from __future__ import unicode_literals
+import itertools
+import hashlib
+import json
import re
from .common import InfoExtractor
-from ..compat import compat_str
+from ..compat import (
+ compat_str,
+ compat_HTTPError,
+)
from ..utils import (
+ ExtractorError,
get_element_by_attribute,
int_or_none,
- limit_length,
lowercase_escape,
+ std_headers,
try_get,
+ url_or_none,
)
video_url = media.get('video_url')
height = int_or_none(media.get('dimensions', {}).get('height'))
width = int_or_none(media.get('dimensions', {}).get('width'))
- description = media.get('caption')
+ description = try_get(
+ media, lambda x: x['edge_media_to_caption']['edges'][0]['node']['text'],
+ compat_str) or media.get('caption')
thumbnail = media.get('display_src')
- timestamp = int_or_none(media.get('date'))
+ timestamp = int_or_none(media.get('taken_at_timestamp') or media.get('date'))
uploader = media.get('owner', {}).get('full_name')
uploader_id = media.get('owner', {}).get('username')
- like_count = int_or_none(media.get('likes', {}).get('count'))
- comment_count = int_or_none(media.get('comments', {}).get('count'))
+
+ def get_count(key, kind):
+ return int_or_none(try_get(
+ media, (lambda x: x['edge_media_%s' % key]['count'],
+ lambda x: x['%ss' % kind]['count'])))
+ like_count = get_count('preview_like', 'like')
+ comment_count = get_count('to_comment', 'comment')
+
comments = [{
'author': comment.get('user', {}).get('username'),
'author_id': comment.get('user', {}).get('id'),
node = try_get(edge, lambda x: x['node'], dict)
if not node:
continue
- node_video_url = try_get(node, lambda x: x['video_url'], compat_str)
+ node_video_url = url_or_none(node.get('video_url'))
if not node_video_url:
continue
entries.append({
}
-class InstagramUserIE(InfoExtractor):
- _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<username>[^/]{2,})/?(?:$|[?#])'
+class InstagramPlaylistIE(InfoExtractor):
+ # A superclass for handling any kind of query based on GraphQL which
+ # results in a playlist.
+
+ _gis_tmpl = None # used to cache GIS request type
+
+ def _parse_graphql(self, webpage, item_id):
+ # Reads a webpage and returns its GraphQL data.
+ return self._parse_json(
+ self._search_regex(
+ r'sharedData\s*=\s*({.+?})\s*;\s*[<\n]', webpage, 'data'),
+ item_id)
+
+ def _extract_graphql(self, data, url):
+ # Parses GraphQL queries containing videos and generates a playlist.
+ def get_count(suffix):
+ return int_or_none(try_get(
+ node, lambda x: x['edge_media_' + suffix]['count']))
+
+ uploader_id = self._match_id(url)
+ csrf_token = data['config']['csrf_token']
+ rhx_gis = data.get('rhx_gis') or '3c7ca9dcefcf966d11dacf1f151335e8'
+
+ cursor = ''
+ for page_num in itertools.count(1):
+ variables = {
+ 'first': 12,
+ 'after': cursor,
+ }
+ variables.update(self._query_vars_for(data))
+ variables = json.dumps(variables)
+
+ if self._gis_tmpl:
+ gis_tmpls = [self._gis_tmpl]
+ else:
+ gis_tmpls = [
+ '%s' % rhx_gis,
+ '',
+ '%s:%s' % (rhx_gis, csrf_token),
+ '%s:%s:%s' % (rhx_gis, csrf_token, std_headers['User-Agent']),
+ ]
+
+ # try all of the ways to generate a GIS query, and not only use the
+ # first one that works, but cache it for future requests
+ for gis_tmpl in gis_tmpls:
+ try:
+ json_data = 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': self._QUERY_HASH,
+ 'variables': variables,
+ })
+ media = self._parse_timeline_from(json_data)
+ self._gis_tmpl = gis_tmpl
+ break
+ except ExtractorError as e:
+ # if it's an error caused by a bad query, and there are
+ # more GIS templates to try, ignore it and keep trying
+ if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
+ if gis_tmpl != gis_tmpls[-1]:
+ continue
+ raise
+
+ edges = media.get('edges')
+ if not edges or not isinstance(edges, list):
+ break
+
+ for edge in edges:
+ node = edge.get('node')
+ if not node or not isinstance(node, dict):
+ continue
+ if node.get('__typename') != 'GraphVideo' and node.get('is_video') is not True:
+ continue
+ video_id = node.get('shortcode')
+ if not video_id:
+ continue
+
+ info = self.url_result(
+ 'https://instagram.com/p/%s/' % video_id,
+ ie=InstagramIE.ie_key(), video_id=video_id)
+
+ description = try_get(
+ node, lambda x: x['edge_media_to_caption']['edges'][0]['node']['text'],
+ compat_str)
+ thumbnail = node.get('thumbnail_src') or node.get('display_src')
+ timestamp = int_or_none(node.get('taken_at_timestamp'))
+
+ comment_count = get_count('to_comment')
+ like_count = get_count('preview_like')
+ view_count = int_or_none(node.get('video_view_count'))
+
+ info.update({
+ 'description': description,
+ 'thumbnail': thumbnail,
+ 'timestamp': timestamp,
+ 'comment_count': comment_count,
+ 'like_count': like_count,
+ 'view_count': view_count,
+ })
+
+ yield info
+
+ page_info = media.get('page_info')
+ if not page_info or not isinstance(page_info, dict):
+ break
+
+ has_next_page = page_info.get('has_next_page')
+ if not has_next_page:
+ break
+
+ cursor = page_info.get('end_cursor')
+ if not cursor or not isinstance(cursor, compat_str):
+ break
+
+ def _real_extract(self, url):
+ user_or_tag = self._match_id(url)
+ webpage = self._download_webpage(url, user_or_tag)
+ data = self._parse_graphql(webpage, user_or_tag)
+
+ self._set_cookie('instagram.com', 'ig_pr', '1')
+
+ return self.playlist_result(
+ self._extract_graphql(data, url), user_or_tag, user_or_tag)
+
+
+class InstagramUserIE(InstagramPlaylistIE):
+ _VALID_URL = r'https?://(?:www\.)?instagram\.com/(?P<id>[^/]{2,})/?(?:$|[?#])'
IE_DESC = 'Instagram user profile'
IE_NAME = 'instagram:user'
_TEST = {
'id': 'porsche',
'title': 'porsche',
},
- 'playlist_mincount': 2,
- 'playlist': [{
- 'info_dict': {
- 'id': '614605558512799803_462752227',
- 'ext': 'mp4',
- 'title': '#Porsche Intelligent Performance.',
- 'thumbnail': r're:^https?://.*\.jpg',
- 'uploader': 'Porsche',
- 'uploader_id': 'porsche',
- 'timestamp': 1387486713,
- 'upload_date': '20131219',
- },
- }],
+ 'playlist_count': 5,
'params': {
'extract_flat': True,
'skip_download': True,
+ 'playlistend': 5,
}
}
- def _real_extract(self, url):
- mobj = re.match(self._VALID_URL, url)
- uploader_id = mobj.group('username')
-
- entries = []
- page_count = 0
- media_url = 'http://instagram.com/%s/media' % uploader_id
- while True:
- page = self._download_json(
- media_url, uploader_id,
- note='Downloading page %d ' % (page_count + 1),
- )
- page_count += 1
-
- for it in page['items']:
- if it.get('type') != 'video':
- continue
- like_count = int_or_none(it.get('likes', {}).get('count'))
- user = it.get('user', {})
-
- formats = [{
- 'format_id': k,
- 'height': v.get('height'),
- 'width': v.get('width'),
- 'url': v['url'],
- } for k, v in it['videos'].items()]
- self._sort_formats(formats)
-
- thumbnails_el = it.get('images', {})
- thumbnail = thumbnails_el.get('thumbnail', {}).get('url')
-
- # In some cases caption is null, which corresponds to None
- # in python. As a result, it.get('caption', {}) gives None
- title = (it.get('caption') or {}).get('text', it['id'])
-
- entries.append({
- 'id': it['id'],
- 'title': limit_length(title, 80),
- 'formats': formats,
- 'thumbnail': thumbnail,
- 'webpage_url': it.get('link'),
- 'uploader': user.get('full_name'),
- 'uploader_id': user.get('username'),
- 'like_count': like_count,
- 'timestamp': int_or_none(it.get('created_time')),
- })
+ _QUERY_HASH = '42323d64886122307be10013ad2dcc44',
- if not page['items']:
- break
- max_id = page['items'][-1]['id'].split('_')[0]
- media_url = (
- 'http://instagram.com/%s/media?max_id=%s' % (
- uploader_id, max_id))
+ @staticmethod
+ def _parse_timeline_from(data):
+ # extracts the media timeline data from a GraphQL result
+ return data['data']['user']['edge_owner_to_timeline_media']
+
+ @staticmethod
+ def _query_vars_for(data):
+ # returns a dictionary of variables to add to the timeline query based
+ # on the GraphQL of the original page
+ return {
+ 'id': data['entry_data']['ProfilePage'][0]['graphql']['user']['id']
+ }
+
+class InstagramTagIE(InstagramPlaylistIE):
+ _VALID_URL = r'https?://(?:www\.)?instagram\.com/explore/tags/(?P<id>[^/]+)'
+ IE_DESC = 'Instagram hashtag search'
+ IE_NAME = 'instagram:tag'
+ _TEST = {
+ 'url': 'https://instagram.com/explore/tags/lolcats',
+ 'info_dict': {
+ 'id': 'lolcats',
+ 'title': 'lolcats',
+ },
+ 'playlist_count': 50,
+ 'params': {
+ 'extract_flat': True,
+ 'skip_download': True,
+ 'playlistend': 50,
+ }
+ }
+
+ _QUERY_HASH = 'f92f56d47dc7a55b606908374b43a314',
+
+ @staticmethod
+ def _parse_timeline_from(data):
+ # extracts the media timeline data from a GraphQL result
+ return data['data']['hashtag']['edge_hashtag_to_media']
+
+ @staticmethod
+ def _query_vars_for(data):
+ # returns a dictionary of variables to add to the timeline query based
+ # on the GraphQL of the original page
return {
- '_type': 'playlist',
- 'entries': entries,
- 'id': uploader_id,
- 'title': uploader_id,
+ 'tag_name':
+ data['entry_data']['TagPage'][0]['graphql']['hashtag']['name']
}
Rapsys Git / youtubedl / blobdiff