1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-22 10:11:52 +00:00
youtube-dl/youtube_dl/extractor/zingmp3.py

161 lines
5.4 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
from .common import InfoExtractor
from ..utils import (
ExtractorError,
int_or_none,
)
class ZingMp3BaseIE(InfoExtractor):
_VALID_URL_TMPL = r'https?://(?:mp3\.zing|zingmp3)\.vn/(?:%s)/[^/]+/(?P<id>\w+)\.html'
_GEO_COUNTRIES = ['VN']
def _extract_item(self, item, fatal):
item_id = item['id']
title = item.get('name') or item['title']
formats = []
for k, v in (item.get('source') or {}).items():
if not v:
continue
if k in ('mp4', 'hls'):
for res, video_url in v.items():
if not video_url:
continue
if k == 'hls':
formats.extend(self._extract_m3u8_formats(
video_url, item_id, 'mp4',
'm3u8_native', m3u8_id=k, fatal=False))
elif k == 'mp4':
formats.append({
'format_id': 'mp4-' + res,
'url': video_url,
'height': int_or_none(self._search_regex(
r'^(\d+)p', res, 'resolution', default=None)),
})
else:
formats.append({
'ext': 'mp3',
'format_id': k,
'tbr': int_or_none(k),
'url': self._proto_relative_url(v),
'vcodec': 'none',
})
if not formats:
if not fatal:
return
msg = item['msg']
if msg == 'Sorry, this content is not available in your country.':
self.raise_geo_restricted(countries=self._GEO_COUNTRIES)
raise ExtractorError(msg, expected=True)
self._sort_formats(formats)
subtitles = None
lyric = item.get('lyric')
if lyric:
subtitles = {
'origin': [{
'url': lyric,
}],
}
album = item.get('album') or {}
return {
'id': item_id,
'title': title,
'formats': formats,
'thumbnail': item.get('thumbnail'),
'subtitles': subtitles,
'duration': int_or_none(item.get('duration')),
'track': title,
'artist': item.get('artists_names'),
'album': album.get('name') or album.get('title'),
'album_artist': album.get('artists_names'),
}
def _real_extract(self, url):
page_id = self._match_id(url)
webpage = self._download_webpage(
url.replace('://zingmp3.vn/', '://mp3.zing.vn/'),
page_id, query={'play_song': 1})
data_path = self._search_regex(
r'data-xml="([^"]+)', webpage, 'data path')
return self._process_data(self._download_json(
'https://mp3.zing.vn/xhr' + data_path, page_id)['data'])
class ZingMp3IE(ZingMp3BaseIE):
_VALID_URL = ZingMp3BaseIE._VALID_URL_TMPL % 'bai-hat|video-clip'
_TESTS = [{
'url': 'http://mp3.zing.vn/bai-hat/Xa-Mai-Xa-Bao-Thy/ZWZB9WAB.html',
'md5': 'ead7ae13693b3205cbc89536a077daed',
'info_dict': {
'id': 'ZWZB9WAB',
'title': 'Xa Mãi Xa',
'ext': 'mp3',
'thumbnail': r're:^https?://.+\.jpg',
'subtitles': {
'origin': [{
'ext': 'lrc',
}]
},
'duration': 255,
'track': 'Xa Mãi Xa',
'artist': 'Bảo Thy',
'album': 'Special Album',
'album_artist': 'Bảo Thy',
},
}, {
'url': 'https://mp3.zing.vn/video-clip/Suong-Hoa-Dua-Loi-K-ICM-RYO/ZO8ZF7C7.html',
'md5': 'e9c972b693aa88301ef981c8151c4343',
'info_dict': {
'id': 'ZO8ZF7C7',
'title': 'Sương Hoa Đưa Lối',
'ext': 'mp4',
'thumbnail': r're:^https?://.+\.jpg',
'duration': 207,
'track': 'Sương Hoa Đưa Lối',
'artist': 'K-ICM, RYO',
},
}, {
'url': 'https://zingmp3.vn/bai-hat/Xa-Mai-Xa-Bao-Thy/ZWZB9WAB.html',
'only_matching': True,
}]
IE_NAME = 'zingmp3'
IE_DESC = 'mp3.zing.vn'
def _process_data(self, data):
return self._extract_item(data, True)
class ZingMp3AlbumIE(ZingMp3BaseIE):
_VALID_URL = ZingMp3BaseIE._VALID_URL_TMPL % 'album|playlist'
_TESTS = [{
'url': 'http://mp3.zing.vn/album/Lau-Dai-Tinh-Ai-Bang-Kieu-Minh-Tuyet/ZWZBWDAF.html',
'info_dict': {
'_type': 'playlist',
'id': 'ZWZBWDAF',
'title': 'Lâu Đài Tình Ái',
},
'playlist_count': 10,
}, {
'url': 'http://mp3.zing.vn/playlist/Duong-Hong-Loan-apollobee/IWCAACCB.html',
'only_matching': True,
}, {
'url': 'https://zingmp3.vn/album/Lau-Dai-Tinh-Ai-Bang-Kieu-Minh-Tuyet/ZWZBWDAF.html',
'only_matching': True,
}]
IE_NAME = 'zingmp3:album'
def _process_data(self, data):
def entries():
for item in (data.get('items') or []):
entry = self._extract_item(item, False)
if entry:
yield entry
info = data.get('info') or {}
return self.playlist_result(
entries(), info.get('id'), info.get('name') or info.get('title'))