests
    4. allUniv = []
    7. def getHTMLText(url):
    8.     try:
    9.         r = requests.get(url, timeout=30)
    10.         r.raise_for_status()
    11.         r.encoding = 'utf-8'
    12.         return r.text
    13.     except:
    14.         return ""
    17. def fillUnivList(soup):
    18.     data = soup.find_all('tr')
    19.     print(data)
    20.     for tr in data:
    21.         ltd = tr.find_all('td')
    22.         print(ltd)
    23.         if len(ltd) == 0:
    24.             continue
    25.         singleUniv = []
    26.         for td in ltd:
    27.             print(td.get_text())
    28.             if td.get_text() != '':
    29.                 singleUniv.append(td.get_text().strip())
    30.         print(singleUniv)
    31.         allUniv.append(singleUniv)
    32.         print(allUniv)
    36. def printUnivList(num):
    37.     print("{:^4}{:^10}{:^5}{:^8}{:^10}".format("排名", "学校名称", "省市", "总分", "培养规模"))
    38.     for i in range(num):
    39.         u = allUniv[i]
    40.         print(u)
    41.         print("{:^4}{:^10}{:^5}{:^8}{:^10}".format(u[0], u[1], u[2], u[4], u[5]))
    42.         print("{:^4}{:^10}{:^5}{:^8}{:^10}".format(u[0], u[1], u[2], u[3], u[6]))
    45. def main():
    46.     url = 'https://www.shanghairanking.cn/api/pub/v1/bcur?bcur_type=11&year=2020'
    47.     html = getHTMLText(url)
    48.     soup = BeautifulSoup(html, "html.parser")
    49.     print(soup)
    50.     # fillUnivList(soup)
    51.     # printUnivList(5)
    54. main()
    1. import requests
    2. from bs4 import BeautifulSoup
    3. import bs4
    5. def getJson(url):
    6.     try:
    7.         r = requests.get(url, timeout=30)
    8.         r.raise_for_status()
    9.         r.encoding = r.apparent_encoding
    10.         return r.json()
    11.     except  Exception as e:
    12.         print(e)
    13.         return {'code':'0','msg':'error'}
    15. def printUnivList(rankings, limit):
    16.     print("{:<4}\t{:^10}\t{:^10}\t{:^8}".format("排名","学校名称","省市","总分"))
    17.     for i in range(limit):
    18.         u=rankings[i]
    19.         print("{:^4}\t{:^10}\t{:^10}\t{:^8}".format(u["ranking"],u["univNameCn"],u["province"],u["score"]))
    21. def main():
    22.     uinfo = []
    23.     url = 'https://www.shanghairanking.cn/api/pub/v1/bcur?bcur_type=11&year=2020'
    24.     json = getJson(url)
    26.     if json["code"]!=200:
    27.         print(json["msg"])
    28.         return
    30.     rankings=json["data"]["rankings"]
    31.     printUnivList(rankings, 20) # 20 univs
    33. if __name__=="__main__":
    34.     main()