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 USER_AGENT = "Mozilla/5.0 (X11; Linux x86_64; rv:67.0) Gecko/20100101 Firefox/67.0"
  19
  20 MIMETYPES = {
  21     "video/mp4": "mp4",
  22     "video/x-flv": "flv",
  23     "video/3gpp": "3gp",
  24 }
  25
  26 QUALITIES = {
  27     "hd1080": 5,
  28     "hd720": 4,
  29     "large": 3,
  30     "medium": 2,
  31     "small": 1,
  32 }
  33
  34
  35 class VideoUnavailable(Exception):
  36     pass
  37
  38 class NotYouTube(Exception):
  39     pass
  40
  41 def print_form(url="", msg=""):
  42     script_url = "https://%s%s" % (os.environ["HTTP_HOST"], os.environ["REQUEST_URI"])
  43     sys.stdout.write("Content-Type: text/html\r\n\r\n")
  44     sys.stdout.write("""
  45 <!DOCTYPE html>
  46 <html>
  47 <head>
  48     <title>delx.net.au - YouTube Scraper</title>
  49     <link rel="stylesheet" type="text/css" href="/style.css">
  50     <style type="text/css">
  51         input[type="text"] {
  52             width: 100%;
  53         }
  54         .error {
  55             color: red;
  56         }
  57     </style>
  58 </head>
  59 <body>
  60     <h1>delx.net.au - YouTube Scraper</h1>
  61     {0}
  62     <form action="" method="get">
  63     <p>This page will let you easily download YouTube videos to watch offline. It
  64     will automatically grab the highest quality version.</p>
  65     <div><input type="text" name="url" value="{1}"/></div>
  66     <div><input type="submit" value="Download!"/></div>
  67     </form>
  68     <p>Tip! Use this bookmarklet: <a href="javascript:(function(){window.location='{2}?url='+escape(location);})()">YouTube Download</a>
  69     to easily download videos. Right-click the link and add it to bookmarks,
  70     then when you're looking at a YouTube page select that bookmark from your
  71     browser's bookmarks menu to download the video straight away.</p>
  72 </body>
  73 </html>
  74 """.replace("{0}", msg).replace("{1}", url).replace("{2}", script_url))
  75
  76 cookiejar = http.cookiejar.CookieJar()
  77 urlopener = urllib.request.build_opener(urllib.request.HTTPCookieProcessor(cookiejar))
  78 referrer = ""
  79
  80 def urlopen(url, offset=None):
  81     if url.startswith("//"):
  82         url = "https:" + url
  83     if not url.startswith("http://") and not url.startswith("https://"):
  84         url = "https://www.youtube.com" + url
  85
  86     global referrer
  87     req = urllib.request.Request(url)
  88     if not referrer:
  89         referrer = url
  90     else:
  91         req.add_header("Referer", referrer)
  92
  93     req.add_header("User-Agent", USER_AGENT)
  94
  95     if offset:
  96         req.add_header("Range", "bytes=%d-" % offset)
  97
  98     res = urlopener.open(req)
  99
 100     content_range = res.getheader("Content-Range")
 101     if content_range:
 102         tokens = content_range.split()
 103         assert tokens[0] == "bytes"
 104         start = int(tokens[1].split("-")[0])
 105         assert start == offset
 106     return res
 107
 108 def validate_url(url):
 109     parsed_url = urllib.parse.urlparse(url)
 110     scheme_ok = parsed_url.scheme == "https"
 111     host_ok = parsed_url.netloc.lstrip("www.") in ["youtube.com", "youtu.be"]
 112
 113     if scheme_ok and host_ok:
 114         return
 115     else:
 116         raise NotYouTube()
 117
 118 def parse_url(url, parser):
 119     f = urlopen(url)
 120     parser.feed(f.read().decode("utf-8"))
 121     parser.close()
 122     f.close()
 123
 124 def append_to_qs(url, params):
 125     r = list(urllib.parse.urlsplit(url))
 126     qs = urllib.parse.parse_qs(r[3])
 127     qs.update(params)
 128     r[3] = urllib.parse.urlencode(qs, True)
 129     url = urllib.parse.urlunsplit(r)
 130     return url
 131
 132 def get_player_config(scripts):
 133     config_strings = [
 134         ("ytplayer.config = {", 1, "};", 1),
 135         ("ytcfg.set({\"", 2, "});", 1),
 136     ]
 137     player_config = {}
 138     for script in scripts:
 139         for line in script.split("\n"):
 140             for s1, off1, s2, off2 in config_strings:
 141                 if s1 in line:
 142                     p1 = line.find(s1) + len(s1) - off1
 143                     p2 = line.find(s2, p1) + off2
 144                     if p1 >= 0 and p2 > 0:
 145                         player_config.update(json.loads(line[p1:p2]))
 146     return player_config
 147
 148 def extract_js(script):
 149     PREFIX = "var _yt_player={};(function(g){var window=this;"
 150     SUFFIX = ";})(_yt_player);\n"
 151     assert script.startswith(PREFIX)
 152     assert script.endswith(SUFFIX)
 153
 154     return script[len(PREFIX):-len(SUFFIX)]
 155
 156 def find_cipher_func(script):
 157     FUNC_NAME = R"([a-zA-Z0-9$]+)"
 158     DECODE_URI_COMPONENT = R"(\(decodeURIComponent)?"
 159     FUNC_PARAMS = R"(\([a-zA-Z,\.]+\.s\))"
 160     TERMINATOR = R"[,;\)]"
 161     PATTERN = FUNC_NAME + DECODE_URI_COMPONENT + FUNC_PARAMS + TERMINATOR
 162
 163     match = re.search(PATTERN, script)
 164     func_name = match.groups()[0]
 165     return func_name
 166
 167 def find_url_func(script):
 168     FUNC_NAME = R"([a-zA-Z0-9$]+)"
 169     PATTERN = R"this\.url\s*=\s*" + FUNC_NAME + R"\s*\(\s*this\s*\)"
 170
 171     match = re.search(PATTERN, script)
 172     func_name = match.groups()[0]
 173     return func_name
 174
 175 def decode_cipher_url(js_url, cipher):
 176     cipher = urllib.parse.parse_qs(cipher)
 177     args = [
 178         cipher["url"][0],
 179         cipher["sp"][0],
 180         cipher["s"][0],
 181     ]
 182
 183     f = urlopen(js_url)
 184     script = f.read().decode("utf-8")
 185     f.close()
 186
 187     cipher_func_name = find_cipher_func(script)
 188     url_func_name = find_url_func(script)
 189
 190     params = {
 191         "cipher_func_name": cipher_func_name,
 192         "url_func_name": url_func_name,
 193         "args": json.dumps(args),
 194         "code": json.dumps(extract_js(script)),
 195     }
 196     p = subprocess.Popen(
 197         "node",
 198         shell=True,
 199         close_fds=True,
 200         stdin=subprocess.PIPE,
 201         stdout=subprocess.PIPE
 202     )
 203     js_decode_script = ("""
 204         const vm = require('vm');
 205
 206         const fakeGlobal = {};
 207         fakeGlobal.window = fakeGlobal;
 208         fakeGlobal.location = {
 209             hash: '',
 210             host: 'www.youtube.com',
 211             hostname: 'www.youtube.com',
 212             href: 'https://www.youtube.com',
 213             origin: 'https://www.youtube.com',
 214             pathname: '/',
 215             protocol: 'https:'
 216         };
 217         fakeGlobal.history = {
 218             pushState: function(){}
 219         };
 220         fakeGlobal.document = {
 221             location: fakeGlobal.location
 222         };
 223         fakeGlobal.document = {};
 224         fakeGlobal.navigator = {
 225             userAgent: ''
 226         };
 227         fakeGlobal.XMLHttpRequest = class XMLHttpRequest {};
 228         fakeGlobal.matchMedia = () => ({matches: () => {}, media: ''});
 229         fakeGlobal.result_url = null;
 230         fakeGlobal.g = function(){}; // this is _yt_player
 231
 232         const code_string = %(code)s + ';';
 233         const exec_string = 'result_url = %(url_func_name)s(%(cipher_func_name)s(...%(args)s));';
 234         vm.runInNewContext(code_string + exec_string, fakeGlobal);
 235
 236         console.log(fakeGlobal.result_url);
 237     """ % params)
 238
 239     p.stdin.write(js_decode_script.encode("utf-8"))
 240     p.stdin.close()
 241
 242     result_url = p.stdout.read().decode("utf-8").strip()
 243     if p.wait() != 0:
 244         raise Exception("js failed to execute: %d" % p.returncode)
 245
 246     return result_url
 247
 248 def get_best_video(player_config):
 249     player_args = player_config["args"]
 250     player_response = json.loads(player_args["player_response"])
 251     formats = player_response["streamingData"]["formats"]
 252
 253     best_url = None
 254     best_quality = None
 255     best_extension = None
 256     for format_data in formats:
 257         mimetype = format_data["mimeType"].split(";")[0]
 258         quality = format_data["quality"]
 259
 260         if quality not in QUALITIES:
 261             continue
 262         if mimetype not in MIMETYPES:
 263             continue
 264
 265         extension = MIMETYPES[mimetype]
 266         quality = QUALITIES.get(quality, -1)
 267
 268         if best_quality is not None and quality < best_quality:
 269             continue
 270
 271         if "signatureCipher" in format_data:
 272             js_url = player_config["PLAYER_JS_URL"]
 273             video_url = decode_cipher_url(js_url, format_data["signatureCipher"])
 274         else:
 275             video_url = format_data["url"]
 276
 277         best_url = video_url
 278         best_quality = quality
 279         best_extension = extension
 280
 281     return best_url, best_extension
 282
 283 def sanitize_filename(filename):
 284     return (
 285         re.sub("\s+", " ", filename.strip())
 286         .replace("\\", "-")
 287         .replace("/", "-")
 288         .replace("\0", " ")
 289     )
 290
 291 def get_video_url(page):
 292     player_config = get_player_config(page.scripts)
 293     if not player_config:
 294         raise VideoUnavailable(page.unavailable_message or "Could not find video URL")
 295
 296     video_url, extension = get_best_video(player_config)
 297     if not video_url:
 298         return None, None
 299
 300     title = player_config["args"].get("title", None)
 301     if not title:
 302         title = json.loads(player_config["args"]["player_response"])["videoDetails"]["title"]
 303     if not title:
 304         title = "Unknown title"
 305
 306     filename = sanitize_filename(title) + "." + extension
 307
 308     return video_url, filename
 309
 310 class YouTubeVideoPageParser(html.parser.HTMLParser):
 311     def __init__(self):
 312         super().__init__()
 313         self.unavailable_message = None
 314         self.scripts = []
 315
 316     def handle_starttag(self, tag, attrs):
 317         attrs = dict(attrs)
 318         self._handle_unavailable_message(tag, attrs)
 319         self._handle_script(tag, attrs)
 320
 321     def handle_endtag(self, tag):
 322         self.handle_data = self._ignore_data
 323
 324     def _ignore_data(self, _):
 325         pass
 326
 327     def _handle_unavailable_message(self, tag, attrs):
 328         if attrs.get("id", None) == "unavailable-message":
 329             self.handle_data = self._handle_unavailable_message_data
 330
 331     def _handle_unavailable_message_data(self, data):
 332         self.unavailable_message = data.strip()
 333
 334     def _handle_script(self, tag, attrs):
 335         if tag == "script":
 336             self.handle_data = self._handle_script_data
 337
 338     def _handle_script_data(self, data):
 339         if data:
 340             self.scripts.append(data)
 341
 342 def write_video(filename, video_data):
 343     quoted_filename = urllib.parse.quote(filename.encode("utf-8"))
 344     sys.stdout.buffer.write(
 345         b"Content-Disposition: attachment; filename*=UTF-8''{0}\r\n"
 346         .replace(b"{0}", quoted_filename.encode("utf-8"))
 347     )
 348     sys.stdout.buffer.write(
 349         b"Content-Length: {0}\r\n"
 350         .replace(b"{0}", video_data.getheader("Content-Length").encode("utf-8"))
 351     )
 352     sys.stdout.buffer.write(b"\r\n")
 353     shutil.copyfileobj(video_data, sys.stdout.buffer)
 354     video_data.close()
 355
 356 def cgimain():
 357     args = cgi.parse()
 358     try:
 359         url = args["url"][0]
 360     except:
 361         print_form(url="https://www.youtube.com/watch?v=FOOBAR")
 362         return
 363
 364     try:
 365         page = YouTubeVideoPageParser()
 366         validate_url(url)
 367         parse_url(url, page)
 368         video_url, filename = get_video_url(page)
 369         video_data = urlopen(video_url)
 370     except VideoUnavailable as e:
 371         print_form(
 372             url=url,
 373             msg="<p class='error'>Sorry, there was an error: %s</p>" % cgi.escape(e.args[0])
 374         )
 375     except NotYouTube:
 376         print_form(
 377             url=url,
 378             msg="<p class='error'>Sorry, that does not look like a YouTube page!</p>"
 379         )
 380     except Exception as e:
 381         print_form(
 382             url=url,
 383             msg="<p class='error'>Sorry, there was an unknown error.</p>"
 384         )
 385         return
 386
 387     write_video(filename, video_data)
 388
 389 def pp_size(size):
 390     suffixes = ["", "KiB", "MiB", "GiB"]
 391     for i, suffix in enumerate(suffixes):
 392         if size < 1024:
 393             break
 394         size /= 1024
 395     return "%.2f %s" % (size, suffix)
 396
 397 def copy_with_progress(content_length, infile, outfile):
 398     def print_status():
 399         rate = 0
 400         if now != last_ts:
 401             rate = last_bytes_read / (now - last_ts)
 402         sys.stdout.write("\33[2K\r")
 403         sys.stdout.write("%s / %s (%s/sec)" % (
 404             pp_size(bytes_read),
 405             pp_size(content_length),
 406             pp_size(rate),
 407         ))
 408         sys.stdout.flush()
 409
 410     last_ts = 0
 411     last_bytes_read = 0
 412     bytes_read = 0
 413     while True:
 414         now = time.time()
 415         if now - last_ts > 0.5:
 416             print_status()
 417             last_ts = now
 418             last_bytes_read = 0
 419
 420         buf = infile.read(32768)
 421         if not buf:
 422             break
 423         outfile.write(buf)
 424         last_bytes_read += len(buf)
 425         bytes_read += len(buf)
 426
 427     # Newline at the end
 428     print_status()
 429     print()
 430
 431 def main():
 432     try:
 433         url = sys.argv[1]
 434     except:
 435         print("Usage: %s https://youtube.com/watch?v=FOOBAR" % sys.argv[0], file=sys.stderr)
 436         sys.exit(1)
 437
 438     page = YouTubeVideoPageParser()
 439     parse_url(url, page)
 440     video_url, filename = get_video_url(page)
 441     print("Downloading", filename)
 442
 443     outfile = open(filename, "ab")
 444     offset = outfile.tell()
 445     if offset > 0:
 446         print("Resuming download from", pp_size(offset))
 447     total_size = None
 448
 449     while True:
 450         try:
 451             video_data = urlopen(video_url, offset)
 452         except urllib.error.HTTPError as e:
 453             if e.code == 416:
 454                 print("File is complete!")
 455                 break
 456             else:
 457                 raise
 458
 459         content_length = int(video_data.getheader("Content-Length"))
 460         if total_size is None:
 461             total_size = content_length
 462
 463         try:
 464             copy_with_progress(content_length, video_data, outfile)
 465         except IOError as e:
 466             print()
 467
 468         video_data.close()
 469         if outfile.tell() != total_size:
 470             old_offset = offset
 471             offset = outfile.tell()
 472             if old_offset == offset:
 473                 time.sleep(1)
 474             print("Restarting download from", pp_size(offset))
 475         else:
 476             break
 477
 478     outfile.close()
 479
 480
 481 if __name__ == "__main__":
 482     if "SCRIPT_NAME" in os.environ:
 483         cgimain()
 484     else:
 485         try:
 486             main()
 487         except KeyboardInterrupt:
 488             print("\nExiting...")
 489             sys.exit(1)
 490