PublicMirrors/you-get
1
0
Fork 0
mirror of https://github.com/soimort/you-get.git synced 2025-01-23 21:45:02 +03:00
Code Issues Actions Packages Projects Releases 1 Wiki Activity

[instagram] fix extraction

Browse Source
This commit is contained in:
Mort Yao 2022-07-01 22:21:47 +02:00
parent a47960f6ed
commit d661c95480
No known key found for this signature in database
GPG Key ID: 07DA00CB78203251
1 changed files with 32 additions and 42 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
src/you_get/extractors/instagram.py
View File

@ -10,28 +10,18 @@ def instagram_download(url, output_dir='.', merge=True, info_only=False, **kwarg
vid = r1(r'instagram.com/\w+/([^/]+)', url) vid = r1(r'instagram.com/\w+/([^/]+)', url)
description = r1(r'<meta property="og:title" content="([^"]*)"', cont) or \ description = r1(r'<meta property="og:title" content="([^"]*)"', cont) or \
r1(r'<title>\s([^<]*)</title>', cont) # with logged-in cookies r1(r'<title>([^<]*)</title>', cont) # with logged-in cookies
title = "{} [{}]".format(description.replace("\n", " "), vid) title = "{} [{}]".format(description.replace("\n", " "), vid)
stream = r1(r'<meta property="og:video" content="([^"]*)"', cont) appId = r1(r'"appId":"(\d+)"', cont)
if stream: media_id = r1(r'"media_id":"(\d+)"', cont)
_, ext, size = url_info(stream)
print_info(site_info, title, ext, size) api_url = 'https://i.instagram.com/api/v1/media/%s/info/' % media_id
if not info_only:
download_urls([stream], title, ext, size, output_dir, merge=merge)
else:
data = re.search(r'window\._sharedData\s*=\s*(.*);</script>', cont)
try: try:
info = json.loads(data.group(1)) api_cont = get_content(api_url, headers={**fake_headers, **{'x-ig-app-id': appId}})
post = info['entry_data']['PostPage'][0]
assert post['items']
except: except:
# with logged-in cookies log.wtf('[Error] Please specify a cookie file.')
data = re.search(r'window\.__additionalDataLoaded\(\'[^\']+\',(.*)\);</script>', cont) post = json.loads(api_cont)
if data is not None:
log.e('[Warning] Cookies needed.')
post = json.loads(data.group(1))
for item in post['items']: for item in post['items']:
code = item['code'] code = item['code']