You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
l1ving_youtube-dl/youtube_dl/extractor/openload.py

360 lines
12 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
import json
import os
import re
import subprocess
import tempfile
from .common import InfoExtractor
from ..compat import (
compat_urlparse,
compat_kwargs,
)
from ..utils import (
check_executable,
determine_ext,
encodeArgument,
ExtractorError,
get_element_by_id,
get_exe_version,
is_outdated_version,
std_headers,
)
def cookie_to_dict(cookie):
cookie_dict = {
'name': cookie.name,
'value': cookie.value,
}
if cookie.port_specified:
cookie_dict['port'] = cookie.port
if cookie.domain_specified:
cookie_dict['domain'] = cookie.domain
if cookie.path_specified:
cookie_dict['path'] = cookie.path
if cookie.expires is not None:
cookie_dict['expires'] = cookie.expires
if cookie.secure is not None:
cookie_dict['secure'] = cookie.secure
if cookie.discard is not None:
cookie_dict['discard'] = cookie.discard
try:
if (cookie.has_nonstandard_attr('httpOnly') or
cookie.has_nonstandard_attr('httponly') or
cookie.has_nonstandard_attr('HttpOnly')):
cookie_dict['httponly'] = True
except TypeError:
pass
return cookie_dict
def cookie_jar_to_list(cookie_jar):
return [cookie_to_dict(cookie) for cookie in cookie_jar]
class PhantomJSwrapper(object):
"""PhantomJS wrapper class
This class is experimental.
"""
_TEMPLATE = r'''
phantom.onError = function(msg, trace) {{
var msgStack = ['PHANTOM ERROR: ' + msg];
if(trace && trace.length) {{
msgStack.push('TRACE:');
trace.forEach(function(t) {{
msgStack.push(' -> ' + (t.file || t.sourceURL) + ': ' + t.line
+ (t.function ? ' (in function ' + t.function +')' : ''));
}});
}}
console.error(msgStack.join('\n'));
phantom.exit(1);
}};
var page = require('webpage').create();
var fs = require('fs');
var read = {{ mode: 'r', charset: 'utf-8' }};
var write = {{ mode: 'w', charset: 'utf-8' }};
JSON.parse(fs.read("{cookies}", read)).forEach(function(x) {{
phantom.addCookie(x);
}});
page.settings.resourceTimeout = {timeout};
page.settings.userAgent = "{ua}";
page.onLoadStarted = function() {{
page.evaluate(function() {{
delete window._phantom;
delete window.callPhantom;
}});
}};
var saveAndExit = function() {{
fs.write("{html}", page.content, write);
fs.write("{cookies}", JSON.stringify(phantom.cookies), write);
phantom.exit();
}};
page.onLoadFinished = function(status) {{
if(page.url === "") {{
page.setContent(fs.read("{html}", read), "{url}");
}}
else {{
{jscode}
}}
}};
page.open("");
'''
_TMP_FILE_NAMES = ['script', 'html', 'cookies']
@staticmethod
def _version():
return get_exe_version('phantomjs', version_re=r'([0-9.]+)')
def __init__(self, extractor, required_version=None, timeout=10000):
self._TMP_FILES = {}
self.exe = check_executable('phantomjs', ['-v'])
if not self.exe:
raise ExtractorError('PhantomJS executable not found in PATH, '
'download it from http://phantomjs.org',
expected=True)
self.extractor = extractor
if required_version:
version = self._version()
if is_outdated_version(version, required_version):
self.extractor._downloader.report_warning(
'Your copy of PhantomJS is outdated, update it to version '
'%s or newer if you encounter any errors.' % required_version)
self.options = {
'timeout': timeout,
}
for name in self._TMP_FILE_NAMES:
tmp = tempfile.NamedTemporaryFile(delete=False)
tmp.close()
self._TMP_FILES[name] = tmp
def __del__(self):
for name in self._TMP_FILE_NAMES:
try:
os.remove(self._TMP_FILES[name].name)
except (IOError, OSError, KeyError):
pass
def _save_cookies(self, url):
cookies = cookie_jar_to_list(self.extractor._downloader.cookiejar)
for cookie in cookies:
if 'path' not in cookie:
cookie['path'] = '/'
if 'domain' not in cookie:
cookie['domain'] = compat_urlparse.urlparse(url).netloc
with open(self._TMP_FILES['cookies'].name, 'wb') as f:
f.write(json.dumps(cookies).encode('utf-8'))
def _load_cookies(self):
with open(self._TMP_FILES['cookies'].name, 'rb') as f:
cookies = json.loads(f.read().decode('utf-8'))
for cookie in cookies:
if cookie['httponly'] is True:
cookie['rest'] = {'httpOnly': None}
if 'expiry' in cookie:
cookie['expire_time'] = cookie['expiry']
self.extractor._set_cookie(**compat_kwargs(cookie))
def get(self, url, html=None, video_id=None, note=None, note2='Executing JS on webpage', headers={}, jscode='saveAndExit();'):
"""
Downloads webpage (if needed) and executes JS
Params:
url: website url
html: optional, html code of website
video_id: video id
note: optional, displayed when downloading webpage
note2: optional, displayed when executing JS
headers: custom http headers
jscode: code to be executed when page is loaded
Returns tuple with:
* downloaded website (after JS execution)
* anything you print with `console.log` (but not inside `page.execute`!)
In most cases you don't need to add any `jscode`.
It is executed in `page.onLoadFinished`.
`saveAndExit();` is mandatory, use it instead of `phantom.exit()`
It is possible to wait for some element on the webpage, for example:
var check = function() {
var elementFound = page.evaluate(function() {
return document.querySelector('#b.done') !== null;
});
if(elementFound)
saveAndExit();
else
window.setTimeout(check, 500);
}
page.evaluate(function(){
document.querySelector('#a').click();
});
check();
"""
if 'saveAndExit();' not in jscode:
raise ExtractorError('`saveAndExit();` not found in `jscode`')
if not html:
html = self.extractor._download_webpage(url, video_id, note=note, headers=headers)
with open(self._TMP_FILES['html'].name, 'wb') as f:
f.write(html.encode('utf-8'))
self._save_cookies(url)
replaces = self.options
replaces['url'] = url
user_agent = headers.get('User-Agent') or std_headers['User-Agent']
replaces['ua'] = user_agent.replace('"', '\\"')
replaces['jscode'] = jscode
for x in self._TMP_FILE_NAMES:
replaces[x] = self._TMP_FILES[x].name.replace('\\', '\\\\').replace('"', '\\"')
with open(self._TMP_FILES['script'].name, 'wb') as f:
f.write(self._TEMPLATE.format(**replaces).encode('utf-8'))
if video_id is None:
self.extractor.to_screen('%s' % (note2,))
else:
self.extractor.to_screen('%s: %s' % (video_id, note2))
p = subprocess.Popen([
self.exe, '--ssl-protocol=any',
self._TMP_FILES['script'].name
], stdout=subprocess.PIPE, stderr=subprocess.PIPE)
out, err = p.communicate()
if p.returncode != 0:
raise ExtractorError(
'Executing JS failed\n:' + encodeArgument(err))
with open(self._TMP_FILES['html'].name, 'rb') as f:
html = f.read().decode('utf-8')
self._load_cookies()
return (html, encodeArgument(out))
class OpenloadIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?(?:openload\.(?:co|io|link)|oload\.(?:tv|stream))/(?:f|embed)/(?P<id>[a-zA-Z0-9-_]+)'
_TESTS = [{
'url': 'https://openload.co/f/kUEfGclsU9o',
'md5': 'bf1c059b004ebc7a256f89408e65c36e',
'info_dict': {
'id': 'kUEfGclsU9o',
'ext': 'mp4',
'title': 'skyrim_no-audio_1080.mp4',
'thumbnail': r're:^https?://.*\.jpg$',
},
}, {
'url': 'https://openload.co/embed/rjC09fkPLYs',
'info_dict': {
'id': 'rjC09fkPLYs',
'ext': 'mp4',
'title': 'movie.mp4',
'thumbnail': r're:^https?://.*\.jpg$',
'subtitles': {
'en': [{
'ext': 'vtt',
}],
},
},
'params': {
'skip_download': True, # test subtitles only
},
}, {
'url': 'https://openload.co/embed/kUEfGclsU9o/skyrim_no-audio_1080.mp4',
'only_matching': True,
}, {
'url': 'https://openload.io/f/ZAn6oz-VZGE/',
'only_matching': True,
}, {
'url': 'https://openload.co/f/_-ztPaZtMhM/',
'only_matching': True,
}, {
# unavailable via https://openload.co/f/Sxz5sADo82g/, different layout
# for title and ext
'url': 'https://openload.co/embed/Sxz5sADo82g/',
'only_matching': True,
}, {
# unavailable via https://openload.co/embed/e-Ixz9ZR5L0/ but available
# via https://openload.co/f/e-Ixz9ZR5L0/
'url': 'https://openload.co/f/e-Ixz9ZR5L0/',
'only_matching': True,
}, {
'url': 'https://oload.tv/embed/KnG-kKZdcfY/',
'only_matching': True,
}, {
'url': 'http://www.openload.link/f/KnG-kKZdcfY',
'only_matching': True,
}, {
'url': 'https://oload.stream/f/KnG-kKZdcfY',
'only_matching': True,
}]
_USER_AGENT = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/56.0.2924.87 Safari/537.36'
@staticmethod
def _extract_urls(webpage):
return re.findall(
r'<iframe[^>]+src=["\']((?:https?://)?(?:openload\.(?:co|io)|oload\.tv)/embed/[a-zA-Z0-9-_]+)',
webpage)
def _real_extract(self, url):
video_id = self._match_id(url)
url_pattern = 'https://openload.co/%%s/%s/' % video_id
headers = {
'User-Agent': self._USER_AGENT,
}
for path in ('embed', 'f'):
page_url = url_pattern % path
last = path == 'f'
webpage = self._download_webpage(
page_url, video_id, 'Downloading %s webpage' % path,
headers=headers, fatal=last)
if not webpage:
continue
if 'File not found' in webpage or 'deleted by the owner' in webpage:
if not last:
continue
raise ExtractorError('File not found', expected=True, video_id=video_id)
break
phantom = PhantomJSwrapper(self, required_version='2.0')
webpage, _ = phantom.get(page_url, html=webpage, video_id=video_id, headers=headers)
decoded_id = (get_element_by_id('streamurl', webpage) or
get_element_by_id('streamuri', webpage))
video_url = 'https://openload.co/stream/%s?mime=true' % decoded_id
title = self._og_search_title(webpage, default=None) or self._search_regex(
r'<span[^>]+class=["\']title["\'][^>]*>([^<]+)', webpage,
'title', default=None) or self._html_search_meta(
'description', webpage, 'title', fatal=True)
entries = self._parse_html5_media_entries(page_url, webpage, video_id)
entry = entries[0] if entries else {}
subtitles = entry.get('subtitles')
info_dict = {
'id': video_id,
'title': title,
'thumbnail': entry.get('thumbnail') or self._og_search_thumbnail(webpage, default=None),
'url': video_url,
# Seems all videos have extensions in their titles
'ext': determine_ext(title, 'mp4'),
'subtitles': subtitles,
'http_headers': headers,
}
return info_dict