1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-05 09:01:21 +00:00
youtube-dl/youtube_dl/extractor/ninecninemedia.py

122 lines
4.5 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..compat import compat_str
from ..utils import (
parse_iso8601,
float_or_none,
ExtractorError,
int_or_none,
)
class NineCNineMediaBaseIE(InfoExtractor):
_API_BASE_TEMPLATE = 'http://capi.9c9media.com/destinations/%s/platforms/desktop/contents/%s/'
class NineCNineMediaStackIE(NineCNineMediaBaseIE):
IE_NAME = '9c9media:stack'
_GEO_COUNTRIES = ['CA']
_VALID_URL = r'9c9media:stack:(?P<destination_code>[^:]+):(?P<content_id>\d+):(?P<content_package>\d+):(?P<id>\d+)'
def _real_extract(self, url):
destination_code, content_id, package_id, stack_id = re.match(self._VALID_URL, url).groups()
stack_base_url_template = self._API_BASE_TEMPLATE + 'contentpackages/%s/stacks/%s/manifest.'
stack_base_url = stack_base_url_template % (destination_code, content_id, package_id, stack_id)
formats = []
formats.extend(self._extract_m3u8_formats(
stack_base_url + 'm3u8', stack_id, 'mp4',
'm3u8_native', m3u8_id='hls', fatal=False))
formats.extend(self._extract_f4m_formats(
stack_base_url + 'f4m', stack_id,
f4m_id='hds', fatal=False))
self._sort_formats(formats)
return {
'id': stack_id,
'formats': formats,
}
class NineCNineMediaIE(NineCNineMediaBaseIE):
IE_NAME = '9c9media'
_VALID_URL = r'9c9media:(?P<destination_code>[^:]+):(?P<id>\d+)'
def _real_extract(self, url):
destination_code, content_id = re.match(self._VALID_URL, url).groups()
api_base_url = self._API_BASE_TEMPLATE % (destination_code, content_id)
content = self._download_json(api_base_url, content_id, query={
'$include': '[Media,Season,ContentPackages]',
})
title = content['Name']
if len(content['ContentPackages']) > 1:
raise ExtractorError('multiple content packages')
content_package = content['ContentPackages'][0]
package_id = content_package['Id']
content_package_url = api_base_url + 'contentpackages/%s/' % package_id
content_package = self._download_json(content_package_url, content_id)
if content_package.get('Constraints', {}).get('Security', {}).get('Type') == 'adobe-drm':
raise ExtractorError('This video is DRM protected.', expected=True)
stacks = self._download_json(content_package_url + 'stacks/', package_id)['Items']
multistacks = len(stacks) > 1
thumbnails = []
for image in content.get('Images', []):
image_url = image.get('Url')
if not image_url:
continue
thumbnails.append({
'url': image_url,
'width': int_or_none(image.get('Width')),
'height': int_or_none(image.get('Height')),
})
tags, categories = [], []
for source_name, container in (('Tags', tags), ('Genres', categories)):
for e in content.get(source_name, []):
e_name = e.get('Name')
if not e_name:
continue
container.append(e_name)
description = content.get('Desc') or content.get('ShortDesc')
season = content.get('Season', {})
base_info = {
'description': description,
'timestamp': parse_iso8601(content.get('BroadcastDateTime')),
'episode_number': int_or_none(content.get('Episode')),
'season': season.get('Name'),
'season_number': season.get('Number'),
'season_id': season.get('Id'),
'series': content.get('Media', {}).get('Name'),
'tags': tags,
'categories': categories,
}
entries = []
for stack in stacks:
stack_id = compat_str(stack['Id'])
entry = {
'_type': 'url_transparent',
'url': '9c9media:stack:%s:%s:%s:%s' % (destination_code, content_id, package_id, stack_id),
'id': stack_id,
'title': '%s_part%s' % (title, stack['Name']) if multistacks else title,
'duration': float_or_none(stack.get('Duration')),
'ie_key': 'NineCNineMediaStack',
}
entry.update(base_info)
entries.append(entry)
return {
'_type': 'multi_video',
'id': content_id,
'title': title,
'description': description,
'entries': entries,
}