[extractor/YahooJapanNews] Fix extractor (#4480)

Authored by: Lesmiscore

diff --git a/yt_dlp/extractor/yahoo.py b/yt_dlp/extractor/yahoo.py
index 8811df6d85a77f5f728b657fdbc70126fe43ec77..f85990e0a586e121268ee0af69201b0a6cf8c3dc 100644 (file)
--- a/yt_dlp/extractor/yahoo.py
+++ b/yt_dlp/extractor/yahoo.py
@@ -1,12 +1,10 @@
 import hashlib
 import itertools
-import re
 import urllib.parse
 
 from .brightcove import BrightcoveNewIE
 from .common import InfoExtractor, SearchInfoExtractor
 from .youtube import YoutubeIE
-from ..compat import compat_str
 from ..utils import (
     ExtractorError,
     clean_html,
@@ -14,6 +12,7 @@
     mimetype2ext,
     parse_iso8601,
     smuggle_url,
+    traverse_obj,
     try_get,
     url_or_none,
 )
@@ -456,33 +455,20 @@ def _real_extract(self, url):
 class YahooJapanNewsIE(InfoExtractor):
     IE_NAME = 'yahoo:japannews'
     IE_DESC = 'Yahoo! Japan News'
-    _VALID_URL = r'https?://(?P<host>(?:news|headlines)\.yahoo\.co\.jp)[^\d]*(?P<id>\d[\d-]*\d)?'
+    _VALID_URL = r'https?://news\.yahoo\.co\.jp/(?:articles|feature)/(?P<id>[a-zA-Z0-9]+)'
     _GEO_COUNTRIES = ['JP']
     _TESTS = [{
-        'url': 'https://headlines.yahoo.co.jp/videonews/ann?a=20190716-00000071-ann-int',
+        'url': 'https://news.yahoo.co.jp/articles/a70fe3a064f1cfec937e2252c7fc6c1ba3201c0e',
         'info_dict': {
-            'id': '1736242',
+            'id': 'a70fe3a064f1cfec937e2252c7fc6c1ba3201c0e',
             'ext': 'mp4',
-            'title': 'ã\83 ã\83³å¤§çµ±é \98ã\81\8c対æ\97¥æ\89¹å\88¤ã\82\92å¼·å\8c\96â\80\9cç\8f¾é\87\91å\8c\96â\80\9då\8a¹æ\9e\9cã\81¯ï¼\9fï¼\88ã\83\86ã\83¬ã\83\93æ\9c\9dæ\97¥ç³»ï¼\88ANNï¼\89ï¼\89 - Yahoo!ã\83\8bã\83¥ã\83¼ã\82¹',
-            'description': '韓国の元徴用工らを巡る裁判の原告が弁護士が差し押さえた三菱重工業の資産を売却して - Yahoo!ニュース(テレビ朝日系（ANN）)',
-            'thumbnail': r're:^https?://.*\.[a-zA-Z\d]{3,4}$',
+            'title': 'ã\80\90ç\8b¬è\87ªã\80\91å®\89å\80\8då\85\83ç·\8fç\90\86ã\80\8cå\9b½è\91¬ã\80\8d中止æ±\82ã\82\81â\80\9cè\84\85è¿«ã\83¡ã\83¼ã\83«â\80\9dâ\80¦ã\80\8cå­\90ã\81©ã\82\82èª\98æ\8b\90ã\80\8dâ\80\9cé\80\81ä¿¡è\80\85â\80\9dã\82\92追跡',
+            'description': 'md5:1c06974575f930f692d8696fbcfdc546',
+            'thumbnail': r're:https://.+',
         },
         'params': {
             'skip_download': True,
         },
-    }, {
-        # geo restricted
-        'url': 'https://headlines.yahoo.co.jp/hl?a=20190721-00000001-oxv-l04',
-        'only_matching': True,
-    }, {
-        'url': 'https://headlines.yahoo.co.jp/videonews/',
-        'only_matching': True,
-    }, {
-        'url': 'https://news.yahoo.co.jp',
-        'only_matching': True,
-    }, {
-        'url': 'https://news.yahoo.co.jp/byline/hashimotojunji/20190628-00131977/',
-        'only_matching': True,
     }, {
         'url': 'https://news.yahoo.co.jp/feature/1356',
         'only_matching': True
@@ -491,11 +477,7 @@ class YahooJapanNewsIE(InfoExtractor):
     def _extract_formats(self, json_data, content_id):
         formats = []
 
-        video_data = try_get(
-            json_data,
-            lambda x: x['ResultSet']['Result'][0]['VideoUrlSet']['VideoUrl'],
-            list)
-        for vid in video_data or []:
+        for vid in traverse_obj(json_data, ('ResultSet', 'Result', ..., 'VideoUrlSet', 'VideoUrl', ...)) or []:
             delivery = vid.get('delivery')
             url = url_or_none(vid.get('Url'))
             if not delivery or not url:
@@ -508,7 +490,7 @@ def _extract_formats(self, json_data, content_id):
             else:
                 formats.append({
                     'url': url,
-                    'format_id': 'http-%s' % compat_str(vid.get('bitrate', '')),
+                    'format_id': f'http-{vid.get("bitrate")}',
                     'height': int_or_none(vid.get('height')),
                     'width': int_or_none(vid.get('width')),
                     'tbr': int_or_none(vid.get('bitrate')),
@@ -519,62 +501,48 @@ def _extract_formats(self, json_data, content_id):
         return formats
 
     def _real_extract(self, url):
-        mobj = self._match_valid_url(url)
-        host = mobj.group('host')
-        display_id = mobj.group('id') or host
-
-        webpage = self._download_webpage(url, display_id)
-
-        title = self._html_search_meta(
-            ['og:title', 'twitter:title'], webpage, 'title', default=None
-        ) or self._html_extract_title(webpage)
-
-        if display_id == host:
-            # Headline page (w/ multiple BC playlists) ('news.yahoo.co.jp', 'headlines.yahoo.co.jp/videonews/', ...)
-            stream_plists = re.findall(r'plist=(\d+)', webpage) or re.findall(r'plist["\']:\s*["\']([^"\']+)', webpage)
-            entries = [
-                self.url_result(
-                    smuggle_url(
-                        'http://players.brightcove.net/5690807595001/HyZNerRl7_default/index.html?playlistId=%s' % plist_id,
-                        {'geo_countries': ['JP']}),
-                    ie='BrightcoveNew', video_id=plist_id)
-                for plist_id in stream_plists]
-            return self.playlist_result(entries, playlist_title=title)
-
-        # Article page
-        description = self._html_search_meta(
-            ['og:description', 'description', 'twitter:description'],
-            webpage, 'description', default=None)
-        thumbnail = self._og_search_thumbnail(
-            webpage, default=None) or self._html_search_meta(
-            'twitter:image', webpage, 'thumbnail', default=None)
-        space_id = self._search_regex([
-            r'<script[^>]+class=["\']yvpub-player["\'][^>]+spaceid=([^&"\']+)',
-            r'YAHOO\.JP\.srch\.\w+link\.onLoad[^;]+spaceID["\' ]*:["\' ]+([^"\']+)',
-            r'<!--\s+SpaceID=(\d+)'
-        ], webpage, 'spaceid')
-
-        content_id = self._search_regex(
-            r'<script[^>]+class=["\']yvpub-player["\'][^>]+contentid=(?P<contentid>[^&"\']+)',
-            webpage, 'contentid', group='contentid')
-
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+        preloaded_state = self._search_json(r'__PRELOADED_STATE__\s*=', webpage, 'preloaded state', video_id)
+
+        content_id = traverse_obj(
+            preloaded_state, ('articleDetail', 'paragraphs', ..., 'objectItems', ..., 'video', 'vid'),
+            get_all=False, expected_type=int)
+        if content_id is None:
+            raise ExtractorError('This article does not contain a video', expected=True)
+
+        HOST = 'news.yahoo.co.jp'
+        space_id = traverse_obj(preloaded_state, ('pageData', 'spaceId'), expected_type=str)
         json_data = self._download_json(
-            'https://feapi-yvpub.yahooapis.jp/v1/content/%s' % content_id,
-            content_id,
-            query={
+            f'https://feapi-yvpub.yahooapis.jp/v1/content/{content_id}',
+            video_id, query={
                 'appid': 'dj0zaiZpPVZMTVFJR0FwZWpiMyZzPWNvbnN1bWVyc2VjcmV0Jng9YjU-',
                 'output': 'json',
-                'space_id': space_id,
-                'domain': host,
-                'ak': hashlib.md5('_'.join((space_id, host)).encode()).hexdigest(),
+                'domain': HOST,
+                'ak': hashlib.md5('_'.join((space_id, HOST)).encode()).hexdigest() if space_id else '',
                 'device_type': '1100',
             })
-        formats = self._extract_formats(json_data, content_id)
+
+        title = (
+            traverse_obj(preloaded_state,
+                         ('articleDetail', 'headline'), ('pageData', 'pageParam', 'title'),
+                         expected_type=str)
+            or self._html_search_meta(('og:title', 'twitter:title'), webpage, 'title', default=None)
+            or self._html_extract_title(webpage))
+        description = (
+            traverse_obj(preloaded_state, ('pageData', 'description'), expected_type=str)
+            or self._html_search_meta(
+                ('og:description', 'description', 'twitter:description'),
+                webpage, 'description', default=None))
+        thumbnail = (
+            traverse_obj(preloaded_state, ('pageData', 'ogpImage'), expected_type=str)
+            or self._og_search_thumbnail(webpage, default=None)
+            or self._html_search_meta('twitter:image', webpage, 'thumbnail', default=None))
 
         return {
-            'id': content_id,
+            'id': video_id,
             'title': title,
             'description': description,
             'thumbnail': thumbnail,
-            'formats': formats,
+            'formats': self._extract_formats(json_data, video_id),
         }