]> jfr.im git - yt-dlp.git/blob - yt_dlp/extractor/openload.py
[ie/orf:on] Improve extraction (#9677)
[yt-dlp.git] / yt_dlp / extractor / openload.py
1 import collections
2 import contextlib
3 import json
4 import os
5 import subprocess
6 import tempfile
7
8 from ..compat import compat_urlparse
9 from ..utils import (
10 ExtractorError,
11 Popen,
12 check_executable,
13 format_field,
14 get_exe_version,
15 is_outdated_version,
16 shell_quote,
17 )
18
19
20 def cookie_to_dict(cookie):
21 cookie_dict = {
22 'name': cookie.name,
23 'value': cookie.value,
24 }
25 if cookie.port_specified:
26 cookie_dict['port'] = cookie.port
27 if cookie.domain_specified:
28 cookie_dict['domain'] = cookie.domain
29 if cookie.path_specified:
30 cookie_dict['path'] = cookie.path
31 if cookie.expires is not None:
32 cookie_dict['expires'] = cookie.expires
33 if cookie.secure is not None:
34 cookie_dict['secure'] = cookie.secure
35 if cookie.discard is not None:
36 cookie_dict['discard'] = cookie.discard
37 with contextlib.suppress(TypeError):
38 if (cookie.has_nonstandard_attr('httpOnly')
39 or cookie.has_nonstandard_attr('httponly')
40 or cookie.has_nonstandard_attr('HttpOnly')):
41 cookie_dict['httponly'] = True
42 return cookie_dict
43
44
45 def cookie_jar_to_list(cookie_jar):
46 return [cookie_to_dict(cookie) for cookie in cookie_jar]
47
48
49 class PhantomJSwrapper:
50 """PhantomJS wrapper class
51
52 This class is experimental.
53 """
54
55 INSTALL_HINT = 'Please download it from https://phantomjs.org/download.html'
56
57 _BASE_JS = R'''
58 phantom.onError = function(msg, trace) {{
59 var msgStack = ['PHANTOM ERROR: ' + msg];
60 if(trace && trace.length) {{
61 msgStack.push('TRACE:');
62 trace.forEach(function(t) {{
63 msgStack.push(' -> ' + (t.file || t.sourceURL) + ': ' + t.line
64 + (t.function ? ' (in function ' + t.function +')' : ''));
65 }});
66 }}
67 console.error(msgStack.join('\n'));
68 phantom.exit(1);
69 }};
70 '''
71
72 _TEMPLATE = R'''
73 var page = require('webpage').create();
74 var fs = require('fs');
75 var read = {{ mode: 'r', charset: 'utf-8' }};
76 var write = {{ mode: 'w', charset: 'utf-8' }};
77 JSON.parse(fs.read("{cookies}", read)).forEach(function(x) {{
78 phantom.addCookie(x);
79 }});
80 page.settings.resourceTimeout = {timeout};
81 page.settings.userAgent = "{ua}";
82 page.onLoadStarted = function() {{
83 page.evaluate(function() {{
84 delete window._phantom;
85 delete window.callPhantom;
86 }});
87 }};
88 var saveAndExit = function() {{
89 fs.write("{html}", page.content, write);
90 fs.write("{cookies}", JSON.stringify(phantom.cookies), write);
91 phantom.exit();
92 }};
93 page.onLoadFinished = function(status) {{
94 if(page.url === "") {{
95 page.setContent(fs.read("{html}", read), "{url}");
96 }}
97 else {{
98 {jscode}
99 }}
100 }};
101 page.open("");
102 '''
103
104 _TMP_FILE_NAMES = ['script', 'html', 'cookies']
105
106 @staticmethod
107 def _version():
108 return get_exe_version('phantomjs', version_re=r'([0-9.]+)')
109
110 def __init__(self, extractor, required_version=None, timeout=10000):
111 self._TMP_FILES = {}
112
113 self.exe = check_executable('phantomjs', ['-v'])
114 if not self.exe:
115 raise ExtractorError(f'PhantomJS not found, {self.INSTALL_HINT}', expected=True)
116
117 self.extractor = extractor
118
119 if required_version:
120 version = self._version()
121 if is_outdated_version(version, required_version):
122 self.extractor._downloader.report_warning(
123 'Your copy of PhantomJS is outdated, update it to version '
124 '%s or newer if you encounter any errors.' % required_version)
125
126 for name in self._TMP_FILE_NAMES:
127 tmp = tempfile.NamedTemporaryFile(delete=False)
128 tmp.close()
129 self._TMP_FILES[name] = tmp
130
131 self.options = collections.ChainMap({
132 'timeout': timeout,
133 }, {
134 x: self._TMP_FILES[x].name.replace('\\', '\\\\').replace('"', '\\"')
135 for x in self._TMP_FILE_NAMES
136 })
137
138 def __del__(self):
139 for name in self._TMP_FILE_NAMES:
140 with contextlib.suppress(OSError, KeyError):
141 os.remove(self._TMP_FILES[name].name)
142
143 def _save_cookies(self, url):
144 cookies = cookie_jar_to_list(self.extractor.cookiejar)
145 for cookie in cookies:
146 if 'path' not in cookie:
147 cookie['path'] = '/'
148 if 'domain' not in cookie:
149 cookie['domain'] = compat_urlparse.urlparse(url).netloc
150 with open(self._TMP_FILES['cookies'].name, 'wb') as f:
151 f.write(json.dumps(cookies).encode('utf-8'))
152
153 def _load_cookies(self):
154 with open(self._TMP_FILES['cookies'].name, 'rb') as f:
155 cookies = json.loads(f.read().decode('utf-8'))
156 for cookie in cookies:
157 if cookie['httponly'] is True:
158 cookie['rest'] = {'httpOnly': None}
159 if 'expiry' in cookie:
160 cookie['expire_time'] = cookie['expiry']
161 self.extractor._set_cookie(**cookie)
162
163 def get(self, url, html=None, video_id=None, note=None, note2='Executing JS on webpage', headers={}, jscode='saveAndExit();'):
164 """
165 Downloads webpage (if needed) and executes JS
166
167 Params:
168 url: website url
169 html: optional, html code of website
170 video_id: video id
171 note: optional, displayed when downloading webpage
172 note2: optional, displayed when executing JS
173 headers: custom http headers
174 jscode: code to be executed when page is loaded
175
176 Returns tuple with:
177 * downloaded website (after JS execution)
178 * anything you print with `console.log` (but not inside `page.execute`!)
179
180 In most cases you don't need to add any `jscode`.
181 It is executed in `page.onLoadFinished`.
182 `saveAndExit();` is mandatory, use it instead of `phantom.exit()`
183 It is possible to wait for some element on the webpage, e.g.
184 var check = function() {
185 var elementFound = page.evaluate(function() {
186 return document.querySelector('#b.done') !== null;
187 });
188 if(elementFound)
189 saveAndExit();
190 else
191 window.setTimeout(check, 500);
192 }
193
194 page.evaluate(function(){
195 document.querySelector('#a').click();
196 });
197 check();
198 """
199 if 'saveAndExit();' not in jscode:
200 raise ExtractorError('`saveAndExit();` not found in `jscode`')
201 if not html:
202 html = self.extractor._download_webpage(url, video_id, note=note, headers=headers)
203 with open(self._TMP_FILES['html'].name, 'wb') as f:
204 f.write(html.encode('utf-8'))
205
206 self._save_cookies(url)
207
208 user_agent = headers.get('User-Agent') or self.extractor.get_param('http_headers')['User-Agent']
209 jscode = self._TEMPLATE.format_map(self.options.new_child({
210 'url': url,
211 'ua': user_agent.replace('"', '\\"'),
212 'jscode': jscode,
213 }))
214
215 stdout = self.execute(jscode, video_id, note=note2)
216
217 with open(self._TMP_FILES['html'].name, 'rb') as f:
218 html = f.read().decode('utf-8')
219 self._load_cookies()
220
221 return html, stdout
222
223 def execute(self, jscode, video_id=None, *, note='Executing JS'):
224 """Execute JS and return stdout"""
225 if 'phantom.exit();' not in jscode:
226 jscode += ';\nphantom.exit();'
227 jscode = self._BASE_JS + jscode
228
229 with open(self._TMP_FILES['script'].name, 'w', encoding='utf-8') as f:
230 f.write(jscode)
231 self.extractor.to_screen(f'{format_field(video_id, None, "%s: ")}{note}')
232
233 cmd = [self.exe, '--ssl-protocol=any', self._TMP_FILES['script'].name]
234 self.extractor.write_debug(f'PhantomJS command line: {shell_quote(cmd)}')
235 try:
236 stdout, stderr, returncode = Popen.run(cmd, timeout=self.options['timeout'] / 1000,
237 text=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
238 except Exception as e:
239 raise ExtractorError(f'{note} failed: Unable to run PhantomJS binary', cause=e)
240 if returncode:
241 raise ExtractorError(f'{note} failed with returncode {returncode}:\n{stderr.strip()}')
242
243 return stdout