[Imgur] Overhaul extractor module (#32612)

Revise extractors for new API and page formats

youtube_dl/extractor/imgur.py

@@ -1,101 +1,267 @@
+# coding: utf-8
 from __future__ import unicode_literals
 
 import re
 
 from .common import InfoExtractor
 from ..utils import (
+    determine_ext,
+    ExtractorError,
+    float_or_none,
     int_or_none,
     js_to_json,
+    merge_dicts,
     mimetype2ext,
-    ExtractorError,
+    parse_iso8601,
+    T,
+    traverse_obj,
+    txt_or_none,
+    url_or_none,
 )
 
 
-class ImgurIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:i\.)?imgur\.com/(?!(?:a|gallery|(?:t(?:opic)?|r)/[^/]+)/)(?P<id>[a-zA-Z0-9]+)'
+class ImgurBaseIE(InfoExtractor):
+    # hard-coded value, as also used by ArchiveTeam
+    _CLIENT_ID = '546c25a59c58ad7'
+
+    @classmethod
+    def _imgur_result(cls, item_id):
+        return cls.url_result('imgur:%s' % item_id, ImgurIE.ie_key(), item_id)
+
+    def _call_api(self, endpoint, video_id, **kwargs):
+        return self._download_json(
+            'https://api.imgur.com/post/v1/%s/%s?client_id=%s&include=media,account' % (endpoint, video_id, self._CLIENT_ID),
+            video_id, **kwargs)
+
+    @staticmethod
+    def get_description(s):
+        if 'Discover the magic of the internet at Imgur' in s:
+            return None
+        return txt_or_none(s)
+
+
+class ImgurIE(ImgurBaseIE):
+    _VALID_URL = r'''(?x)
+        (?:
+            https?://(?:i\.)?imgur\.com/(?!(?:a|gallery|t|topic|r)/)|
+            imgur:
+        )(?P<id>[a-zA-Z0-9]+)
+    '''
 
     _TESTS = [{
-        'url': 'https://i.imgur.com/A61SaA1.gifv',
+        'url': 'https://imgur.com/A61SaA1',
         'info_dict': {
             'id': 'A61SaA1',
             'ext': 'mp4',
             'title': 're:Imgur GIF$|MRW gifv is up and running without any bugs$',
+            'timestamp': 1416446068,
+            'upload_date': '20141120',
         },
     }, {
-        'url': 'https://imgur.com/A61SaA1',
+        'url': 'https://i.imgur.com/A61SaA1.gifv',
         'only_matching': True,
     }, {
         'url': 'https://i.imgur.com/crGpqCV.mp4',
         'only_matching': True,
     }, {
-        # no title
+        # previously, no title
         'url': 'https://i.imgur.com/jxBXAMC.gifv',
-        'only_matching': True,
+        'info_dict': {
+            'id': 'jxBXAMC',
+            'ext': 'mp4',
+            'title': 'Fahaka puffer feeding',
+            'timestamp': 1533835503,
+            'upload_date': '20180809',
+        },
     }]
 
+    def _extract_twitter_formats(self, html, tw_id='twitter', **kwargs):
+        fatal = kwargs.pop('fatal', False)
+        tw_stream = self._html_search_meta('twitter:player:stream', html, fatal=fatal, **kwargs)
+        if not tw_stream:
+            return []
+        ext = mimetype2ext(self._html_search_meta(
+            'twitter:player:stream:content_type', html, default=None))
+        width, height = (int_or_none(self._html_search_meta('twitter:player:' + v, html, default=None))
+                         for v in ('width', 'height'))
+        return [{
+            'format_id': tw_id,
+            'url': tw_stream,
+            'ext': ext or determine_ext(tw_stream),
+            'width': width,
+            'height': height,
+        }]
+
     def _real_extract(self, url):
         video_id = self._match_id(url)
+        data = self._call_api('media', video_id, fatal=False, expected_status=404)
         webpage = self._download_webpage(
-            'https://i.imgur.com/{id}.gifv'.format(id=video_id), video_id)
+            'https://i.imgur.com/{id}.gifv'.format(id=video_id), video_id, fatal=not data) or ''
+
+        if not traverse_obj(data, ('media', 0, (
+                ('type', T(lambda t: t == 'video' or None)),
+                ('metadata', 'is_animated'))), get_all=False):
+            raise ExtractorError(
+                '%s is not a video or animated image' % video_id,
+                expected=True)
+
+        media_fmt = traverse_obj(data, ('media', 0, {
+            'url': ('url', T(url_or_none)),
+            'ext': 'ext',
+            'width': ('width', T(int_or_none)),
+            'height': ('height', T(int_or_none)),
+            'filesize': ('size', T(int_or_none)),
+            'acodec': ('metadata', 'has_sound', T(lambda b: None if b else 'none')),
+        }))
 
-        width = int_or_none(self._og_search_property(
-            'video:width', webpage, default=None))
-        height = int_or_none(self._og_search_property(
-            'video:height', webpage, default=None))
+        media_url = traverse_obj(media_fmt, 'url')
+        if media_url:
+            if not media_fmt.get('ext'):
+                media_fmt['ext'] = mimetype2ext(traverse_obj(
+                    data, ('media', 0, 'mime_type'))) or determine_ext(media_url)
+            if traverse_obj(data, ('media', 0, 'type')) == 'image':
+                media_fmt['acodec'] = 'none'
+                media_fmt.setdefault('preference', -10)
+
+        tw_formats = self._extract_twitter_formats(webpage)
+        if traverse_obj(tw_formats, (0, 'url')) == media_url:
+            tw_formats = []
+        else:
+            # maybe this isn't an animated image/video?
+            self._check_formats(tw_formats, video_id)
 
         video_elements = self._search_regex(
             r'(?s)<div class="video-elements">(.*?)</div>',
             webpage, 'video elements', default=None)
-        if not video_elements:
+        if not (video_elements or tw_formats or media_url):
             raise ExtractorError(
-                'No sources found for video %s. Maybe an image?' % video_id,
+                'No sources found for video %s. Maybe a plain image?' % video_id,
                 expected=True)
 
-        formats = []
-        for m in re.finditer(r'<source\s+src="(?P<src>[^"]+)"\s+type="(?P<type>[^"]+)"', video_elements):
-            formats.append({
-                'format_id': m.group('type').partition('/')[2],
-                'url': self._proto_relative_url(m.group('src')),
-                'ext': mimetype2ext(m.group('type')),
-                'width': width,
-                'height': height,
+        def mung_format(fmt, *extra):
+            fmt.update({
                 'http_headers': {
                     'User-Agent': 'youtube-dl (like wget)',
                 },
             })
+            for d in extra:
+                fmt.update(d)
+            return fmt
 
-        gif_json = self._search_regex(
-            r'(?s)var\s+videoItem\s*=\s*(\{.*?\})',
-            webpage, 'GIF code', fatal=False)
-        if gif_json:
-            gifd = self._parse_json(
-                gif_json, video_id, transform_source=js_to_json)
-            formats.append({
-                'format_id': 'gif',
-                'preference': -10,
-                'width': width,
-                'height': height,
-                'ext': 'gif',
-                'acodec': 'none',
-                'vcodec': 'gif',
-                'container': 'gif',
-                'url': self._proto_relative_url(gifd['gifUrl']),
-                'filesize': gifd.get('size'),
-                'http_headers': {
-                    'User-Agent': 'youtube-dl (like wget)',
-                },
-            })
+        if video_elements:
+            def og_get_size(media_type):
+                return dict((p, int_or_none(self._og_search_property(
+                    ':'.join((media_type, p)), webpage, default=None)))
+                    for p in ('width', 'height'))
+
+            size = og_get_size('video')
+            if all(v is None for v in size.values()):
+                size = og_get_size('image')
+
+            formats = traverse_obj(
+                re.finditer(r'<source\s+src="(?P<src>[^"]+)"\s+type="(?P<type>[^"]+)"', video_elements),
+                (Ellipsis, {
+                    'format_id': ('type', T(lambda s: s.partition('/')[2])),
+                    'url': ('src', T(self._proto_relative_url)),
+                    'ext': ('type', T(mimetype2ext)),
+                }, T(lambda f: mung_format(f, size))))
+
+            gif_json = self._search_regex(
+                r'(?s)var\s+videoItem\s*=\s*(\{.*?\})',
+                webpage, 'GIF code', fatal=False)
+            MUST_BRANCH = (None, T(lambda _: None))
+            formats.extend(traverse_obj(gif_json, (
+                T(lambda j: self._parse_json(
+                    j, video_id, transform_source=js_to_json, fatal=False)), {
+                        'url': ('gifUrl', T(self._proto_relative_url)),
+                        'filesize': ('size', T(int_or_none)),
+                }, T(lambda f: mung_format(f, size, {
+                    'format_id': 'gif',
+                    'preference': -10,  # gifs are worse than videos
+                    'ext': 'gif',
+                    'acodec': 'none',
+                    'vcodec': 'gif',
+                    'container': 'gif',
+                })), MUST_BRANCH)))
+        else:
+            formats = []
+
+        # maybe add formats from JSON or page Twitter metadata
+        if not any((u == media_url) for u in traverse_obj(formats, (Ellipsis, 'url'))):
+            formats.append(mung_format(media_fmt))
+        tw_url = traverse_obj(tw_formats, (0, 'url'))
+        if not any((u == tw_url) for u in traverse_obj(formats, (Ellipsis, 'url'))):
+            formats.extend(mung_format(f) for f in tw_formats)
 
         self._sort_formats(formats)
 
-        return {
+        return merge_dicts(traverse_obj(data, {
+            'uploader_id': ('account_id', T(txt_or_none),
+                            T(lambda a: a if int_or_none(a) != 0 else None)),
+            'uploader': ('account', 'username', T(txt_or_none)),
+            'uploader_url': ('account', 'avatar_url', T(url_or_none)),
+            'like_count': ('upvote_count', T(int_or_none)),
+            'dislike_count': ('downvote_count', T(int_or_none)),
+            'comment_count': ('comment_count', T(int_or_none)),
+            'age_limit': ('is_mature', T(lambda x: 18 if x else None)),
+            'timestamp': (('updated_at', 'created_at'), T(parse_iso8601)),
+            'release_timestamp': ('created_at', T(parse_iso8601)),
+        }, get_all=False), traverse_obj(data, ('media', 0, 'metadata', {
+            'title': ('title', T(txt_or_none)),
+            'description': ('description', T(self.get_description)),
+            'duration': ('duration', T(float_or_none)),
+            'timestamp': (('updated_at', 'created_at'), T(parse_iso8601)),
+            'release_timestamp': ('created_at', T(parse_iso8601)),
+        })), {
             'id': video_id,
             'formats': formats,
-            'title': self._og_search_title(webpage, default=video_id),
-        }
+            'title': self._og_search_title(webpage, default='Imgur video ' + video_id),
+            'description': self.get_description(self._og_search_description(webpage)),
+            'thumbnail': url_or_none(self._html_search_meta('thumbnailUrl', webpage, default=None)),
+        })
+
+
+class ImgurGalleryBaseIE(ImgurBaseIE):
+    _GALLERY = True
+
+    def _real_extract(self, url):
+        gallery_id = self._match_id(url)
 
+        data = self._call_api('albums', gallery_id, fatal=False, expected_status=404)
 
-class ImgurGalleryIE(InfoExtractor):
+        info = traverse_obj(data, {
+            'title': ('title', T(txt_or_none)),
+            'description': ('description', T(self.get_description)),
+        })
+
+        if traverse_obj(data, 'is_album'):
+
+            def yield_media_ids():
+                for m_id in traverse_obj(data, (
+                        'media', lambda _, v: v.get('type') == 'video' or v['metadata']['is_animated'],
+                        'id', T(txt_or_none))):
+                    yield m_id
+
+            # if a gallery with exactly one video, apply album metadata to video
+            media_id = (
+                self._GALLERY
+                and traverse_obj(data, ('image_count', T(lambda c: c == 1)))
+                and next(yield_media_ids(), None))
+
+            if not media_id:
+                result = self.playlist_result(
+                    map(self._imgur_result, yield_media_ids()), gallery_id)
+                result.update(info)
+                return result
+            gallery_id = media_id
+
+        result = self._imgur_result(gallery_id)
+        info['_type'] = 'url_transparent'
+        result.update(info)
+        return result
+
+
+class ImgurGalleryIE(ImgurGalleryBaseIE):
     IE_NAME = 'imgur:gallery'
     _VALID_URL = r'https?://(?:i\.)?imgur\.com/(?:gallery|(?:t(?:opic)?|r)/[^/]+)/(?P<id>[a-zA-Z0-9]+)'
 
@@ -106,49 +272,93 @@ class ImgurGalleryIE(InfoExtractor):
             'title': 'Adding faces make every GIF better',
         },
         'playlist_count': 25,
+        'skip': 'Zoinks! You\'ve taken a wrong turn.',
     }, {
+        # TODO: static images - replace with animated/video gallery
         'url': 'http://imgur.com/topic/Aww/ll5Vk',
         'only_matching': True,
     }, {
         'url': 'https://imgur.com/gallery/YcAQlkx',
+        'add_ies': ['Imgur'],
         'info_dict': {
             'id': 'YcAQlkx',
             'ext': 'mp4',
             'title': 'Classic Steve Carell gif...cracks me up everytime....damn the repost downvotes....',
-        }
+            'timestamp': 1358554297,
+            'upload_date': '20130119',
+            'uploader_id': '1648642',
+            'uploader': 'wittyusernamehere',
+        },
     }, {
+        # TODO: static image - replace with animated/video gallery
         'url': 'http://imgur.com/topic/Funny/N8rOudd',
         'only_matching': True,
     }, {
         'url': 'http://imgur.com/r/aww/VQcQPhM',
-        'only_matching': True,
+        'add_ies': ['Imgur'],
+        'info_dict': {
+            'id': 'VQcQPhM',
+            'ext': 'mp4',
+            'title': 'The boss is here',
+            'timestamp': 1476494751,
+            'upload_date': '20161015',
+            'uploader_id': '19138530',
+            'uploader': 'thematrixcam',
+        },
+    },
+        # from PR #16674
+        {
+        'url': 'https://imgur.com/t/unmuted/6lAn9VQ',
+        'info_dict': {
+            'id': '6lAn9VQ',
+            'title': 'Penguins !',
+        },
+        'playlist_count': 3,
+    }, {
+        'url': 'https://imgur.com/t/unmuted/kx2uD3C',
+        'add_ies': ['Imgur'],
+        'info_dict': {
+            'id': 'ZVMv45i',
+            'ext': 'mp4',
+            'title': 'Intruder',
+            'timestamp': 1528129683,
+            'upload_date': '20180604',
+        },
+    }, {
+        'url': 'https://imgur.com/t/unmuted/wXSK0YH',
+        'add_ies': ['Imgur'],
+        'info_dict': {
+            'id': 'JCAP4io',
+            'ext': 'mp4',
+            'title': 're:I got the blues$',
+            'description': 'Luka’s vocal stylings.\n\nFP edit: don’t encourage me. I’ll never stop posting Luka and friends.',
+            'timestamp': 1527809525,
+            'upload_date': '20180531',
+        },
     }]
 
-    def _real_extract(self, url):
-        gallery_id = self._match_id(url)
-
-        data = self._download_json(
-            'https://imgur.com/gallery/%s.json' % gallery_id,
-            gallery_id)['data']['image']
 
-        if data.get('is_album'):
-            entries = [
-                self.url_result('http://imgur.com/%s' % image['hash'], ImgurIE.ie_key(), image['hash'])
-                for image in data['album_images']['images'] if image.get('hash')]
-            return self.playlist_result(entries, gallery_id, data.get('title'), data.get('description'))
-
-        return self.url_result('http://imgur.com/%s' % gallery_id, ImgurIE.ie_key(), gallery_id)
-
-
-class ImgurAlbumIE(ImgurGalleryIE):
+class ImgurAlbumIE(ImgurGalleryBaseIE):
     IE_NAME = 'imgur:album'
     _VALID_URL = r'https?://(?:i\.)?imgur\.com/a/(?P<id>[a-zA-Z0-9]+)'
-
+    _GALLERY = False
     _TESTS = [{
+        # TODO: only static images - replace with animated/video gallery
         'url': 'http://imgur.com/a/j6Orj',
+        'only_matching': True,
+    },
+        # from PR #21693
+        {
+        'url': 'https://imgur.com/a/iX265HX',
+        'info_dict': {
+            'id': 'iX265HX',
+            'title': 'enen-no-shouboutai'
+        },
+        'playlist_count': 2,
+    }, {
+        'url': 'https://imgur.com/a/8pih2Ed',
         'info_dict': {
-            'id': 'j6Orj',
-            'title': 'A Literary Analysis of "Star Wars: The Force Awakens"',
+            'id': '8pih2Ed'
         },
-        'playlist_count': 12,
+        'playlist_mincount': 1,
     }]