yt_dlp/extractor/substack.py

   1 import re
   2 import urllib.parse
   3
   4 from .common import InfoExtractor
   5 from ..networking import HEADRequest
   6 from ..utils import (
   7     determine_ext,
   8     js_to_json,
   9     str_or_none,
  10 )
  11 from ..utils.traversal import traverse_obj
  12
  13
  14 class SubstackIE(InfoExtractor):
  15     _VALID_URL = r'https?://(?P<username>[\w-]+)\.substack\.com/p/(?P<id>[\w-]+)'
  16     _TESTS = [{
  17         'url': 'https://haleynahman.substack.com/p/i-made-a-vlog?s=r',
  18         'md5': 'f27e4fc6252001d48d479f45e65cdfd5',
  19         'info_dict': {
  20             'id': '47660949',
  21             'ext': 'mp4',
  22             'title': 'I MADE A VLOG',
  23             'description': 'md5:9248af9a759321e1027226f988f54d96',
  24             'thumbnail': 'md5:bec758a34d8ee9142d43bcebdf33af18',
  25             'uploader': 'Maybe Baby',
  26             'uploader_id': '33628',
  27         },
  28     }, {
  29         'url': 'https://haleynahman.substack.com/p/-dear-danny-i-found-my-boyfriends?s=r',
  30         'md5': '0a63eacec877a1171a62cfa69710fcea',
  31         'info_dict': {
  32             'id': '51045592',
  33             'ext': 'mpga',
  34             'title': "🎧 Dear Danny: I found my boyfriend's secret Twitter account",
  35             'description': 'md5:a57f2439319e56e0af92dd0c95d75797',
  36             'thumbnail': 'md5:daa40b6b79249417c14ff8103db29639',
  37             'uploader': 'Maybe Baby',
  38             'uploader_id': '33628',
  39         },
  40     }, {
  41         'url': 'https://andrewzimmern.substack.com/p/mussels-with-black-bean-sauce-recipe',
  42         'md5': 'fd3c07077b02444ff0130715b5f632bb',
  43         'info_dict': {
  44             'id': '47368578',
  45             'ext': 'mp4',
  46             'title': 'Mussels with Black Bean Sauce: Recipe of the Week #7',
  47             'description': 'md5:b96234a2906c7d854d5229818d889515',
  48             'thumbnail': 'md5:e30bfaa9da40e82aa62354263a9dd232',
  49             'uploader': "Andrew Zimmern's Spilled Milk ",
  50             'uploader_id': '577659',
  51         },
  52     }, {
  53         # Podcast that needs its file extension resolved to mp3
  54         'url': 'https://persuasion1.substack.com/p/summers',
  55         'md5': '1456a755d46084744facdfac9edf900f',
  56         'info_dict': {
  57             'id': '141970405',
  58             'ext': 'mp3',
  59             'title': 'Larry Summers on What Went Wrong on Campus',
  60             'description': 'Yascha Mounk and Larry Summers also discuss the promise and perils of artificial intelligence.',
  61             'thumbnail': r're:https://substackcdn\.com/image/.+\.jpeg',
  62             'uploader': 'Persuasion',
  63             'uploader_id': '61579',
  64         },
  65     }]
  66
  67     @classmethod
  68     def _extract_embed_urls(cls, url, webpage):
  69         if not re.search(r'<script[^>]+src=["\']https://substackcdn.com/[^"\']+\.js', webpage):
  70             return
  71
  72         mobj = re.search(r'{[^}]*\\?["\']subdomain\\?["\']\s*:\s*\\?["\'](?P<subdomain>[^\\"\']+)', webpage)
  73         if mobj:
  74             parsed = urllib.parse.urlparse(url)
  75             yield parsed._replace(netloc=f'{mobj.group("subdomain")}.substack.com').geturl()
  76             raise cls.StopExtraction
  77
  78     def _extract_video_formats(self, video_id, url):
  79         formats, subtitles = [], {}
  80         for video_format in ('hls', 'mp4'):
  81             video_url = urllib.parse.urljoin(url, f'/api/v1/video/upload/{video_id}/src?type={video_format}')
  82
  83             if video_format == 'hls':
  84                 fmts, subs = self._extract_m3u8_formats_and_subtitles(video_url, video_id, 'mp4', fatal=False)
  85                 formats.extend(fmts)
  86                 self._merge_subtitles(subs, target=subtitles)
  87             else:
  88                 formats.append({
  89                     'url': video_url,
  90                     'ext': video_format,
  91                 })
  92
  93         return formats, subtitles
  94
  95     def _real_extract(self, url):
  96         display_id, username = self._match_valid_url(url).group('id', 'username')
  97         webpage = self._download_webpage(url, display_id)
  98
  99         webpage_info = self._parse_json(self._search_json(
 100             r'window\._preloads\s*=\s*JSON\.parse\(', webpage, 'json string',
 101             display_id, transform_source=js_to_json, contains_pattern=r'"{(?s:.+)}"'), display_id)
 102
 103         canonical_url = url
 104         domain = traverse_obj(webpage_info, ('domainInfo', 'customDomain', {str}))
 105         if domain:
 106             canonical_url = urllib.parse.urlparse(url)._replace(netloc=domain).geturl()
 107
 108         post_type = webpage_info['post']['type']
 109         formats, subtitles = [], {}
 110         if post_type == 'podcast':
 111             fmt = {'url': webpage_info['post']['podcast_url']}
 112             if not determine_ext(fmt['url'], default_ext=None):
 113                 # The redirected format URL expires but the original URL doesn't,
 114                 # so we only want to extract the extension from this request
 115                 fmt['ext'] = determine_ext(self._request_webpage(
 116                     HEADRequest(fmt['url']), display_id,
 117                     'Resolving podcast file extension',
 118                     'Podcast URL is invalid').url)
 119             formats.append(fmt)
 120         elif post_type == 'video':
 121             formats, subtitles = self._extract_video_formats(webpage_info['post']['videoUpload']['id'], canonical_url)
 122         else:
 123             self.raise_no_formats(f'Page type "{post_type}" is not supported')
 124
 125         return {
 126             'id': str(webpage_info['post']['id']),
 127             'formats': formats,
 128             'subtitles': subtitles,
 129             'title': traverse_obj(webpage_info, ('post', 'title')),
 130             'description': traverse_obj(webpage_info, ('post', 'description')),
 131             'thumbnail': traverse_obj(webpage_info, ('post', 'cover_image')),
 132             'uploader': traverse_obj(webpage_info, ('pub', 'name')),
 133             'uploader_id': str_or_none(traverse_obj(webpage_info, ('post', 'publication_id'))),
 134             'webpage_url': canonical_url,
 135         }