From 84213ea8d41d5fe1608333a16ac578dccdf9a915 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Sat, 6 Jun 2020 04:16:31 +0700 Subject: [PATCH] [youtube] Extract chapters from JSON (closes #24819) --- test/test_youtube_chapters.py | 2 +- youtube_dl/extractor/youtube.py | 63 +++++++++++++++++++++++++++++++-- 2 files changed, 62 insertions(+), 3 deletions(-) diff --git a/test/test_youtube_chapters.py b/test/test_youtube_chapters.py index 324ca8525..e69c57377 100644 --- a/test/test_youtube_chapters.py +++ b/test/test_youtube_chapters.py @@ -267,7 +267,7 @@ def test_youtube_chapters(self): for description, duration, expected_chapters in self._TEST_CASES: ie = YoutubeIE() expect_value( - self, ie._extract_chapters(description, duration), + self, ie._extract_chapters_from_description(description, duration), expected_chapters, None) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index fec17987b..54ec76db5 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1652,8 +1652,63 @@ def extract_id(cls, url): video_id = mobj.group(2) return video_id + def _extract_chapters_from_json(self, webpage, video_id, duration): + if not webpage: + return + player = self._parse_json( + self._search_regex( + r'RELATED_PLAYER_ARGS["\']\s*:\s*({.+})\s*,?\s*\n', webpage, + 'player args', default='{}'), + video_id, fatal=False) + if not player or not isinstance(player, dict): + return + watch_next_response = player.get('watch_next_response') + if not isinstance(watch_next_response, compat_str): + return + response = self._parse_json(watch_next_response, video_id, fatal=False) + if not response or not isinstance(response, dict): + return + chapters_list = try_get( + response, + lambda x: x['playerOverlays'] + ['playerOverlayRenderer'] + ['decoratedPlayerBarRenderer'] + ['decoratedPlayerBarRenderer'] + ['playerBar'] + ['chapteredPlayerBarRenderer'] + ['chapters'], + list) + if not chapters_list: + return + + def chapter_time(chapter): + return float_or_none( + try_get( + chapter, + lambda x: x['chapterRenderer']['timeRangeStartMillis'], + int), + scale=1000) + chapters = [] + for next_num, chapter in enumerate(chapters_list, start=1): + start_time = chapter_time(chapter) + if start_time is None: + continue + end_time = (chapter_time(chapters_list[next_num]) + if next_num < len(chapters_list) else duration) + if end_time is None: + continue + title = try_get( + chapter, lambda x: x['chapterRenderer']['title']['simpleText'], + compat_str) + chapters.append({ + 'start_time': start_time, + 'end_time': end_time, + 'title': title, + }) + return chapters + @staticmethod - def _extract_chapters(description, duration): + def _extract_chapters_from_description(description, duration): if not description: return None chapter_lines = re.findall( @@ -1687,6 +1742,10 @@ def _extract_chapters(description, duration): }) return chapters + def _extract_chapters(self, webpage, description, video_id, duration): + return (self._extract_chapters_from_json(webpage, video_id, duration) + or self._extract_chapters_from_description(description, duration)) + def _real_extract(self, url): url, smuggled_data = unsmuggle_url(url, {}) @@ -2324,7 +2383,7 @@ def _extract_count(count_name): errnote='Unable to download video annotations', fatal=False, data=urlencode_postdata({xsrf_field_name: xsrf_token})) - chapters = self._extract_chapters(description_original, video_duration) + chapters = self._extract_chapters(video_webpage, description_original, video_id, video_duration) # Look for the DASH manifest if self._downloader.params.get('youtube_include_dash_manifest', True):