]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/udemy.py
Fix extraction from youtube.
[youtubedl] / youtube_dl / extractor / udemy.py
index f5b5e7fd6775b4194ca18265f4c78f639a5f092a..105826e9bb04d8b0e39e2126f637607cc8f6159b 100644 (file)
@@ -1,23 +1,42 @@
 from __future__ import unicode_literals
 
+import re
+
 from .common import InfoExtractor
 from ..compat import (
     compat_HTTPError,
-    compat_urllib_parse,
+    compat_kwargs,
+    compat_str,
     compat_urllib_request,
+    compat_urlparse,
 )
 from ..utils import (
+    determine_ext,
+    extract_attributes,
     ExtractorError,
     float_or_none,
     int_or_none,
+    js_to_json,
     sanitized_Request,
+    try_get,
     unescapeHTML,
+    url_or_none,
+    urlencode_postdata,
 )
 
 
 class UdemyIE(InfoExtractor):
     IE_NAME = 'udemy'
-    _VALID_URL = r'https?://www\.udemy\.com/(?:[^#]+#/lecture/|lecture/view/?\?lectureId=)(?P<id>\d+)'
+    _VALID_URL = r'''(?x)
+                    https?://
+                        www\.udemy\.com/
+                        (?:
+                            [^#]+\#/lecture/|
+                            lecture/view/?\?lectureId=|
+                            [^/]+/learn/v4/t/lecture/
+                        )
+                        (?P<id>\d+)
+                    '''
     _LOGIN_URL = 'https://www.udemy.com/join/login-popup/?displayType=ajax&showSkipButton=1'
     _ORIGIN_URL = 'https://www.udemy.com'
     _NETRC_MACHINE = 'udemy'
@@ -33,36 +52,63 @@ class UdemyIE(InfoExtractor):
             'duration': 579.29,
         },
         'skip': 'Requires udemy account credentials',
+    }, {
+        # new URL schema
+        'url': 'https://www.udemy.com/electric-bass-right-from-the-start/learn/v4/t/lecture/4580906',
+        'only_matching': True,
+    }, {
+        # no url in outputs format entry
+        'url': 'https://www.udemy.com/learn-web-development-complete-step-by-step-guide-to-success/learn/v4/t/lecture/4125812',
+        'only_matching': True,
+    }, {
+        # only outputs rendition
+        'url': 'https://www.udemy.com/how-you-can-help-your-local-community-5-amazing-examples/learn/v4/t/lecture/3225750?start=0',
+        'only_matching': True,
     }]
 
-    def _enroll_course(self, webpage, course_id):
+    def _extract_course_info(self, webpage, video_id):
+        course = self._parse_json(
+            unescapeHTML(self._search_regex(
+                r'ng-init=["\'].*\bcourse=({.+?})[;"\']',
+                webpage, 'course', default='{}')),
+            video_id, fatal=False) or {}
+        course_id = course.get('id') or self._search_regex(
+            r'data-course-id=["\'](\d+)', webpage, 'course id')
+        return course_id, course.get('title')
+
+    def _enroll_course(self, base_url, webpage, course_id):
+        def combine_url(base_url, url):
+            return compat_urlparse.urljoin(base_url, url) if not url.startswith('http') else url
+
         checkout_url = unescapeHTML(self._search_regex(
-            r'href=(["\'])(?P<url>https?://(?:www\.)?udemy\.com/payment/checkout/.+?)\1',
+            r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/(?:payment|cart)/checkout/.+?)\1',
             webpage, 'checkout url', group='url', default=None))
         if checkout_url:
             raise ExtractorError(
                 'Course %s is not free. You have to pay for it before you can download. '
-                'Use this URL to confirm purchase: %s' % (course_id, checkout_url), expected=True)
+                'Use this URL to confirm purchase: %s'
+                % (course_id, combine_url(base_url, checkout_url)),
+                expected=True)
 
         enroll_url = unescapeHTML(self._search_regex(
-            r'href=(["\'])(?P<url>https?://(?:www\.)?udemy\.com/course/subscribe/.+?)\1',
+            r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/course/subscribe/.+?)\1',
             webpage, 'enroll url', group='url', default=None))
         if enroll_url:
-            webpage = self._download_webpage(enroll_url, course_id, 'Enrolling in the course')
+            webpage = self._download_webpage(
+                combine_url(base_url, enroll_url),
+                course_id, 'Enrolling in the course',
+                headers={'Referer': base_url})
             if '>You have enrolled in' in webpage:
                 self.to_screen('%s: Successfully enrolled in the course' % course_id)
 
     def _download_lecture(self, course_id, lecture_id):
         return self._download_json(
-            'https://www.udemy.com/api-2.0/users/me/subscribed-courses/%s/lectures/%s?%s' % (
-                course_id, lecture_id, compat_urllib_parse.urlencode({
-                    'video_only': '',
-                    'auto_play': '',
-                    'fields[lecture]': 'title,description,asset',
-                    'fields[asset]': 'asset_type,stream_url,thumbnail_url,download_urls,data',
-                    'instructorPreviewMode': 'False',
-                })),
-            lecture_id, 'Downloading lecture JSON')
+            'https://www.udemy.com/api-2.0/users/me/subscribed-courses/%s/lectures/%s?'
+            % (course_id, lecture_id),
+            lecture_id, 'Downloading lecture JSON', query={
+                'fields[lecture]': 'title,description,view_html,asset',
+                'fields[asset]': 'asset_type,stream_url,thumbnail_url,download_urls,stream_urls,captions,data',
+            })
 
     def _handle_error(self, response):
         if not isinstance(response, dict):
@@ -75,7 +121,14 @@ class UdemyIE(InfoExtractor):
                 error_str += ' - %s' % error_data.get('formErrors')
             raise ExtractorError(error_str, expected=True)
 
-    def _download_json(self, url_or_request, video_id, note='Downloading JSON metadata'):
+    def _download_webpage_handle(self, *args, **kwargs):
+        headers = kwargs.get('headers', {}).copy()
+        headers['User-Agent'] = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_5) AppleWebKit/603.2.4 (KHTML, like Gecko) Version/10.1.1 Safari/603.2.4'
+        kwargs['headers'] = headers
+        return super(UdemyIE, self)._download_webpage_handle(
+            *args, **compat_kwargs(kwargs))
+
+    def _download_json(self, url_or_request, *args, **kwargs):
         headers = {
             'X-Udemy-Snail-Case': 'true',
             'X-Requested-With': 'XMLHttpRequest',
@@ -93,7 +146,7 @@ class UdemyIE(InfoExtractor):
         else:
             url_or_request = sanitized_Request(url_or_request, headers=headers)
 
-        response = super(UdemyIE, self)._download_json(url_or_request, video_id, note)
+        response = super(UdemyIE, self)._download_json(url_or_request, *args, **kwargs)
         self._handle_error(response)
         return response
 
@@ -101,7 +154,7 @@ class UdemyIE(InfoExtractor):
         self._login()
 
     def _login(self):
-        (username, password) = self._get_login_info()
+        username, password = self._get_login_info()
         if username is None:
             return
 
@@ -109,7 +162,9 @@ class UdemyIE(InfoExtractor):
             self._LOGIN_URL, None, 'Downloading login popup')
 
         def is_logged(webpage):
-            return any(p in webpage for p in ['href="https://www.udemy.com/user/logout/', '>Logout<'])
+            return any(re.search(p, webpage) for p in (
+                r'href=["\'](?:https://www\.udemy\.com)?/user/logout/',
+                r'>Logout<'))
 
         # already logged in
         if is_logged(login_popup):
@@ -118,17 +173,17 @@ class UdemyIE(InfoExtractor):
         login_form = self._form_hidden_inputs('login-form', login_popup)
 
         login_form.update({
-            'email': username.encode('utf-8'),
-            'password': password.encode('utf-8'),
+            'email': username,
+            'password': password,
         })
 
-        request = sanitized_Request(
-            self._LOGIN_URL, compat_urllib_parse.urlencode(login_form).encode('utf-8'))
-        request.add_header('Referer', self._ORIGIN_URL)
-        request.add_header('Origin', self._ORIGIN_URL)
-
         response = self._download_webpage(
-            request, None, 'Logging in as %s' % username)
+            self._LOGIN_URL, None, 'Logging in',
+            data=urlencode_postdata(login_form),
+            headers={
+                'Referer': self._ORIGIN_URL,
+                'Origin': self._ORIGIN_URL,
+            })
 
         if not is_logged(response):
             error = self._html_search_regex(
@@ -143,15 +198,14 @@ class UdemyIE(InfoExtractor):
 
         webpage = self._download_webpage(url, lecture_id)
 
-        course_id = self._search_regex(
-            r'data-course-id=["\'](\d+)', webpage, 'course id')
+        course_id, _ = self._extract_course_info(webpage, lecture_id)
 
         try:
             lecture = self._download_lecture(course_id, lecture_id)
         except ExtractorError as e:
             # Error could possibly mean we are not enrolled in the course
             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
-                self._enroll_course(webpage, course_id)
+                self._enroll_course(url, webpage, course_id)
                 lecture = self._download_lecture(course_id, lecture_id)
             else:
                 raise
@@ -161,56 +215,179 @@ class UdemyIE(InfoExtractor):
 
         asset = lecture['asset']
 
-        asset_type = asset.get('assetType') or asset.get('asset_type')
+        asset_type = asset.get('asset_type') or asset.get('assetType')
         if asset_type != 'Video':
             raise ExtractorError(
                 'Lecture %s is not a video' % lecture_id, expected=True)
 
-        stream_url = asset.get('streamUrl') or asset.get('stream_url')
+        stream_url = asset.get('stream_url') or asset.get('streamUrl')
         if stream_url:
             youtube_url = self._search_regex(
                 r'(https?://www\.youtube\.com/watch\?v=.*)', stream_url, 'youtube URL', default=None)
             if youtube_url:
                 return self.url_result(youtube_url, 'Youtube')
 
-        video_id = asset['id']
-        thumbnail = asset.get('thumbnailUrl') or asset.get('thumbnail_url')
+        video_id = compat_str(asset['id'])
+        thumbnail = asset.get('thumbnail_url') or asset.get('thumbnailUrl')
         duration = float_or_none(asset.get('data', {}).get('duration'))
-        outputs = asset.get('data', {}).get('outputs', {})
+
+        subtitles = {}
+        automatic_captions = {}
 
         formats = []
-        for format_ in asset.get('download_urls', {}).get('Video', []):
-            video_url = format_.get('file')
-            if not video_url:
-                continue
-            format_id = format_.get('label')
-            f = {
-                'url': format_['file'],
-                'height': int_or_none(format_id),
+
+        def extract_output_format(src, f_id):
+            return {
+                'url': src.get('url'),
+                'format_id': '%sp' % (src.get('height') or f_id),
+                'width': int_or_none(src.get('width')),
+                'height': int_or_none(src.get('height')),
+                'vbr': int_or_none(src.get('video_bitrate_in_kbps')),
+                'vcodec': src.get('video_codec'),
+                'fps': int_or_none(src.get('frame_rate')),
+                'abr': int_or_none(src.get('audio_bitrate_in_kbps')),
+                'acodec': src.get('audio_codec'),
+                'asr': int_or_none(src.get('audio_sample_rate')),
+                'tbr': int_or_none(src.get('total_bitrate_in_kbps')),
+                'filesize': int_or_none(src.get('file_size_in_bytes')),
             }
-            if format_id:
-                # Some videos contain additional metadata (e.g.
-                # https://www.udemy.com/ios9-swift/learn/#/lecture/3383208)
-                output = outputs.get(format_id)
-                if isinstance(output, dict):
-                    f.update({
-                        'format_id': '%sp' % (output.get('label') or format_id),
-                        'width': int_or_none(output.get('width')),
-                        'height': int_or_none(output.get('height')),
-                        'vbr': int_or_none(output.get('video_bitrate_in_kbps')),
-                        'vcodec': output.get('video_codec'),
-                        'fps': int_or_none(output.get('frame_rate')),
-                        'abr': int_or_none(output.get('audio_bitrate_in_kbps')),
-                        'acodec': output.get('audio_codec'),
-                        'asr': int_or_none(output.get('audio_sample_rate')),
-                        'tbr': int_or_none(output.get('total_bitrate_in_kbps')),
-                        'filesize': int_or_none(output.get('file_size_in_bytes')),
-                    })
-                else:
-                    f['format_id'] = '%sp' % format_id
-            formats.append(f)
 
-        self._sort_formats(formats)
+        outputs = asset.get('data', {}).get('outputs')
+        if not isinstance(outputs, dict):
+            outputs = {}
+
+        def add_output_format_meta(f, key):
+            output = outputs.get(key)
+            if isinstance(output, dict):
+                output_format = extract_output_format(output, key)
+                output_format.update(f)
+                return output_format
+            return f
+
+        def extract_formats(source_list):
+            if not isinstance(source_list, list):
+                return
+            for source in source_list:
+                video_url = url_or_none(source.get('file') or source.get('src'))
+                if not video_url:
+                    continue
+                if source.get('type') == 'application/x-mpegURL' or determine_ext(video_url) == 'm3u8':
+                    formats.extend(self._extract_m3u8_formats(
+                        video_url, video_id, 'mp4', entry_protocol='m3u8_native',
+                        m3u8_id='hls', fatal=False))
+                    continue
+                format_id = source.get('label')
+                f = {
+                    'url': video_url,
+                    'format_id': '%sp' % format_id,
+                    'height': int_or_none(format_id),
+                }
+                if format_id:
+                    # Some videos contain additional metadata (e.g.
+                    # https://www.udemy.com/ios9-swift/learn/#/lecture/3383208)
+                    f = add_output_format_meta(f, format_id)
+                formats.append(f)
+
+        def extract_subtitles(track_list):
+            if not isinstance(track_list, list):
+                return
+            for track in track_list:
+                if not isinstance(track, dict):
+                    continue
+                if track.get('kind') != 'captions':
+                    continue
+                src = url_or_none(track.get('src'))
+                if not src:
+                    continue
+                lang = track.get('language') or track.get(
+                    'srclang') or track.get('label')
+                sub_dict = automatic_captions if track.get(
+                    'autogenerated') is True else subtitles
+                sub_dict.setdefault(lang, []).append({
+                    'url': src,
+                })
+
+        for url_kind in ('download', 'stream'):
+            urls = asset.get('%s_urls' % url_kind)
+            if isinstance(urls, dict):
+                extract_formats(urls.get('Video'))
+
+        captions = asset.get('captions')
+        if isinstance(captions, list):
+            for cc in captions:
+                if not isinstance(cc, dict):
+                    continue
+                cc_url = url_or_none(cc.get('url'))
+                if not cc_url:
+                    continue
+                lang = try_get(cc, lambda x: x['locale']['locale'], compat_str)
+                sub_dict = (automatic_captions if cc.get('source') == 'auto'
+                            else subtitles)
+                sub_dict.setdefault(lang or 'en', []).append({
+                    'url': cc_url,
+                })
+
+        view_html = lecture.get('view_html')
+        if view_html:
+            view_html_urls = set()
+            for source in re.findall(r'<source[^>]+>', view_html):
+                attributes = extract_attributes(source)
+                src = attributes.get('src')
+                if not src:
+                    continue
+                res = attributes.get('data-res')
+                height = int_or_none(res)
+                if src in view_html_urls:
+                    continue
+                view_html_urls.add(src)
+                if attributes.get('type') == 'application/x-mpegURL' or determine_ext(src) == 'm3u8':
+                    m3u8_formats = self._extract_m3u8_formats(
+                        src, video_id, 'mp4', entry_protocol='m3u8_native',
+                        m3u8_id='hls', fatal=False)
+                    for f in m3u8_formats:
+                        m = re.search(r'/hls_(?P<height>\d{3,4})_(?P<tbr>\d{2,})/', f['url'])
+                        if m:
+                            if not f.get('height'):
+                                f['height'] = int(m.group('height'))
+                            if not f.get('tbr'):
+                                f['tbr'] = int(m.group('tbr'))
+                    formats.extend(m3u8_formats)
+                else:
+                    formats.append(add_output_format_meta({
+                        'url': src,
+                        'format_id': '%dp' % height if height else None,
+                        'height': height,
+                    }, res))
+
+            # react rendition since 2017.04.15 (see
+            # https://github.com/rg3/youtube-dl/issues/12744)
+            data = self._parse_json(
+                self._search_regex(
+                    r'videojs-setup-data=(["\'])(?P<data>{.+?})\1', view_html,
+                    'setup data', default='{}', group='data'), video_id,
+                transform_source=unescapeHTML, fatal=False)
+            if data and isinstance(data, dict):
+                extract_formats(data.get('sources'))
+                if not duration:
+                    duration = int_or_none(data.get('duration'))
+                extract_subtitles(data.get('tracks'))
+
+            if not subtitles and not automatic_captions:
+                text_tracks = self._parse_json(
+                    self._search_regex(
+                        r'text-tracks=(["\'])(?P<data>\[.+?\])\1', view_html,
+                        'text tracks', default='{}', group='data'), video_id,
+                    transform_source=lambda s: js_to_json(unescapeHTML(s)),
+                    fatal=False)
+                extract_subtitles(text_tracks)
+
+        if not formats and outputs:
+            for format_id, output in outputs.items():
+                f = extract_output_format(output, format_id)
+                if f.get('url'):
+                    formats.append(f)
+
+        self._sort_formats(formats, field_preference=('height', 'width', 'tbr', 'format_id'))
 
         return {
             'id': video_id,
@@ -218,13 +395,15 @@ class UdemyIE(InfoExtractor):
             'description': description,
             'thumbnail': thumbnail,
             'duration': duration,
-            'formats': formats
+            'formats': formats,
+            'subtitles': subtitles,
+            'automatic_captions': automatic_captions,
         }
 
 
 class UdemyCourseIE(UdemyIE):
     IE_NAME = 'udemy:course'
-    _VALID_URL = r'https?://www\.udemy\.com/(?P<id>[\da-z-]+)'
+    _VALID_URL = r'https?://(?:www\.)?udemy\.com/(?P<id>[^/?#&]+)'
     _TESTS = []
 
     @classmethod
@@ -236,29 +415,34 @@ class UdemyCourseIE(UdemyIE):
 
         webpage = self._download_webpage(url, course_path)
 
-        response = self._download_json(
-            'https://www.udemy.com/api-1.1/courses/%s' % course_path,
-            course_path, 'Downloading course JSON')
-
-        course_id = response['id']
-        course_title = response.get('title')
+        course_id, title = self._extract_course_info(webpage, course_path)
 
-        self._enroll_course(webpage, course_id)
+        self._enroll_course(url, webpage, course_id)
 
         response = self._download_json(
-            'https://www.udemy.com/api-1.1/courses/%s/curriculum' % course_id,
-            course_id, 'Downloading course curriculum')
+            'https://www.udemy.com/api-2.0/courses/%s/cached-subscriber-curriculum-items' % course_id,
+            course_id, 'Downloading course curriculum', query={
+                'fields[chapter]': 'title,object_index',
+                'fields[lecture]': 'title,asset',
+                'page_size': '1000',
+            })
 
         entries = []
-        chapter, chapter_number = None, None
-        for asset in response:
-            asset_type = asset.get('assetType') or asset.get('asset_type')
-            if asset_type == 'Video':
-                asset_id = asset.get('id')
-                if asset_id:
+        chapter, chapter_number = [None] * 2
+        for entry in response['results']:
+            clazz = entry.get('_class')
+            if clazz == 'lecture':
+                asset = entry.get('asset')
+                if isinstance(asset, dict):
+                    asset_type = asset.get('asset_type') or asset.get('assetType')
+                    if asset_type != 'Video':
+                        continue
+                lecture_id = entry.get('id')
+                if lecture_id:
                     entry = {
                         '_type': 'url_transparent',
-                        'url': 'https://www.udemy.com/%s/#/lecture/%s' % (course_path, asset['id']),
+                        'url': 'https://www.udemy.com/%s/learn/v4/t/lecture/%s' % (course_path, entry['id']),
+                        'title': entry.get('title'),
                         'ie_key': UdemyIE.ie_key(),
                     }
                     if chapter_number:
@@ -266,8 +450,8 @@ class UdemyCourseIE(UdemyIE):
                     if chapter:
                         entry['chapter'] = chapter
                     entries.append(entry)
-            elif asset.get('type') == 'chapter':
-                chapter_number = asset.get('index') or asset.get('object_index')
-                chapter = asset.get('title')
+            elif clazz == 'chapter':
+                chapter_number = entry.get('object_index')
+                chapter = entry.get('title')
 
-        return self.playlist_result(entries, course_id, course_title)
+        return self.playlist_result(entries, course_id, title)