[extractor/JStream] Add extractor (#6252)

Authored by: Lesmiscore

2 files changed, 74 insertions, 0 deletions

diff --git a/yt_dlp/extractor/_extractors.py b/yt_dlp/extractor/_extractors.py
index 0f65f1cc7..d560ed91c 100644
--- a/yt_dlp/extractor/_extractors.py
+++ b/yt_dlp/extractor/_extractors.py
@@ -878,6 +878,7 @@ from .japandiet import (
 from .jeuxvideo import JeuxVideoIE
 from .jove import JoveIE
 from .joj import JojIE
+from .jstream import JStreamIE
 from .jwplatform import JWPlatformIE
 from .kakao import KakaoIE
 from .kaltura import KalturaIE
diff --git a/yt_dlp/extractor/jstream.py b/yt_dlp/extractor/jstream.py
new file mode 100644
index 000000000..3e2e62712
--- /dev/null
+++ b/yt_dlp/extractor/jstream.py
@@ -0,0 +1,73 @@
+import base64
+import re
+import json
+
+from .common import InfoExtractor
+from ..utils import (
+    float_or_none,
+    js_to_json,
+    remove_start,
+)
+
+
+class JStreamIE(InfoExtractor):
+    # group "id" only exists for compliance, not directly used in requests
+    # also all components are mandatory
+    _VALID_URL = r'jstream:(?P<host>www\d+):(?P<id>(?P<publisher>[a-z0-9]+):(?P<mid>\d+))'
+
+    _TESTS = [{
+        'url': 'jstream:www50:eqd638pvwx:752',
+        'info_dict': {
+            'id': 'eqd638pvwx:752',
+            'ext': 'mp4',
+            'title': '阪神淡路大震災 激震の記録2020年版　解説動画',
+            'duration': 672,
+            'thumbnail': r're:https?://eqd638pvwx\.eq\.webcdn\.stream\.ne\.jp/.+\.jpg',
+        },
+    }]
+
+    def _parse_jsonp(self, callback, string, video_id):
+        return self._search_json(rf'\s*{re.escape(callback)}\s*\(', string, callback, video_id)
+
+    def _find_formats(self, video_id, movie_list_hls, host, publisher, subtitles):
+        for value in movie_list_hls:
+            text = value.get('text') or ''
+            if not text.startswith('auto'):
+                continue
+            m3u8_id = remove_start(remove_start(text, 'auto'), '_') or None
+            fmts, subs = self._extract_m3u8_formats_and_subtitles(
+                f'https://{publisher}.eq.webcdn.stream.ne.jp/{host}/{publisher}/jmc_pub/{value.get("url")}', video_id, 'mp4', m3u8_id=m3u8_id)
+            self._merge_subtitles(subs, target=subtitles)
+            yield from fmts
+
+    def _real_extract(self, url):
+        host, publisher, mid, video_id = self._match_valid_url(url).group('host', 'publisher', 'mid', 'id')
+        video_info_jsonp = self._download_webpage(
+            f'https://{publisher}.eq.webcdn.stream.ne.jp/{host}/{publisher}/jmc_pub/eq_meta/v1/{mid}.jsonp',
+            video_id, 'Requesting video info')
+        video_info = self._parse_jsonp('metaDataResult', video_info_jsonp, video_id)['movie']
+        subtitles = {}
+        formats = list(self._find_formats(video_id, video_info.get('movie_list_hls'), host, publisher, subtitles))
+        self._remove_duplicate_formats(formats)
+        return {
+            'id': video_id,
+            'title': video_info.get('title'),
+            'duration': float_or_none(video_info.get('duration')),
+            'thumbnail': video_info.get('thumbnail_url'),
+            'formats': formats,
+            'subtitles': subtitles,
+        }
+
+    @classmethod
+    def _extract_embed_urls(cls, url, webpage):
+        # check for eligiblity of webpage
+        # https://support.eq.stream.co.jp/hc/ja/articles/115008388147-%E3%83%97%E3%83%AC%E3%82%A4%E3%83%A4%E3%83%BCAPI%E3%81%AE%E3%82%B5%E3%83%B3%E3%83%97%E3%83%AB%E3%82%B3%E3%83%BC%E3%83%89
+        script_tag = re.search(r'<script\s*[^>]+?src="https://ssl-cache\.stream\.ne\.jp/(?P<host>www\d+)/(?P<publisher>[a-z0-9]+)/[^"]+?/if\.js"', webpage)
+        if not script_tag:
+            return
+        host, publisher = script_tag.groups()
+        for m in re.finditer(r'(?s)PlayerFactoryIF\.create\(\s*({[^\}]+?})\s*\)\s*;', webpage):
+            # TODO: using json.loads here as InfoExtractor._parse_json is not classmethod
+            info = json.loads(js_to_json(m.group(1)))
+            mid = base64.b64decode(info.get('m')).decode()
+            yield f'jstream:{host}:{publisher}:{mid}'