PublicMirrors/you-get
1
0
Fork 0
mirror of https://github.com/soimort/you-get.git synced 2025-02-02 16:24:00 +03:00
Code Issues Actions Packages Projects Releases 1 Wiki Activity

[twitter] fix extraction

Browse Source
This commit is contained in:
Mort Yao 2023-05-09 15:22:19 +02:00
parent 2aaa877a9b
commit ad5825a8f6
No known key found for this signature in database
GPG Key ID: 07DA00CB78203251
1 changed files with 12 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
src/you_get/extractors/twitter.py
View File

@ -34,7 +34,18 @@ def twitter_download(url, output_dir='.', merge=True, info_only=False, **kwargs)
**kwargs)
return
html = get_html(url, faker=True) # now it seems faker must be enabled
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:88.0) Gecko/20100101 Firefox/88.0'
}
host = 'www.twitter.com'
html, set_cookie = getHttps(host, url, headers=headers)
# "Found. Redirecting to..."
guest_id = r1('guest_id=([^;]+);', set_cookie)
headers['Cookie'] = 'guest_id=%s' % guest_id
html = get_content(url, headers=headers)
screen_name = r1(r'twitter\.com/([^/]+)', url) or r1(r'data-screen-name="([^"]*)"', html) or \
r1(r'<meta name="twitter:title" content="([^"]*)"', html)
item_id = r1(r'twitter\.com/[^/]+/status/(\d+)', url) or r1(r'data-item-id="([^"]*)"', html) or \