yt-dlp/yt_dlp/extractor/japandiet.py

278 lines
10 KiB
Python
Raw Permalink Normal View History

import re
from .common import InfoExtractor
from ..utils import (
ExtractorError,
clean_html,
int_or_none,
join_nonempty,
parse_qs,
smuggle_url,
traverse_obj,
try_call,
unsmuggle_url,
)
def _parse_japanese_date(text):
if not text:
return None
ERA_TABLE = {
'明治': 1868,
'大正': 1912,
'昭和': 1926,
'平成': 1989,
'令和': 2019,
}
ERA_RE = '|'.join(map(re.escape, ERA_TABLE.keys()))
mobj = re.search(rf'({ERA_RE})?(\d+)年(\d+)月(\d+)日', re.sub(r'[\s\u3000]+', '', text))
if not mobj:
return None
era, year, month, day = mobj.groups()
year, month, day = map(int, (year, month, day))
if era:
# example input: 令和5年3月34日
# even though each era have their end, don't check here
year += ERA_TABLE[era]
return '%04d%02d%02d' % (year, month, day)
def _parse_japanese_duration(text):
mobj = re.search(r'(?:(\d+)日間?)?(?:(\d+)時間?)?(?:(\d+)分)?(?:(\d+)秒)?', re.sub(r'[\s\u3000]+', '', text or ''))
if not mobj:
return
days, hours, mins, secs = (int_or_none(x, default=0) for x in mobj.groups())
return secs + mins * 60 + hours * 60 * 60 + days * 24 * 60 * 60
class ShugiinItvBaseIE(InfoExtractor):
_INDEX_ROOMS = None
@classmethod
def _find_rooms(cls, webpage):
return [{
'_type': 'url',
'id': x.group(1),
'title': clean_html(x.group(2)).strip(),
'url': smuggle_url(f'https://www.shugiintv.go.jp/jp/index.php?room_id={x.group(1)}', {'g': x.groups()}),
'ie_key': ShugiinItvLiveIE.ie_key(),
} for x in re.finditer(r'(?s)<a\s+href="[^"]+\?room_id=(room\d+)"\s*class="play_live".+?class="s12_14">(.+?)</td>', webpage)]
def _fetch_rooms(self):
if not self._INDEX_ROOMS:
webpage = self._download_webpage(
'https://www.shugiintv.go.jp/jp/index.php', None,
encoding='euc-jp', note='Downloading proceedings info')
ShugiinItvBaseIE._INDEX_ROOMS = self._find_rooms(webpage)
return self._INDEX_ROOMS
class ShugiinItvLiveIE(ShugiinItvBaseIE):
_VALID_URL = r'https?://(?:www\.)?shugiintv\.go\.jp/(?:jp|en)(?:/index\.php)?$'
IE_DESC = '衆議院インターネット審議中継'
_TESTS = [{
'url': 'https://www.shugiintv.go.jp/jp/index.php',
'info_dict': {
'_type': 'playlist',
'title': 'All proceedings for today',
},
# expect at least one proceedings is running
'playlist_mincount': 1,
}]
@classmethod
def suitable(cls, url):
return super().suitable(url) and not any(x.suitable(url) for x in (ShugiinItvLiveRoomIE, ShugiinItvVodIE))
def _real_extract(self, url):
self.to_screen(
'Downloading all running proceedings. To specify one proceeding, use direct link from the website')
return self.playlist_result(self._fetch_rooms(), playlist_title='All proceedings for today')
class ShugiinItvLiveRoomIE(ShugiinItvBaseIE):
_VALID_URL = r'https?://(?:www\.)?shugiintv\.go\.jp/(?:jp|en)/index\.php\?room_id=(?P<id>room\d+)'
IE_DESC = '衆議院インターネット審議中継 (中継)'
_TESTS = [{
'url': 'https://www.shugiintv.go.jp/jp/index.php?room_id=room01',
'info_dict': {
'id': 'room01',
'title': '内閣委員会',
},
'skip': 'this runs for a time and not every day',
}, {
'url': 'https://www.shugiintv.go.jp/jp/index.php?room_id=room11',
'info_dict': {
'id': 'room11',
'title': '外務委員会',
},
'skip': 'this runs for a time and not every day',
}]
def _real_extract(self, url):
url, smug = unsmuggle_url(url, default={})
if smug.get('g'):
room_id, title = smug['g']
else:
room_id = self._match_id(url)
title = traverse_obj(self._fetch_rooms(), (lambda k, v: v['id'] == room_id, 'title'), get_all=False)
formats, subtitles = self._extract_m3u8_formats_and_subtitles(
f'https://hlslive.shugiintv.go.jp/{room_id}/amlst:{room_id}/playlist.m3u8',
room_id, ext='mp4')
return {
'id': room_id,
'title': title,
'formats': formats,
'subtitles': subtitles,
'is_live': True,
}
class ShugiinItvVodIE(ShugiinItvBaseIE):
_VALID_URL = r'https?://(?:www\.)?shugiintv\.go\.jp/(?:jp|en)/index\.php\?ex=VL(?:\&[^=]+=[^&]*)*\&deli_id=(?P<id>\d+)'
IE_DESC = '衆議院インターネット審議中継 (ビデオライブラリ)'
_TESTS = [{
'url': 'https://www.shugiintv.go.jp/jp/index.php?ex=VL&media_type=&deli_id=53846',
'info_dict': {
'id': '53846',
'title': 'ウクライナ大統領国会演説(オンライン)',
'release_date': '20220323',
'chapters': 'count:4',
},
}, {
'url': 'https://www.shugiintv.go.jp/en/index.php?ex=VL&media_type=&deli_id=53846',
'only_matching': True,
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(
f'https://www.shugiintv.go.jp/jp/index.php?ex=VL&media_type=&deli_id={video_id}', video_id,
encoding='euc-jp')
m3u8_url = self._search_regex(
r'id="vtag_src_base_vod"\s*value="(http.+?\.m3u8)"', webpage, 'm3u8 url')
m3u8_url = re.sub(r'^http://', 'https://', m3u8_url)
formats, subtitles = self._extract_m3u8_formats_and_subtitles(
m3u8_url, video_id, ext='mp4')
title = self._html_search_regex(
(r'<td\s+align="left">(.+)\s*\(\d+分\)',
r'<TD.+?<IMG\s*src=".+?/spacer\.gif".+?height="15">(.+?)<IMG'), webpage, 'title', fatal=False)
release_date = _parse_japanese_date(self._html_search_regex(
r'開会日</td>\s*<td.+?/td>\s*<TD>(.+?)</TD>',
webpage, 'title', fatal=False))
chapters = []
for chp in re.finditer(r'(?i)<A\s+HREF="([^"]+?)"\s*class="play_vod">(?!<img)(.+)</[Aa]>', webpage):
chapters.append({
'title': clean_html(chp.group(2)).strip(),
'start_time': try_call(lambda: float(parse_qs(chp.group(1))['time'][0].strip())),
})
# NOTE: there are blanks at the first and the end of the videos,
# so getting/providing the video duration is not possible
# also, the exact end_time for the last chapter is unknown (we can get at most minutes of granularity)
last_tr = re.findall(r'(?s)<TR\s*class="s14_24">(.+?)</TR>', webpage)[-1]
if last_tr and chapters:
last_td = re.findall(r'<TD.+?</TD>', last_tr)[-1]
if last_td:
chapters[-1]['end_time'] = chapters[-1]['start_time'] + _parse_japanese_duration(clean_html(last_td))
return {
'id': video_id,
'title': title,
'release_date': release_date,
'chapters': chapters,
'formats': formats,
'subtitles': subtitles,
}
class SangiinInstructionIE(InfoExtractor):
_VALID_URL = r'https?://www\.webtv\.sangiin\.go\.jp/webtv/index\.php'
IE_DESC = False # this shouldn't be listed as a supported site
def _real_extract(self, url):
raise ExtractorError(
'Copy the link from the button below the video description/player '
'and use that link to download. If there is no button in the frame, '
'get the URL of the frame showing the video.', expected=True)
class SangiinIE(InfoExtractor):
_VALID_URL = r'https?://www\.webtv\.sangiin\.go\.jp/webtv/detail\.php\?sid=(?P<id>\d+)'
IE_DESC = '参議院インターネット審議中継 (archive)'
_TESTS = [{
'url': 'https://www.webtv.sangiin.go.jp/webtv/detail.php?sid=7052',
'info_dict': {
'id': '7052',
'title': '2022年10月7日 本会議',
'description': 'md5:0a5fed523f95c88105a0b0bf1dd71489',
'upload_date': '20221007',
'ext': 'mp4',
},
}, {
'url': 'https://www.webtv.sangiin.go.jp/webtv/detail.php?sid=7037',
'info_dict': {
'id': '7037',
'title': '2022年10月3日 開会式',
'upload_date': '20221003',
'ext': 'mp4',
},
}, {
'url': 'https://www.webtv.sangiin.go.jp/webtv/detail.php?sid=7076',
'info_dict': {
'id': '7076',
'title': '2022年10月27日 法務委員会',
'upload_date': '20221027',
'ext': 'mp4',
'is_live': True,
},
'skip': 'this live is turned into archive after it ends',
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
date = self._html_search_regex(
r'<dt[^>]*>\s*開会日\s*</dt>\s*<dd[^>]*>\s*(.+?)\s*</dd>', webpage,
'date', fatal=False)
upload_date = _parse_japanese_date(date)
title = self._html_search_regex(
r'<dt[^>]*>\s*会議名\s*</dt>\s*<dd[^>]*>\s*(.+?)\s*</dd>', webpage,
'date', fatal=False)
# some videos don't have the elements, so assume it's missing
description = self._html_search_regex(
r'会議の経過\s*</h3>\s*<span[^>]*>(.+?)</span>', webpage,
'description', default=None)
# this row appears only when it's livestream
is_live = bool(self._html_search_regex(
r'<dt[^>]*>\s*公報掲載時刻\s*</dt>\s*<dd[^>]*>\s*(.+?)\s*</dd>', webpage,
'is_live', default=None))
m3u8_url = self._search_regex(
r'var\s+videopath\s*=\s*(["\'])([^"\']+)\1', webpage,
'm3u8 url', group=2)
formats, subs = self._extract_m3u8_formats_and_subtitles(m3u8_url, video_id, 'mp4')
return {
'id': video_id,
'title': join_nonempty(date, title, delim=' '),
'description': description,
'upload_date': upload_date,
'formats': formats,
'subtitles': subs,
'is_live': is_live,
}