Python-сообщество

 import requests
from bs4 import BeautifulSoup
import pandas as pd
# url = 'https://www.tennisabstract.com/cgi-bin/wplayer.cgi?p=EliseMertens#career-splits-h'
#
headers = {
    "user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/108.0.0.0 YaBrowser/23.1.1.1138 Yowser/2.5 Safari/537.36"
}
#
# response1 = requests.get(url, headers=headers)
# soup1 = BeautifulSoup(response1.content, 'html.parser')

https://www.minorleaguesplits.com/tennisabstract/cgi-bin/frags/EliseMertens.js

  
>>> import urllib.request
>>> 
>>> url = 'https://www.minorleaguesplits.com/tennisabstract/cgi-bin/frags/EliseMertens.js'
>>> 
>>> headers = {
...     'User-Agent': 'Firefox'
... }
>>> 
>>> req = urllib.request.Request(url=url, headers=headers)
>>> 
>>> with urllib.request.urlopen(req) as data:
...     text = data.read().decode('utf-8')
... 
>>> text[:100]
'var player_frag = `<p><h1 id="recent-results-h">Recent Results&nbsp;&nbsp;&nbsp;<span style="font-si'
>>>

  
>>> import re
>>> 
>>> text = """var player_frag = `<p><h1 id="recent-results-h"> ... </h1>"""
>>> 
>>> textnew1 = re.sub(r'^var[^`]*`', r'', text)
>>> textnew2 = re.sub(r'^<([^>]+)>.*$', r'\g<0></\1>', textnew1)
>>> 
>>> out = textnew2
>>> 
>>> text
'var player_frag = `<p><h1 id="recent-results-h"> ... </h1>'
>>> textnew1
'<p><h1 id="recent-results-h"> ... </h1>'
>>> textnew2
'<p><h1 id="recent-results-h"> ... </h1></p>'
>>> 
>>> out
'<p><h1 id="recent-results-h"> ... </h1></p>'
>>>