Merge tag 'upstream/2016.08.17'

[youtubedl] / youtube_dl / extractor / facebook.py

diff --git a/youtube_dl/extractor/facebook.py b/youtube_dl/extractor/facebook.py
index f8bdfc2d33c9f00b9f902a4303eb7024f4646312..0fb781a733f4c19780ed88f8f5b24c1102b10a44 100644 (file)
--- a/youtube_dl/extractor/facebook.py
+++ b/youtube_dl/extractor/facebook.py
@@ -1,54 +1,166 @@
+from __future__ import unicode_literals
+

 import json
 import json

-import netrc

 import re
 import socket
 
 from .common import InfoExtractor
 import re
 import socket
 
 from .common import InfoExtractor

-from ..utils import (
+from ..compat import (
+    compat_etree_fromstring,

     compat_http_client,
     compat_http_client,

-    compat_str,

     compat_urllib_error,
     compat_urllib_error,

-    compat_urllib_parse,
-    compat_urllib_request,
-
+    compat_urllib_parse_unquote,
+    compat_urllib_parse_unquote_plus,
+)
+from ..utils import (
+    error_to_compat_str,

     ExtractorError,
     ExtractorError,

+    limit_length,
+    sanitized_Request,
+    urlencode_postdata,
+    get_element_by_id,
+    clean_html,

 )
 
 
 class FacebookIE(InfoExtractor):
 )
 
 
 class FacebookIE(InfoExtractor):

-    """Information Extractor for Facebook"""
-
-    _VALID_URL = r'^(?:https?://)?(?:\w+\.)?facebook\.com/(?:video/video|photo)\.php\?(?:.*?)v=(?P<ID>\d+)(?:.*)'
+    _VALID_URL = r'''(?x)
+                (?:
+                    https?://
+                        (?:[\w-]+\.)?facebook\.com/
+                        (?:[^#]*?\#!/)?
+                        (?:
+                            (?:
+                                video/video\.php|
+                                photo\.php|
+                                video\.php|
+                                video/embed|
+                                story\.php
+                            )\?(?:.*?)(?:v|video_id|story_fbid)=|
+                            [^/]+/videos/(?:[^/]+/)?|
+                            [^/]+/posts/|
+                            groups/[^/]+/permalink/
+                        )|
+                    facebook:
+                )
+                (?P<id>[0-9]+)
+                '''

     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
     _NETRC_MACHINE = 'facebook'
     _LOGIN_URL = 'https://www.facebook.com/login.php?next=http%3A%2F%2Ffacebook.com%2Fhome.php&login_attempt=1'
     _CHECKPOINT_URL = 'https://www.facebook.com/checkpoint/?next=http%3A%2F%2Ffacebook.com%2Fhome.php&_fb_noscript=1'
     _NETRC_MACHINE = 'facebook'

-    IE_NAME = u'facebook'
-    _TEST = {
-        u'url': u'https://www.facebook.com/photo.php?v=120708114770723',
-        u'file': u'120708114770723.mp4',
-        u'md5': u'48975a41ccc4b7a581abd68651c1a5a8',
-        u'info_dict': {
-            u"duration": 279, 
-            u"title": u"PEOPLE ARE AWESOME 2013"
+    IE_NAME = 'facebook'
+
+    _CHROME_USER_AGENT = 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.97 Safari/537.36'
+
+    _VIDEO_PAGE_TEMPLATE = 'https://www.facebook.com/video/video.php?v=%s'
+
+    _TESTS = [{
+        'url': 'https://www.facebook.com/video.php?v=637842556329505&fref=nf',
+        'md5': '6a40d33c0eccbb1af76cf0485a052659',
+        'info_dict': {
+            'id': '637842556329505',
+            'ext': 'mp4',
+            'title': 're:Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam',
+            'uploader': 'Tennis on Facebook',

         }
         }

-    }
+    }, {
+        'note': 'Video without discernible title',
+        'url': 'https://www.facebook.com/video.php?v=274175099429670',
+        'info_dict': {
+            'id': '274175099429670',
+            'ext': 'mp4',
+            'title': 'Facebook video #274175099429670',
+            'uploader': 'Asif Nawab Butt',
+        },
+        'expected_warnings': [
+            'title'
+        ]
+    }, {
+        'note': 'Video with DASH manifest',
+        'url': 'https://www.facebook.com/video.php?v=957955867617029',
+        'md5': '54706e4db4f5ad58fbad82dde1f1213f',
+        'info_dict': {
+            'id': '957955867617029',
+            'ext': 'mp4',
+            'title': 'When you post epic content on instagram.com/433 8 million followers, this is ...',
+            'uploader': 'Demy de Zeeuw',
+        },
+    }, {
+        'url': 'https://www.facebook.com/maxlayn/posts/10153807558977570',
+        'md5': '037b1fa7f3c2d02b7a0d7bc16031ecc6',
+        'info_dict': {
+            'id': '544765982287235',
+            'ext': 'mp4',
+            'title': '"What are you doing running in the snow?"',
+            'uploader': 'FailArmy',
+        }
+    }, {
+        'url': 'https://m.facebook.com/story.php?story_fbid=1035862816472149&id=116132035111903',
+        'md5': '1deb90b6ac27f7efcf6d747c8a27f5e3',
+        'info_dict': {
+            'id': '1035862816472149',
+            'ext': 'mp4',
+            'title': 'What the Flock Is Going On In New Zealand  Credit: ViralHog',
+            'uploader': 'S. Saint',
+        },
+    }, {
+        'note': 'swf params escaped',
+        'url': 'https://www.facebook.com/barackobama/posts/10153664894881749',
+        'md5': '97ba073838964d12c70566e0085c2b91',
+        'info_dict': {
+            'id': '10153664894881749',
+            'ext': 'mp4',
+            'title': 'Facebook video #10153664894881749',
+        },
+    }, {
+        'url': 'https://www.facebook.com/video.php?v=10204634152394104',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.facebook.com/amogood/videos/1618742068337349/?fref=nf',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.facebook.com/ChristyClarkForBC/videos/vb.22819070941/10153870694020942/?type=2&theater',
+        'only_matching': True,
+    }, {
+        'url': 'facebook:544765982287235',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.facebook.com/groups/164828000315060/permalink/764967300301124/',
+        'only_matching': True,
+    }, {
+        'url': 'https://zh-hk.facebook.com/peoplespower/videos/1135894589806027/',
+        'only_matching': True,
+    }]
+
+    @staticmethod
+    def _extract_url(webpage):
+        mobj = re.search(
+            r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
+        if mobj is not None:
+            return mobj.group('url')

 
 

-    def report_login(self):
-        """Report attempt to log in."""
-        self.to_screen(u'Logging in')
+        # Facebook API embed
+        # see https://developers.facebook.com/docs/plugins/embedded-video-player
+        mobj = re.search(r'''(?x)<div[^>]+
+                class=(?P<q1>[\'"])[^\'"]*\bfb-(?:video|post)\b[^\'"]*(?P=q1)[^>]+
+                data-href=(?P<q2>[\'"])(?P<url>(?:https?:)?//(?:www\.)?facebook.com/.+?)(?P=q2)''', webpage)
+        if mobj is not None:
+            return mobj.group('url')

 
     def _login(self):
         (useremail, password) = self._get_login_info()
         if useremail is None:
             return
 
 
     def _login(self):
         (useremail, password) = self._get_login_info()
         if useremail is None:
             return
 

-        login_page_req = compat_urllib_request.Request(self._LOGIN_URL)
-        login_page_req.add_header('Cookie', 'locale=en_US')
-        self.report_login()
-        login_page = self._download_webpage(login_page_req, None, note=False,
-            errnote=u'Unable to download login page')
-        lsd = self._search_regex(r'"lsd":"(\w*?)"', login_page, u'lsd')
-        lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, u'lgnrnd')
+        login_page_req = sanitized_Request(self._LOGIN_URL)
+        self._set_cookie('facebook.com', 'locale', 'en_US')
+        login_page = self._download_webpage(login_page_req, None,
+                                            note='Downloading login page',
+                                            errnote='Unable to download login page')
+        lsd = self._search_regex(
+            r'<input type="hidden" name="lsd" value="([^"]*)"',
+            login_page, 'lsd')
+        lgnrnd = self._search_regex(r'name="lgnrnd" value="([^"]*?)"', login_page, 'lgnrnd')

 
         login_form = {
             'email': useremail,
 
         login_form = {
             'email': useremail,


@@ -60,74 +172,171 @@ class FacebookIE(InfoExtractor):
             'legacy_return': '1',
             'timezone': '-60',
             'trynum': '1',
             'legacy_return': '1',
             'timezone': '-60',
             'trynum': '1',

-            }
-        request = compat_urllib_request.Request(self._LOGIN_URL, compat_urllib_parse.urlencode(login_form))
+        }
+        request = sanitized_Request(self._LOGIN_URL, urlencode_postdata(login_form))

         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
         try:
         request.add_header('Content-Type', 'application/x-www-form-urlencoded')
         try:

-            login_results = compat_urllib_request.urlopen(request).read()
+            login_results = self._download_webpage(request, None,
+                                                   note='Logging in', errnote='unable to fetch login page')

             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:
             if re.search(r'<form(.*)name="login"(.*)</form>', login_results) is not None:

-                self._downloader.report_warning(u'unable to log in: bad username/password, or exceded login rate limit (~3/min). Check credentials or wait.')
+                error = self._html_search_regex(
+                    r'(?s)<div[^>]+class=(["\']).*?login_error_box.*?\1[^>]*><div[^>]*>.*?</div><div[^>]*>(?P<error>.+?)</div>',
+                    login_results, 'login error', default=None, group='error')
+                if error:
+                    raise ExtractorError('Unable to login: %s' % error, expected=True)
+                self._downloader.report_warning('unable to log in: bad username/password, or exceeded login rate limit (~3/min). Check credentials or wait.')
+                return
+
+            fb_dtsg = self._search_regex(
+                r'name="fb_dtsg" value="(.+?)"', login_results, 'fb_dtsg', default=None)
+            h = self._search_regex(
+                r'name="h"\s+(?:\w+="[^"]+"\s+)*?value="([^"]+)"', login_results, 'h', default=None)
+
+            if not fb_dtsg or not h:

                 return
 
             check_form = {
                 return
 
             check_form = {

-                'fb_dtsg': self._search_regex(r'"fb_dtsg":"(.*?)"', login_results, u'fb_dtsg'),
-                'nh': self._search_regex(r'name="nh" value="(\w*?)"', login_results, u'nh'),
+                'fb_dtsg': fb_dtsg,
+                'h': h,

                 'name_action_selected': 'dont_save',
                 'name_action_selected': 'dont_save',

-                'submit[Continue]': self._search_regex(r'<input value="(.*?)" name="submit\[Continue\]"', login_results, u'continue'),

             }
             }

-            check_req = compat_urllib_request.Request(self._CHECKPOINT_URL, compat_urllib_parse.urlencode(check_form))
+            check_req = sanitized_Request(self._CHECKPOINT_URL, urlencode_postdata(check_form))

             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')
             check_req.add_header('Content-Type', 'application/x-www-form-urlencoded')

-            check_response = compat_urllib_request.urlopen(check_req).read()
+            check_response = self._download_webpage(check_req, None,
+                                                    note='Confirming login')

             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:
             if re.search(r'id="checkpointSubmitButton"', check_response) is not None:

-                self._downloader.report_warning(u'Unable to confirm login, you have to login in your brower and authorize the login.')
+                self._downloader.report_warning('Unable to confirm login, you have to login in your browser and authorize the login.')

         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:

-            self._downloader.report_warning(u'unable to log in: %s' % compat_str(err))
+            self._downloader.report_warning('unable to log in: %s' % error_to_compat_str(err))

             return
 
     def _real_initialize(self):
         self._login()
 
             return
 
     def _real_initialize(self):
         self._login()
 

-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        if mobj is None:
-            raise ExtractorError(u'Invalid URL: %s' % url)
-        video_id = mobj.group('ID')
+    def _extract_from_url(self, url, video_id, fatal_if_no_video=True):
+        req = sanitized_Request(url)
+        req.add_header('User-Agent', self._CHROME_USER_AGENT)
+        webpage = self._download_webpage(req, video_id)

 
 

-        url = 'https://www.facebook.com/video/video.php?v=%s' % video_id
-        webpage = self._download_webpage(url, video_id)
+        video_data = None

 
 

-        BEFORE = '{swf.addParam(param[0], param[1]);});\n'
+        BEFORE = '{swf.addParam(param[0], param[1]);});'

         AFTER = '.forEach(function(variable) {swf.addVariable(variable[0], variable[1]);});'
         AFTER = '.forEach(function(variable) {swf.addVariable(variable[0], variable[1]);});'

-        m = re.search(re.escape(BEFORE) + '(.*?)' + re.escape(AFTER), webpage)
-        if not m:
+        PATTERN = re.escape(BEFORE) + '(?:\n|\\\\n)(.*?)' + re.escape(AFTER)
+
+        for m in re.findall(PATTERN, webpage):
+            swf_params = m.replace('\\\\', '\\').replace('\\"', '"')
+            data = dict(json.loads(swf_params))
+            params_raw = compat_urllib_parse_unquote(data['params'])
+            video_data_candidate = json.loads(params_raw)['video_data']
+            for _, f in video_data_candidate.items():
+                if not f:
+                    continue
+                if isinstance(f, dict):
+                    f = [f]
+                if not isinstance(f, list):
+                    continue
+                if f[0].get('video_id') == video_id:
+                    video_data = video_data_candidate
+                    break
+            if video_data:
+                break
+
+        def video_data_list2dict(video_data):
+            ret = {}
+            for item in video_data:
+                format_id = item['stream_type']
+                ret.setdefault(format_id, []).append(item)
+            return ret
+
+        if not video_data:
+            server_js_data = self._parse_json(self._search_regex(
+                r'handleServerJS\(({.+})\);', webpage, 'server js data', default='{}'), video_id)
+            for item in server_js_data.get('instances', []):
+                if item[1][0] == 'VideoConfig':
+                    video_data = video_data_list2dict(item[2][0]['videoData'])
+                    break
+
+        if not video_data:
+            if not fatal_if_no_video:
+                return webpage, False

             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
             if m_msg is not None:
                 raise ExtractorError(
             m_msg = re.search(r'class="[^"]*uiInterstitialContent[^"]*"><div>(.*?)</div>', webpage)
             if m_msg is not None:
                 raise ExtractorError(

-                    u'The video is not available, Facebook said: "%s"' % m_msg.group(1),
+                    'The video is not available, Facebook said: "%s"' % m_msg.group(1),

                     expected=True)
             else:
                     expected=True)
             else:

-                raise ExtractorError(u'Cannot parse data')
-        data = dict(json.loads(m.group(1)))
-        params_raw = compat_urllib_parse.unquote(data['params'])
-        params = json.loads(params_raw)
-        video_data = params['video_data'][0]
-        video_url = video_data.get('hd_src')
-        if not video_url:
-            video_url = video_data['sd_src']
-        if not video_url:
-            raise ExtractorError(u'Cannot find video URL')
-        video_duration = int(video_data['video_duration'])
-        thumbnail = video_data['thumbnail_src']
+                raise ExtractorError('Cannot parse data')
+
+        formats = []
+        for format_id, f in video_data.items():
+            if f and isinstance(f, dict):
+                f = [f]
+            if not f or not isinstance(f, list):
+                continue
+            for quality in ('sd', 'hd'):
+                for src_type in ('src', 'src_no_ratelimit'):
+                    src = f[0].get('%s_%s' % (quality, src_type))
+                    if src:
+                        preference = -10 if format_id == 'progressive' else 0
+                        if quality == 'hd':
+                            preference += 5
+                        formats.append({
+                            'format_id': '%s_%s_%s' % (format_id, quality, src_type),
+                            'url': src,
+                            'preference': preference,
+                        })
+            dash_manifest = f[0].get('dash_manifest')
+            if dash_manifest:
+                formats.extend(self._parse_mpd_formats(
+                    compat_etree_fromstring(compat_urllib_parse_unquote_plus(dash_manifest))))
+        if not formats:
+            raise ExtractorError('Cannot find video formats')
+
+        self._sort_formats(formats)

 
         video_title = self._html_search_regex(
 
         video_title = self._html_search_regex(

-            r'<h2 class="uiHeaderTitle">([^<]*)</h2>', webpage, u'title')
+            r'<h2\s+[^>]*class="uiHeaderTitle"[^>]*>([^<]*)</h2>', webpage, 'title',
+            default=None)
+        if not video_title:
+            video_title = self._html_search_regex(
+                r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
+                webpage, 'alternative title', default=None)
+            video_title = limit_length(video_title, 80)
+        if not video_title:
+            video_title = 'Facebook video #%s' % video_id
+        uploader = clean_html(get_element_by_id('fbPhotoPageAuthorName', webpage))

 
 

-        info = {
+        info_dict = {

             'id': video_id,
             'title': video_title,
             'id': video_id,
             'title': video_title,

-            'url': video_url,
-            'ext': 'mp4',
-            'duration': video_duration,
-            'thumbnail': thumbnail,
+            'formats': formats,
+            'uploader': uploader,

         }
         }

-        return [info]
+
+        return webpage, info_dict
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+
+        real_url = self._VIDEO_PAGE_TEMPLATE % video_id if url.startswith('facebook:') else url
+        webpage, info_dict = self._extract_from_url(real_url, video_id, fatal_if_no_video=False)
+
+        if info_dict:
+            return info_dict
+
+        if '/posts/' in url:
+            entries = [
+                self.url_result('facebook:%s' % vid, FacebookIE.ie_key())
+                for vid in self._parse_json(
+                    self._search_regex(
+                        r'(["\'])video_ids\1\s*:\s*(?P<ids>\[.+?\])',
+                        webpage, 'video ids', group='ids'),
+                    video_id)]
+
+            return self.playlist_result(entries, video_id)
+        else:
+            _, info_dict = self._extract_from_url(
+                self._VIDEO_PAGE_TEMPLATE % video_id,
+                video_id, fatal_if_no_video=True)
+            return info_dict