code.delx.au - youtube-cgi/blob - youtube.cgi

   1 #!/usr/bin/env python3
   2
   3 import cgi
   4 import html.parser
   5 import http.cookiejar
   6 import json
   7 import os
   8 import re
   9 import shutil
  10 import subprocess
  11 import sys
  12 import time
  13 import urllib.error
  14 import urllib.parse
  15 import urllib.request
  16
  17
  18 MOZILLA_RELEASE_URL = "https://www.mozilla.org/en-US/firefox/releases/"
  19 USER_AGENT_TEMPLATE = "Mozilla/5.0 (X11; Linux x86_64; rv:83.0) Gecko/20100101 Firefox/%s"
  20
  21 MIMETYPES = {
  22     "video/mp4": "mp4",
  23     "video/x-flv": "flv",
  24     "video/3gpp": "3gp",
  25 }
  26
  27 QUALITIES = {
  28     "hd1080": 5,
  29     "hd720": 4,
  30     "large": 3,
  31     "medium": 2,
  32     "small": 1,
  33 }
  34
  35
  36 class VideoUnavailable(Exception):
  37     pass
  38
  39 class NotYouTube(Exception):
  40     pass
  41
  42 def print_form(url="", msg=""):
  43     script_url = "https://%s%s" % (os.environ["HTTP_HOST"], os.environ["REQUEST_URI"])
  44     sys.stdout.write("Content-Type: text/html\r\n\r\n")
  45     sys.stdout.write("""
  46 <!DOCTYPE html>
  47 <html>
  48 <head>
  49     <title>delx.net.au - YouTube Scraper</title>
  50     <link rel="stylesheet" type="text/css" href="/style.css">
  51     <style type="text/css">
  52         input[type="text"] {
  53             width: 100%;
  54         }
  55         .error {
  56             color: red;
  57         }
  58     </style>
  59 </head>
  60 <body>
  61     <h1>delx.net.au - YouTube Scraper</h1>
  62     {0}
  63     <form action="" method="get">
  64     <p>This page will let you easily download YouTube videos to watch offline. It
  65     will automatically grab the highest quality version.</p>
  66     <div><input type="text" name="url" value="{1}"/></div>
  67     <div><input type="submit" value="Download!"/></div>
  68     </form>
  69     <p>Tip! Use this bookmarklet: <a href="javascript:(function(){window.location='{2}?url='+escape(location);})()">YouTube Download</a>
  70     to easily download videos. Right-click the link and add it to bookmarks,
  71     then when you're looking at a YouTube page select that bookmark from your
  72     browser's bookmarks menu to download the video straight away.</p>
  73 </body>
  74 </html>
  75 """.replace("{0}", msg).replace("{1}", url).replace("{2}", script_url))
  76
  77 cookiejar = http.cookiejar.CookieJar()
  78 urlopener = urllib.request.build_opener(urllib.request.HTTPCookieProcessor(cookiejar))
  79 referrer = ""
  80 user_agent = None
  81
  82 def urlopen(url, offset=None):
  83     global user_agent
  84     if not user_agent:
  85         page = MozillaReleasesPageParser()
  86         with urllib.request.urlopen(MOZILLA_RELEASE_URL) as f:
  87             page.feed(f.read().decode("utf-8"))
  88             page.close()
  89         user_agent = USER_AGENT_TEMPLATE % page.latest_release
  90
  91     if url.startswith("//"):
  92         url = "https:" + url
  93     if not url.startswith("http://") and not url.startswith("https://"):
  94         url = "https://www.youtube.com" + url
  95
  96     global referrer
  97     req = urllib.request.Request(url)
  98     if not referrer:
  99         referrer = url
 100     else:
 101         req.add_header("Referer", referrer)
 102
 103     req.add_header("User-Agent", user_agent)
 104
 105     if offset:
 106         req.add_header("Range", "bytes=%d-" % offset)
 107
 108     res = urlopener.open(req)
 109
 110     content_range = res.getheader("Content-Range")
 111     if content_range:
 112         tokens = content_range.split()
 113         assert tokens[0] == "bytes"
 114         start = int(tokens[1].split("-")[0])
 115         assert start == offset
 116     return res
 117
 118 def validate_url(url):
 119     parsed_url = urllib.parse.urlparse(url)
 120     scheme_ok = parsed_url.scheme == "https"
 121     host = parsed_url.netloc.lstrip("www.").lstrip("m.")
 122     host_ok = host in ["youtube.com", "youtu.be"]
 123
 124     if scheme_ok and host_ok:
 125         return
 126     else:
 127         raise NotYouTube()
 128
 129 def load_parse_url(url, parser):
 130     f = urlopen(url)
 131     parser.feed(f.read().decode("utf-8"))
 132     parser.close()
 133     f.close()
 134
 135 def append_to_qs(url, params):
 136     r = list(urllib.parse.urlsplit(url))
 137     qs = urllib.parse.parse_qs(r[3])
 138     qs.update(params)
 139     r[3] = urllib.parse.urlencode(qs, True)
 140     url = urllib.parse.urlunsplit(r)
 141     return url
 142
 143 def get_player_config(scripts):
 144     config_strings = [
 145         ("ytcfg.set({\"", 2, "});", 1),
 146         ("ytInitialPlayerResponse = {\"", 2, "};", 1),
 147     ]
 148     player_config = {}
 149     for script in scripts:
 150         for line in script.split("\n"):
 151             for s1, off1, s2, off2 in config_strings:
 152                 if s1 in line:
 153                     p1 = line.find(s1) + len(s1) - off1
 154                     p2 = line.find(s2, p1) + off2
 155                     if p1 >= 0 and p2 > 0:
 156                         player_config.update(json.loads(line[p1:p2]))
 157     return player_config
 158
 159 def extract_js(script):
 160     PREFIX = "var _yt_player={};(function(g){var window=this;"
 161     SUFFIX = ";})(_yt_player);\n"
 162     assert script.startswith(PREFIX)
 163     assert script.endswith(SUFFIX)
 164
 165     return script[len(PREFIX):-len(SUFFIX)]
 166
 167 def find_cipher_func(script):
 168     FUNC_NAME = R"([a-zA-Z0-9$]+)"
 169     DECODE_URI_COMPONENT = R"(\(decodeURIComponent)?"
 170     FUNC_PARAMS = R"(\([a-zA-Z,\.]+\.s\))"
 171     TERMINATOR = R"[,;\)]"
 172     PATTERN = FUNC_NAME + DECODE_URI_COMPONENT + FUNC_PARAMS + TERMINATOR
 173
 174     match = re.search(PATTERN, script)
 175     func_name = match.groups()[0]
 176     return func_name
 177
 178 def construct_url_from_cipher_result(cipher_result):
 179     for k, v in cipher_result.items():
 180         if isinstance(v, str) and v.startswith("https://"):
 181             temp_url = v
 182             break
 183     else:
 184         raise Exception("Could not find URL-like string in cipher result!")
 185
 186     for k, v in cipher_result.items():
 187         if isinstance(v, dict):
 188             params = {}
 189             for k2, v2 in v.items():
 190                 params[k2] = urllib.parse.unquote(v2)
 191             return append_to_qs(temp_url, params)
 192     else:
 193         raise Exception("Could not find params-like structure in cipher result!")
 194
 195 def decode_cipher_url(js_url, cipher):
 196     cipher = urllib.parse.parse_qs(cipher)
 197     args = [
 198         cipher["url"][0],
 199         cipher["sp"][0],
 200         cipher["s"][0],
 201     ]
 202
 203     f = urlopen(js_url)
 204     script = f.read().decode("utf-8")
 205     f.close()
 206
 207     cipher_func_name = find_cipher_func(script)
 208
 209     params = {
 210         "cipher_func_name": cipher_func_name,
 211         "args": json.dumps(args),
 212         "code": json.dumps(extract_js(script)),
 213     }
 214     p = subprocess.Popen(
 215         "node",
 216         shell=True,
 217         close_fds=True,
 218         stdin=subprocess.PIPE,
 219         stdout=subprocess.PIPE
 220     )
 221     js_decode_script = ("""
 222         const vm = require('vm');
 223
 224         const fakeGlobal = {};
 225         fakeGlobal.window = fakeGlobal;
 226         fakeGlobal.location = {
 227             hash: '',
 228             host: 'www.youtube.com',
 229             hostname: 'www.youtube.com',
 230             href: 'https://www.youtube.com',
 231             origin: 'https://www.youtube.com',
 232             pathname: '/',
 233             protocol: 'https:'
 234         };
 235         fakeGlobal.history = {
 236             pushState: function(){}
 237         };
 238         fakeGlobal.document = {
 239             location: fakeGlobal.location
 240         };
 241         fakeGlobal.document = {};
 242         fakeGlobal.navigator = {
 243             userAgent: ''
 244         };
 245         fakeGlobal.XMLHttpRequest = class XMLHttpRequest {};
 246         fakeGlobal.matchMedia = () => ({matches: () => {}, media: ''});
 247         fakeGlobal.result = null;
 248         fakeGlobal.g = function(){}; // this is _yt_player
 249         fakeGlobal.TimeRanges = function(){};
 250
 251         const code_string = %(code)s + ';';
 252         const exec_string = 'result = %(cipher_func_name)s(...%(args)s);';
 253         vm.runInNewContext(code_string + exec_string, fakeGlobal);
 254
 255         console.log(JSON.stringify(fakeGlobal.result));
 256     """ % params)
 257
 258     p.stdin.write(js_decode_script.encode("utf-8"))
 259     p.stdin.close()
 260
 261     result = json.load(p.stdout)
 262     if p.wait() != 0:
 263         raise Exception("js failed to execute: %d" % p.returncode)
 264
 265     result_url = construct_url_from_cipher_result(result)
 266     return result_url
 267
 268 def get_best_video(player_config):
 269     formats = player_config["streamingData"]["formats"]
 270
 271     best_url = None
 272     best_quality = None
 273     best_extension = None
 274     for format_data in formats:
 275         mimetype = format_data["mimeType"].split(";")[0]
 276         quality = format_data["quality"]
 277
 278         if quality not in QUALITIES:
 279             continue
 280         if mimetype not in MIMETYPES:
 281             continue
 282
 283         extension = MIMETYPES[mimetype]
 284         quality = QUALITIES.get(quality, -1)
 285
 286         if best_quality is not None and quality < best_quality:
 287             continue
 288
 289         if "signatureCipher" in format_data:
 290             js_url = player_config["PLAYER_JS_URL"]
 291             video_url = decode_cipher_url(js_url, format_data["signatureCipher"])
 292         else:
 293             video_url = format_data["url"]
 294
 295         best_url = video_url
 296         best_quality = quality
 297         best_extension = extension
 298
 299     return best_url, best_extension
 300
 301 def sanitize_filename(filename):
 302     return (
 303         re.sub("\s+", " ", filename.strip())
 304         .replace("\\", "-")
 305         .replace("/", "-")
 306         .replace("\0", " ")
 307     )
 308
 309 def get_video_url(page):
 310     player_config = get_player_config(page.scripts)
 311     if not player_config:
 312         raise VideoUnavailable(page.unavailable_message or "Could not find video URL")
 313
 314     video_url, extension = get_best_video(player_config)
 315     if not video_url:
 316         return None, None
 317
 318     title = player_config["videoDetails"].get("title", None)
 319     if not title:
 320         title = "Unknown title"
 321
 322     filename = sanitize_filename(title) + "." + extension
 323
 324     return video_url, filename
 325
 326 class YouTubeVideoPageParser(html.parser.HTMLParser):
 327     def __init__(self):
 328         super().__init__()
 329         self.unavailable_message = None
 330         self.scripts = []
 331
 332     def handle_starttag(self, tag, attrs):
 333         attrs = dict(attrs)
 334         self._handle_unavailable_message(tag, attrs)
 335         self._handle_script(tag, attrs)
 336
 337     def handle_endtag(self, tag):
 338         self.handle_data = self._ignore_data
 339
 340     def _ignore_data(self, _):
 341         pass
 342
 343     def _handle_unavailable_message(self, tag, attrs):
 344         if attrs.get("id", None) == "unavailable-message":
 345             self.handle_data = self._handle_unavailable_message_data
 346
 347     def _handle_unavailable_message_data(self, data):
 348         self.unavailable_message = data.strip()
 349
 350     def _handle_script(self, tag, attrs):
 351         if tag == "script":
 352             self.handle_data = self._handle_script_data
 353
 354     def _handle_script_data(self, data):
 355         if data:
 356             self.scripts.append(data)
 357
 358 class MozillaReleasesPageParser(html.parser.HTMLParser):
 359     def __init__(self):
 360         super().__init__()
 361         self.latest_release = "1.0"
 362
 363     def handle_starttag(self, tag, attrs):
 364         attrs = dict(attrs)
 365         if attrs.get("data-latest-firefox", None):
 366             self.latest_release = attrs.get("data-latest-firefox", None)
 367
 368 def write_video(filename, video_data):
 369     quoted_filename = urllib.parse.quote(filename.encode("utf-8"))
 370     sys.stdout.buffer.write(
 371         b"Content-Disposition: attachment; filename*=UTF-8''{0}\r\n"
 372         .replace(b"{0}", quoted_filename.encode("utf-8"))
 373     )
 374     sys.stdout.buffer.write(
 375         b"Content-Length: {0}\r\n"
 376         .replace(b"{0}", video_data.getheader("Content-Length").encode("utf-8"))
 377     )
 378     sys.stdout.buffer.write(b"\r\n")
 379     shutil.copyfileobj(video_data, sys.stdout.buffer)
 380     video_data.close()
 381
 382 def cgimain():
 383     args = cgi.parse()
 384     try:
 385         url = args["url"][0]
 386     except:
 387         print_form(url="https://www.youtube.com/watch?v=FOOBAR")
 388         return
 389
 390     try:
 391         page = YouTubeVideoPageParser()
 392         validate_url(url)
 393         with urlopen(url) as f:
 394             page.feed(f.read().decode("utf-8"))
 395             page.close()
 396         video_url, filename = get_video_url(page)
 397         video_data = urlopen(video_url)
 398     except VideoUnavailable as e:
 399         print_form(
 400             url=url,
 401             msg="<p class='error'>Sorry, there was an error: %s</p>" % cgi.escape(e.args[0])
 402         )
 403     except NotYouTube:
 404         print_form(
 405             url=url,
 406             msg="<p class='error'>Sorry, that does not look like a YouTube page!</p>"
 407         )
 408     except Exception as e:
 409         print_form(
 410             url=url,
 411             msg="<p class='error'>Sorry, there was an unknown error.</p>"
 412         )
 413         return
 414
 415     write_video(filename, video_data)
 416
 417 def pp_size(size):
 418     suffixes = ["", "KiB", "MiB", "GiB"]
 419     for i, suffix in enumerate(suffixes):
 420         if size < 1024:
 421             break
 422         size /= 1024
 423     return "%.2f %s" % (size, suffix)
 424
 425 def copy_with_progress(content_length, infile, outfile):
 426     def print_status():
 427         rate = 0
 428         if now != last_ts:
 429             rate = last_bytes_read / (now - last_ts)
 430         sys.stdout.write("\33[2K\r")
 431         sys.stdout.write("%s / %s (%s/sec)" % (
 432             pp_size(bytes_read),
 433             pp_size(content_length),
 434             pp_size(rate),
 435         ))
 436         sys.stdout.flush()
 437
 438     last_ts = 0
 439     last_bytes_read = 0
 440     bytes_read = 0
 441     while True:
 442         now = time.time()
 443         if now - last_ts > 0.5:
 444             print_status()
 445             last_ts = now
 446             last_bytes_read = 0
 447
 448         buf = infile.read(32768)
 449         if not buf:
 450             break
 451         outfile.write(buf)
 452         last_bytes_read += len(buf)
 453         bytes_read += len(buf)
 454
 455     # Newline at the end
 456     print_status()
 457     print()
 458
 459 def main():
 460     try:
 461         url = sys.argv[1]
 462     except:
 463         print("Usage: %s https://youtube.com/watch?v=FOOBAR" % sys.argv[0], file=sys.stderr)
 464         sys.exit(1)
 465
 466     page = YouTubeVideoPageParser()
 467     with urlopen(url) as f:
 468         page.feed(f.read().decode("utf-8"))
 469         page.close()
 470     video_url, filename = get_video_url(page)
 471     print("Downloading", filename)
 472
 473     outfile = open(filename, "ab")
 474     offset = outfile.tell()
 475     if offset > 0:
 476         print("Resuming download from", pp_size(offset))
 477     total_size = None
 478
 479     while True:
 480         try:
 481             video_data = urlopen(video_url, offset)
 482         except urllib.error.HTTPError as e:
 483             if e.code == 416:
 484                 print("File is complete!")
 485                 break
 486             else:
 487                 raise
 488
 489         content_length = int(video_data.getheader("Content-Length"))
 490         if total_size is None:
 491             total_size = content_length
 492
 493         try:
 494             copy_with_progress(content_length, video_data, outfile)
 495         except IOError as e:
 496             print()
 497
 498         video_data.close()
 499         if outfile.tell() != total_size:
 500             old_offset = offset
 501             offset = outfile.tell()
 502             if old_offset == offset:
 503                 time.sleep(1)
 504             print("Restarting download from", pp_size(offset))
 505         else:
 506             break
 507
 508     outfile.close()
 509
 510
 511 if __name__ == "__main__":
 512     if "SCRIPT_NAME" in os.environ:
 513         cgimain()
 514     else:
 515         try:
 516             main()
 517         except KeyboardInterrupt:
 518             print("\nExiting...")
 519             sys.exit(1)
 520