]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/udemy.py
Merge tag 'upstream/2016.12.01'
[youtubedl] / youtube_dl / extractor / udemy.py
index 054f427252341306edd7698e6d150bcb619b64fc..cce29c6e07134c58425f5486b15ce6d9e017d0d7 100644 (file)
@@ -3,17 +3,38 @@ from __future__ import unicode_literals
 import re
 
 from .common import InfoExtractor
-from ..utils import (
-    compat_urllib_parse,
+from ..compat import (
+    compat_HTTPError,
+    compat_str,
     compat_urllib_request,
+    compat_urlparse,
+)
+from ..utils import (
+    determine_ext,
+    extract_attributes,
     ExtractorError,
+    float_or_none,
+    int_or_none,
+    sanitized_Request,
+    unescapeHTML,
+    urlencode_postdata,
 )
 
 
 class UdemyIE(InfoExtractor):
     IE_NAME = 'udemy'
-    _VALID_URL = r'https?://www\.udemy\.com/(?:[^#]+#/lecture/|lecture/view/?\?lectureId=)(?P<id>\d+)'
-    _LOGIN_URL = 'https://www.udemy.com/join/login-submit/'
+    _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'
 
     _TESTS = [{
@@ -27,8 +48,56 @@ 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,
     }]
 
+    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'&quot;id&quot;\s*:\s*(\d+)', 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',
+            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, 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',
+            webpage, 'enroll url', group='url', default=None))
+        if enroll_url:
+            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?'
+            % (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,data',
+            })
+
     def _handle_error(self, response):
         if not isinstance(response, dict):
             return
@@ -40,8 +109,25 @@ class UdemyIE(InfoExtractor):
                 error_str += ' - %s' % error_data.get('formErrors')
             raise ExtractorError(error_str, expected=True)
 
-    def _download_json(self, url, video_id, note='Downloading JSON metadata'):
-        response = super(UdemyIE, self)._download_json(url, video_id, note)
+    def _download_json(self, url_or_request, *args, **kwargs):
+        headers = {
+            'X-Udemy-Snail-Case': 'true',
+            'X-Requested-With': 'XMLHttpRequest',
+        }
+        for cookie in self._downloader.cookiejar:
+            if cookie.name == 'client_id':
+                headers['X-Udemy-Client-Id'] = cookie.value
+            elif cookie.name == 'access_token':
+                headers['X-Udemy-Bearer-Token'] = cookie.value
+                headers['X-Udemy-Authorization'] = 'Bearer %s' % cookie.value
+
+        if isinstance(url_or_request, compat_urllib_request.Request):
+            for header, value in headers.items():
+                url_or_request.add_header(header, value)
+        else:
+            url_or_request = sanitized_Request(url_or_request, headers=headers)
+
+        response = super(UdemyIE, self)._download_json(url_or_request, *args, **kwargs)
         self._handle_error(response)
         return response
 
@@ -51,68 +137,164 @@ class UdemyIE(InfoExtractor):
     def _login(self):
         (username, password) = self._get_login_info()
         if username is None:
-            raise ExtractorError(
-                'Udemy account is required, use --username and --password options to provide account credentials.',
-                expected=True)
+            return
 
         login_popup = self._download_webpage(
-            'https://www.udemy.com/join/login-popup?displayType=ajax&showSkipButton=1', None,
-            'Downloading login popup')
+            self._LOGIN_URL, None, 'Downloading login popup')
+
+        def is_logged(webpage):
+            return any(re.search(p, webpage) for p in (
+                r'href=["\'](?:https://www\.udemy\.com)?/user/logout/',
+                r'>Logout<'))
 
-        if login_popup == '<div class="run-command close-popup redirect" data-url="https://www.udemy.com/"></div>':
+        # already logged in
+        if is_logged(login_popup):
             return
 
-        csrf = self._html_search_regex(r'<input type="hidden" name="csrf" value="(.+?)"', login_popup, 'csrf token')
+        login_form = self._form_hidden_inputs('login-form', login_popup)
 
-        login_form = {
+        login_form.update({
             'email': username,
             'password': password,
-            'csrf': csrf,
-            'displayType': 'json',
-            'isSubmitted': '1',
-        }
-        request = compat_urllib_request.Request(self._LOGIN_URL, compat_urllib_parse.urlencode(login_form))
-        response = self._download_json(request, None, 'Logging in as %s' % username)
-
-        if 'returnUrl' not in response:
+        })
+
+        response = self._download_webpage(
+            self._LOGIN_URL, None, 'Logging in as %s' % username,
+            data=urlencode_postdata(login_form),
+            headers={
+                'Referer': self._ORIGIN_URL,
+                'Origin': self._ORIGIN_URL,
+            })
+
+        if not is_logged(response):
+            error = self._html_search_regex(
+                r'(?s)<div[^>]+class="form-errors[^"]*">(.+?)</div>',
+                response, 'error message', default=None)
+            if error:
+                raise ExtractorError('Unable to login: %s' % error, expected=True)
             raise ExtractorError('Unable to log in')
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        lecture_id = mobj.group('id')
-
-        lecture = self._download_json(
-            'https://www.udemy.com/api-1.1/lectures/%s' % lecture_id, lecture_id, 'Downloading lecture JSON')
+        lecture_id = self._match_id(url)
 
-        if lecture['assetType'] != 'Video':
-            raise ExtractorError('Lecture %s is not a video' % lecture_id, expected=True)
-
-        asset = lecture['asset']
+        webpage = self._download_webpage(url, lecture_id)
 
-        stream_url = asset['streamUrl']
-        mobj = re.search(r'(https?://www\.youtube\.com/watch\?v=.*)', stream_url)
-        if mobj:
-            return self.url_result(mobj.group(1), 'Youtube')
+        course_id, _ = self._extract_course_info(webpage, lecture_id)
 
-        video_id = asset['id']
-        thumbnail = asset['thumbnailUrl']
-        duration = asset['data']['duration']
+        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(url, webpage, course_id)
+                lecture = self._download_lecture(course_id, lecture_id)
+            else:
+                raise
 
-        download_url = asset['downloadUrl']
+        title = lecture['title']
+        description = lecture.get('description')
 
-        formats = [
-            {
-                'url': download_url['Video480p'][0],
-                'format_id': '360p',
-            },
-            {
-                'url': download_url['Video'][0],
-                'format_id': '720p',
-            },
-        ]
+        asset = lecture['asset']
 
-        title = lecture['title']
-        description = lecture['description']
+        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('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 = compat_str(asset['id'])
+        thumbnail = asset.get('thumbnail_url') or asset.get('thumbnailUrl')
+        duration = float_or_none(asset.get('data', {}).get('duration'))
+
+        formats = []
+
+        def extract_output_format(src):
+            return {
+                'url': src['url'],
+                'format_id': '%sp' % (src.get('height') or format_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')),
+            }
+
+        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)
+                output_format.update(f)
+                return output_format
+            return f
+
+        download_urls = asset.get('download_urls')
+        if isinstance(download_urls, dict):
+            video = download_urls.get('Video')
+            if isinstance(video, list):
+                for format_ in video:
+                    video_url = format_.get('file')
+                    if not video_url:
+                        continue
+                    format_id = format_.get('label')
+                    f = {
+                        'url': format_['file'],
+                        '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)
+
+        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))
+
+        self._sort_formats(formats, field_preference=('height', 'width', 'tbr', 'format_id'))
 
         return {
             'id': video_id,
@@ -126,9 +308,7 @@ class UdemyIE(InfoExtractor):
 
 class UdemyCourseIE(UdemyIE):
     IE_NAME = 'udemy:course'
-    _VALID_URL = r'https?://www\.udemy\.com/(?P<coursepath>[\da-z-]+)'
-    _SUCCESSFULLY_ENROLLED = '>You have enrolled in this course!<'
-    _ALREADY_ENROLLED = '>You are already taking this course.<'
+    _VALID_URL = r'https?://(?:www\.)?udemy\.com/(?P<id>[^/?#&]+)'
     _TESTS = []
 
     @classmethod
@@ -136,29 +316,47 @@ class UdemyCourseIE(UdemyIE):
         return False if UdemyIE.suitable(url) else super(UdemyCourseIE, cls).suitable(url)
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        course_path = mobj.group('coursepath')
-
-        response = self._download_json(
-            'https://www.udemy.com/api-1.1/courses/%s' % course_path, course_path, 'Downloading course JSON')
+        course_path = self._match_id(url)
 
-        course_id = int(response['id'])
-        course_title = response['title']
+        webpage = self._download_webpage(url, course_path)
 
-        webpage = self._download_webpage(
-            'https://www.udemy.com/course/subscribe/?courseId=%s' % course_id, course_id, 'Enrolling in the course')
+        course_id, title = self._extract_course_info(webpage, course_path)
 
-        if self._SUCCESSFULLY_ENROLLED in webpage:
-            self.to_screen('%s: Successfully enrolled in' % course_id)
-        elif self._ALREADY_ENROLLED in webpage:
-            self.to_screen('%s: Already enrolled in' % 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')
-
-        entries = [
-            self.url_result('https://www.udemy.com/%s/#/lecture/%s' % (course_path, asset['id']), 'Udemy')
-            for asset in response if asset.get('assetType') == 'Video'
-        ]
-
-        return self.playlist_result(entries, course_id, course_title)
\ No newline at end of file
+        response = self._download_json(
+            '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] * 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/learn/v4/t/lecture/%s' % (course_path, entry['id']),
+                        'title': entry.get('title'),
+                        'ie_key': UdemyIE.ie_key(),
+                    }
+                    if chapter_number:
+                        entry['chapter_number'] = chapter_number
+                    if chapter:
+                        entry['chapter'] = chapter
+                    entries.append(entry)
+            elif clazz == 'chapter':
+                chapter_number = entry.get('object_index')
+                chapter = entry.get('title')
+
+        return self.playlist_result(entries, course_id, title)