[wat] fix format extraction(closes #27901)

This commit is contained in:
Remita Amine 2021-01-21 17:20:32 +01:00
parent cff72b4cc0
commit 8d286bd5b6

View file

@ -1,12 +1,9 @@
# coding: utf-8 # coding: utf-8
from __future__ import unicode_literals from __future__ import unicode_literals
import re
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import compat_str from ..compat import compat_str
from ..utils import ( from ..utils import (
ExtractorError,
unified_strdate, unified_strdate,
HEADRequest, HEADRequest,
int_or_none, int_or_none,
@ -97,46 +94,20 @@ class WatIE(InfoExtractor):
return red_url return red_url
return None return None
def remove_bitrate_limit(manifest_url):
return re.sub(r'(?:max|min)_bitrate=\d+&?', '', manifest_url)
formats = [] formats = []
try:
alt_urls = lambda manifest_url: [re.sub(r'(?:wdv|ssm)?\.ism/', repl + '.ism/', manifest_url) for repl in ('', 'ssm')]
manifest_urls = self._download_json( manifest_urls = self._download_json(
'http://www.wat.tv/get/webhtml/' + video_id, video_id) 'http://www.wat.tv/get/webhtml/' + video_id, video_id)
m3u8_url = manifest_urls.get('hls') m3u8_url = manifest_urls.get('hls')
if m3u8_url: if m3u8_url:
m3u8_url = remove_bitrate_limit(m3u8_url)
for m3u8_alt_url in alt_urls(m3u8_url):
formats.extend(self._extract_m3u8_formats( formats.extend(self._extract_m3u8_formats(
m3u8_alt_url, video_id, 'mp4', m3u8_url, video_id, 'mp4',
'm3u8_native', m3u8_id='hls', fatal=False)) 'm3u8_native', m3u8_id='hls', fatal=False))
formats.extend(self._extract_f4m_formats(
m3u8_alt_url.replace('ios', 'web').replace('.m3u8', '.f4m'),
video_id, f4m_id='hds', fatal=False))
mpd_url = manifest_urls.get('mpd') mpd_url = manifest_urls.get('mpd')
if mpd_url: if mpd_url:
mpd_url = remove_bitrate_limit(mpd_url)
for mpd_alt_url in alt_urls(mpd_url):
formats.extend(self._extract_mpd_formats( formats.extend(self._extract_mpd_formats(
mpd_alt_url, video_id, mpd_id='dash', fatal=False)) mpd_url.replace('://das-q1.tf1.fr/', '://das-q1-ssl.tf1.fr/'),
video_id, mpd_id='dash', fatal=False))
self._sort_formats(formats) self._sort_formats(formats)
except ExtractorError:
abr = 64
for vbr, width, height in self._FORMATS:
tbr = vbr + abr
format_id = 'http-%s' % tbr
fmt_url = 'http://dnl.adv.tf1.fr/2/USP-0x0/%s/%s/%s/ssm/%s-%s-64k.mp4' % (video_id[-4:-2], video_id[-2:], video_id, video_id, vbr)
if self._is_valid_url(fmt_url, video_id, format_id):
formats.append({
'format_id': format_id,
'url': fmt_url,
'vbr': vbr,
'abr': abr,
'width': width,
'height': height,
})
date_diffusion = first_chapter.get('date_diffusion') or video_data.get('configv4', {}).get('estatS4') date_diffusion = first_chapter.get('date_diffusion') or video_data.get('configv4', {}).get('estatS4')
upload_date = unified_strdate(date_diffusion) if date_diffusion else None upload_date = unified_strdate(date_diffusion) if date_diffusion else None