mirror of
https://github.com/ytdl-org/youtube-dl
synced 2024-12-25 10:45:50 +01:00
167 lines
6.7 KiB
Python
167 lines
6.7 KiB
Python
# coding: utf-8
|
|
from __future__ import unicode_literals
|
|
|
|
import re
|
|
|
|
from .common import InfoExtractor
|
|
from ..compat import compat_str
|
|
from ..utils import (
|
|
ExtractorError,
|
|
unified_strdate,
|
|
HEADRequest,
|
|
int_or_none,
|
|
)
|
|
|
|
|
|
class WatIE(InfoExtractor):
|
|
_VALID_URL = r'(?:wat:|https?://(?:www\.)?wat\.tv/video/.*-)(?P<id>[0-9a-z]+)'
|
|
IE_NAME = 'wat.tv'
|
|
_TESTS = [
|
|
{
|
|
'url': 'http://www.wat.tv/video/soupe-figues-l-orange-aux-epices-6z1uz_2hvf7_.html',
|
|
'md5': '83d882d9de5c9d97f0bb2c6273cde56a',
|
|
'info_dict': {
|
|
'id': '11713067',
|
|
'ext': 'mp4',
|
|
'title': 'Soupe de figues à l\'orange et aux épices',
|
|
'description': 'Retrouvez l\'émission "Petits plats en équilibre", diffusée le 18 août 2014.',
|
|
'upload_date': '20140819',
|
|
'duration': 120,
|
|
},
|
|
},
|
|
{
|
|
'url': 'http://www.wat.tv/video/gregory-lemarchal-voix-ange-6z1v7_6ygkj_.html',
|
|
'md5': '34bdfa5ca9fd3c7eb88601b635b0424c',
|
|
'info_dict': {
|
|
'id': '11713075',
|
|
'ext': 'mp4',
|
|
'title': 'Grégory Lemarchal, une voix d\'ange depuis 10 ans (1/3)',
|
|
'upload_date': '20140816',
|
|
},
|
|
'expected_warnings': ["Ce contenu n'est pas disponible pour l'instant."],
|
|
},
|
|
]
|
|
|
|
_FORMATS = (
|
|
(200, 416, 234),
|
|
(400, 480, 270),
|
|
(600, 640, 360),
|
|
(1200, 640, 360),
|
|
(1800, 960, 540),
|
|
(2500, 1280, 720),
|
|
)
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
video_id = video_id if video_id.isdigit() and len(video_id) > 6 else compat_str(int(video_id, 36))
|
|
|
|
# 'contentv4' is used in the website, but it also returns the related
|
|
# videos, we don't need them
|
|
video_data = self._download_json(
|
|
'http://www.wat.tv/interface/contentv4s/' + video_id, video_id)
|
|
video_info = video_data['media']
|
|
|
|
error_desc = video_info.get('error_desc')
|
|
if error_desc:
|
|
self.report_warning(
|
|
'%s returned error: %s' % (self.IE_NAME, error_desc))
|
|
|
|
chapters = video_info['chapters']
|
|
if chapters:
|
|
first_chapter = chapters[0]
|
|
|
|
def video_id_for_chapter(chapter):
|
|
return chapter['tc_start'].split('-')[0]
|
|
|
|
if video_id_for_chapter(first_chapter) != video_id:
|
|
self.to_screen('Multipart video detected')
|
|
entries = [self.url_result('wat:%s' % video_id_for_chapter(chapter)) for chapter in chapters]
|
|
return self.playlist_result(entries, video_id, video_info['title'])
|
|
# Otherwise we can continue and extract just one part, we have to use
|
|
# the video id for getting the video url
|
|
else:
|
|
first_chapter = video_info
|
|
|
|
title = first_chapter['title']
|
|
|
|
def extract_url(path_template, url_type):
|
|
req_url = 'http://www.wat.tv/get/%s' % (path_template % video_id)
|
|
head = self._request_webpage(HEADRequest(req_url), video_id, 'Extracting %s url' % url_type, fatal=False)
|
|
if head:
|
|
red_url = head.geturl()
|
|
if req_url != red_url:
|
|
return red_url
|
|
return None
|
|
|
|
def remove_bitrate_limit(manifest_url):
|
|
return re.sub(r'(?:max|min)_bitrate=\d+&?', '', manifest_url)
|
|
|
|
formats = []
|
|
try:
|
|
manifest_urls = self._download_json(
|
|
'http://www.wat.tv/get/webhtml/' + video_id, video_id)
|
|
m3u8_url = manifest_urls.get('hls')
|
|
if m3u8_url:
|
|
m3u8_url = remove_bitrate_limit(m3u8_url)
|
|
m3u8_formats = self._extract_m3u8_formats(
|
|
m3u8_url, video_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False)
|
|
if m3u8_formats:
|
|
formats.extend(m3u8_formats)
|
|
formats.extend(self._extract_f4m_formats(
|
|
m3u8_url.replace('ios', 'web').replace('.m3u8', '.f4m'),
|
|
video_id, f4m_id='hds', fatal=False))
|
|
http_url = extract_url('android5/%s.mp4', 'http')
|
|
if http_url:
|
|
for m3u8_format in m3u8_formats:
|
|
vbr, abr = m3u8_format.get('vbr'), m3u8_format.get('abr')
|
|
if not vbr or not abr:
|
|
continue
|
|
format_id = m3u8_format['format_id'].replace('hls', 'http')
|
|
fmt_url = re.sub(r'%s-\d+00-\d+' % video_id, '%s-%d00-%d' % (video_id, round(vbr / 100), round(abr)), http_url)
|
|
if self._is_valid_url(fmt_url, video_id, format_id):
|
|
f = m3u8_format.copy()
|
|
f.update({
|
|
'url': fmt_url,
|
|
'format_id': format_id,
|
|
'protocol': 'http',
|
|
})
|
|
formats.append(f)
|
|
mpd_url = manifest_urls.get('mpd')
|
|
if mpd_url:
|
|
formats.extend(self._extract_mpd_formats(remove_bitrate_limit(
|
|
mpd_url), video_id, mpd_id='dash', fatal=False))
|
|
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')
|
|
upload_date = unified_strdate(date_diffusion) if date_diffusion else None
|
|
duration = None
|
|
files = video_info['files']
|
|
if files:
|
|
duration = int_or_none(files[0].get('duration'))
|
|
|
|
return {
|
|
'id': video_id,
|
|
'title': title,
|
|
'thumbnail': first_chapter.get('preview'),
|
|
'description': first_chapter.get('description'),
|
|
'view_count': int_or_none(video_info.get('views')),
|
|
'upload_date': upload_date,
|
|
'duration': duration,
|
|
'formats': formats,
|
|
}
|