mirror of
				https://github.com/yt-dlp/yt-dlp.git
				synced 2025-11-04 08:30:46 +00:00 
			
		
		
		
	Using https://github.com/asottile/pyupgrade 1. `__future__` imports and `coding: utf-8` were removed 2. Files were rewritten with `pyupgrade --py36-plus --keep-percent-format` 3. f-strings were cherry-picked from `pyupgrade --py36-plus` Extractors are left untouched (except removing header) to avoid unnecessary merge conflicts
		
			
				
	
	
		
			178 lines
		
	
	
		
			6.0 KiB
		
	
	
	
		
			Python
		
	
	
	
	
	
			
		
		
	
	
			178 lines
		
	
	
		
			6.0 KiB
		
	
	
	
		
			Python
		
	
	
	
	
	
import re
 | 
						|
 | 
						|
from .common import InfoExtractor
 | 
						|
from .vimeo import VimeoIE
 | 
						|
from ..compat import compat_str
 | 
						|
from ..utils import (
 | 
						|
    ExtractorError,
 | 
						|
    int_or_none,
 | 
						|
    merge_dicts,
 | 
						|
    try_get,
 | 
						|
    unescapeHTML,
 | 
						|
    unified_timestamp,
 | 
						|
    urljoin,
 | 
						|
)
 | 
						|
 | 
						|
 | 
						|
class RayWenderlichIE(InfoExtractor):
 | 
						|
    _VALID_URL = r'''(?x)
 | 
						|
                    https?://
 | 
						|
                        (?:
 | 
						|
                            videos\.raywenderlich\.com/courses|
 | 
						|
                            (?:www\.)?raywenderlich\.com
 | 
						|
                        )/
 | 
						|
                        (?P<course_id>[^/]+)/lessons/(?P<id>\d+)
 | 
						|
                    '''
 | 
						|
 | 
						|
    _TESTS = [{
 | 
						|
        'url': 'https://www.raywenderlich.com/3530-testing-in-ios/lessons/1',
 | 
						|
        'info_dict': {
 | 
						|
            'id': '248377018',
 | 
						|
            'ext': 'mp4',
 | 
						|
            'title': 'Introduction',
 | 
						|
            'description': 'md5:804d031b3efa9fcb49777d512d74f722',
 | 
						|
            'timestamp': 1513906277,
 | 
						|
            'upload_date': '20171222',
 | 
						|
            'duration': 133,
 | 
						|
            'uploader': 'Ray Wenderlich',
 | 
						|
            'uploader_id': 'user3304672',
 | 
						|
        },
 | 
						|
        'params': {
 | 
						|
            'noplaylist': True,
 | 
						|
            'skip_download': True,
 | 
						|
        },
 | 
						|
        'add_ie': [VimeoIE.ie_key()],
 | 
						|
        'expected_warnings': ['HTTP Error 403: Forbidden'],
 | 
						|
    }, {
 | 
						|
        'url': 'https://videos.raywenderlich.com/courses/105-testing-in-ios/lessons/1',
 | 
						|
        'only_matching': True,
 | 
						|
    }]
 | 
						|
 | 
						|
    @staticmethod
 | 
						|
    def _extract_video_id(data, lesson_id):
 | 
						|
        if not data:
 | 
						|
            return
 | 
						|
        groups = try_get(data, lambda x: x['groups'], list) or []
 | 
						|
        if not groups:
 | 
						|
            return
 | 
						|
        for group in groups:
 | 
						|
            if not isinstance(group, dict):
 | 
						|
                continue
 | 
						|
            contents = try_get(data, lambda x: x['contents'], list) or []
 | 
						|
            for content in contents:
 | 
						|
                if not isinstance(content, dict):
 | 
						|
                    continue
 | 
						|
                ordinal = int_or_none(content.get('ordinal'))
 | 
						|
                if ordinal != lesson_id:
 | 
						|
                    continue
 | 
						|
                video_id = content.get('identifier')
 | 
						|
                if video_id:
 | 
						|
                    return compat_str(video_id)
 | 
						|
 | 
						|
    def _real_extract(self, url):
 | 
						|
        mobj = self._match_valid_url(url)
 | 
						|
        course_id, lesson_id = mobj.group('course_id', 'id')
 | 
						|
        display_id = '%s/%s' % (course_id, lesson_id)
 | 
						|
 | 
						|
        webpage = self._download_webpage(url, display_id)
 | 
						|
 | 
						|
        thumbnail = self._og_search_thumbnail(
 | 
						|
            webpage, default=None) or self._html_search_meta(
 | 
						|
            'twitter:image', webpage, 'thumbnail')
 | 
						|
 | 
						|
        if '>Subscribe to unlock' in webpage:
 | 
						|
            raise ExtractorError(
 | 
						|
                'This content is only available for subscribers',
 | 
						|
                expected=True)
 | 
						|
 | 
						|
        info = {
 | 
						|
            'thumbnail': thumbnail,
 | 
						|
        }
 | 
						|
 | 
						|
        vimeo_id = self._search_regex(
 | 
						|
            r'data-vimeo-id=["\'](\d+)', webpage, 'vimeo id', default=None)
 | 
						|
 | 
						|
        if not vimeo_id:
 | 
						|
            data = self._parse_json(
 | 
						|
                self._search_regex(
 | 
						|
                    r'data-collection=(["\'])(?P<data>{.+?})\1', webpage,
 | 
						|
                    'data collection', default='{}', group='data'),
 | 
						|
                display_id, transform_source=unescapeHTML, fatal=False)
 | 
						|
            video_id = self._extract_video_id(
 | 
						|
                data, lesson_id) or self._search_regex(
 | 
						|
                r'/videos/(\d+)/', thumbnail, 'video id')
 | 
						|
            headers = {
 | 
						|
                'Referer': url,
 | 
						|
                'X-Requested-With': 'XMLHttpRequest',
 | 
						|
            }
 | 
						|
            csrf_token = self._html_search_meta(
 | 
						|
                'csrf-token', webpage, 'csrf token', default=None)
 | 
						|
            if csrf_token:
 | 
						|
                headers['X-CSRF-Token'] = csrf_token
 | 
						|
            video = self._download_json(
 | 
						|
                'https://videos.raywenderlich.com/api/v1/videos/%s.json'
 | 
						|
                % video_id, display_id, headers=headers)['video']
 | 
						|
            vimeo_id = video['clips'][0]['provider_id']
 | 
						|
            info.update({
 | 
						|
                '_type': 'url_transparent',
 | 
						|
                'title': video.get('name'),
 | 
						|
                'description': video.get('description') or video.get(
 | 
						|
                    'meta_description'),
 | 
						|
                'duration': int_or_none(video.get('duration')),
 | 
						|
                'timestamp': unified_timestamp(video.get('created_at')),
 | 
						|
            })
 | 
						|
 | 
						|
        return merge_dicts(info, self.url_result(
 | 
						|
            VimeoIE._smuggle_referrer(
 | 
						|
                'https://player.vimeo.com/video/%s' % vimeo_id, url),
 | 
						|
            ie=VimeoIE.ie_key(), video_id=vimeo_id))
 | 
						|
 | 
						|
 | 
						|
class RayWenderlichCourseIE(InfoExtractor):
 | 
						|
    _VALID_URL = r'''(?x)
 | 
						|
                    https?://
 | 
						|
                        (?:
 | 
						|
                            videos\.raywenderlich\.com/courses|
 | 
						|
                            (?:www\.)?raywenderlich\.com
 | 
						|
                        )/
 | 
						|
                        (?P<id>[^/]+)
 | 
						|
                    '''
 | 
						|
 | 
						|
    _TEST = {
 | 
						|
        'url': 'https://www.raywenderlich.com/3530-testing-in-ios',
 | 
						|
        'info_dict': {
 | 
						|
            'title': 'Testing in iOS',
 | 
						|
            'id': '3530-testing-in-ios',
 | 
						|
        },
 | 
						|
        'params': {
 | 
						|
            'noplaylist': False,
 | 
						|
        },
 | 
						|
        'playlist_count': 29,
 | 
						|
    }
 | 
						|
 | 
						|
    @classmethod
 | 
						|
    def suitable(cls, url):
 | 
						|
        return False if RayWenderlichIE.suitable(url) else super(
 | 
						|
            RayWenderlichCourseIE, cls).suitable(url)
 | 
						|
 | 
						|
    def _real_extract(self, url):
 | 
						|
        course_id = self._match_id(url)
 | 
						|
 | 
						|
        webpage = self._download_webpage(url, course_id)
 | 
						|
 | 
						|
        entries = []
 | 
						|
        lesson_urls = set()
 | 
						|
        for lesson_url in re.findall(
 | 
						|
                r'<a[^>]+\bhref=["\'](/%s/lessons/\d+)' % course_id, webpage):
 | 
						|
            if lesson_url in lesson_urls:
 | 
						|
                continue
 | 
						|
            lesson_urls.add(lesson_url)
 | 
						|
            entries.append(self.url_result(
 | 
						|
                urljoin(url, lesson_url), ie=RayWenderlichIE.ie_key()))
 | 
						|
 | 
						|
        title = self._og_search_title(
 | 
						|
            webpage, default=None) or self._html_search_meta(
 | 
						|
            'twitter:title', webpage, 'title', default=None)
 | 
						|
 | 
						|
        return self.playlist_result(entries, course_id, title)
 |