w/youtube-dl
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

[instagram:user] Fix extraction (closes #16119)

Browse source
This commit is contained in:
Sergey M․ 2018-04-17 22:37:50 +07:00
parent 522d6b5c96
commit 238d42cf5d
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D
1 changed files with 38 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

49
youtube_dl/extractor/instagram.py
View file

@ -6,11 +6,16 @@ import json
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import compat_str from ..compat import (
compat_str,
compat_HTTPError,
)
from ..utils import ( from ..utils import (
ExtractorError,
get_element_by_attribute, get_element_by_attribute,
int_or_none, int_or_none,
lowercase_escape, lowercase_escape,
std_headers,
try_get, try_get,
) )
@ -239,6 +244,8 @@ class InstagramUserIE(InfoExtractor):
} }
} }
_gis_tmpl = None
def _entries(self, data): def _entries(self, data):
def get_count(suffix): def get_count(suffix):
return int_or_none(try_get( return int_or_none(try_get(
@ -257,16 +264,36 @@ class InstagramUserIE(InfoExtractor):
'first': 100, 'first': 100,
'after': cursor, 'after': cursor,
}) })
s = '%s:%s:%s' % (rhx_gis, csrf_token, variables)
media = self._download_json( if self._gis_tmpl:
'https://www.instagram.com/graphql/query/', uploader_id, gis_tmpls = [self._gis_tmpl]
'Downloading JSON page %d' % page_num, headers={ else:
'X-Requested-With': 'XMLHttpRequest', gis_tmpls = [
'X-Instagram-GIS': hashlib.md5(s.encode('utf-8')).hexdigest(), '%s' % rhx_gis,
}, query={ '',
'query_hash': '472f257a40c653c64c666ce877d59d2b', '%s:%s' % (rhx_gis, csrf_token),
'variables': variables, '%s:%s:%s' % (rhx_gis, csrf_token, std_headers['User-Agent']),
})['data']['user']['edge_owner_to_timeline_media'] ]
for gis_tmpl in gis_tmpls:
try:
media = self._download_json(
'https://www.instagram.com/graphql/query/', uploader_id,
'Downloading JSON page %d' % page_num, headers={
'X-Requested-With': 'XMLHttpRequest',
'X-Instagram-GIS': hashlib.md5(
('%s:%s' % (gis_tmpl, variables)).encode('utf-8')).hexdigest(),
}, query={
'query_hash': '42323d64886122307be10013ad2dcc44',
'variables': variables,
})['data']['user']['edge_owner_to_timeline_media']
self._gis_tmpl = gis_tmpl
break
except ExtractorError as e:
if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
if gis_tmpl != gis_tmpls[-1]:
continue
raise
edges = media.get('edges') edges = media.get('edges')
if not edges or not isinstance(edges, list): if not edges or not isinstance(edges, list):