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     player_config = None
 134     for script in scripts:
 135         for line in script.split("\n"):
 136             s = "ytplayer.config = {"
 137             if s in line:
 138                 p1 = line.find(s) + len(s) - 1
 139                 p2 = line.find("};", p1) + 1
 140                 if p1 >= 0 and p2 > 0:
 141                     return json.loads(line[p1:p2])
 142
 143 def extract_js(script):
 144     PREFIX = "var _yt_player={};(function(g){var window=this;"
 145     SUFFIX = ";})(_yt_player);\n"
 146     assert script.startswith(PREFIX)
 147     assert script.endswith(SUFFIX)
 148
 149     return script[len(PREFIX):-len(SUFFIX)]
 150
 151 def find_cipher_func(script):
 152     FUNC_NAME = R"([a-zA-Z0-9$]+)"
 153     DECODE_URI_COMPONENT = R"(\(decodeURIComponent)?"
 154     FUNC_PARAMS = R"(\([a-zA-Z,\.]+\.s\))"
 155     TERMINATOR = R"[,;\)]"
 156     PATTERN = FUNC_NAME + DECODE_URI_COMPONENT + FUNC_PARAMS + TERMINATOR
 157
 158     match = re.search(PATTERN, script)
 159     func_name = match.groups()[0]
 160     return func_name
 161
 162 def find_url_func(script):
 163     FUNC_NAME = R"([a-zA-Z0-9$]+)"
 164     PATTERN = R"this\.url\s*=\s*" + FUNC_NAME + R"\s*\(\s*this\s*\)"
 165
 166     match = re.search(PATTERN, script)
 167     func_name = match.groups()[0]
 168     return func_name
 169
 170 def decode_cipher_url(js_url, cipher):
 171     cipher = urllib.parse.parse_qs(cipher)
 172     args = [
 173         cipher["url"][0],
 174         cipher["sp"][0],
 175         cipher["s"][0],
 176     ]
 177
 178     f = urlopen(js_url)
 179     script = f.read().decode("utf-8")
 180     f.close()
 181
 182     cipher_func_name = find_cipher_func(script)
 183     url_func_name = find_url_func(script)
 184
 185     params = {
 186         "cipher_func_name": cipher_func_name,
 187         "url_func_name": url_func_name,
 188         "args": json.dumps(args),
 189         "code": json.dumps(extract_js(script)),
 190     }
 191     p = subprocess.Popen(
 192         "node",
 193         shell=True,
 194         close_fds=True,
 195         stdin=subprocess.PIPE,
 196         stdout=subprocess.PIPE
 197     )
 198     js_decode_script = ("""
 199         const vm = require('vm');
 200
 201         const fakeGlobal = {};
 202         fakeGlobal.window = fakeGlobal;
 203         fakeGlobal.location = {
 204             hash: '',
 205             host: 'www.youtube.com',
 206             hostname: 'www.youtube.com',
 207             href: 'https://www.youtube.com',
 208             origin: 'https://www.youtube.com',
 209             pathname: '/',
 210             protocol: 'https:'
 211         };
 212         fakeGlobal.history = {
 213             pushState: function(){}
 214         };
 215         fakeGlobal.document = {
 216             location: fakeGlobal.location
 217         };
 218         fakeGlobal.document = {};
 219         fakeGlobal.navigator = {
 220             userAgent: ''
 221         };
 222         fakeGlobal.XMLHttpRequest = class XMLHttpRequest {};
 223         fakeGlobal.matchMedia = () => ({matches: () => {}, media: ''});
 224         fakeGlobal.result_url = null;
 225         fakeGlobal.g = function(){}; // this is _yt_player
 226
 227         const code_string = %(code)s + ';';
 228         const exec_string = 'result_url = %(url_func_name)s(%(cipher_func_name)s(...%(args)s));';
 229         vm.runInNewContext(code_string + exec_string, fakeGlobal);
 230
 231         console.log(fakeGlobal.result_url);
 232     """ % params)
 233
 234     p.stdin.write(js_decode_script.encode("utf-8"))
 235     p.stdin.close()
 236
 237     result_url = p.stdout.read().decode("utf-8").strip()
 238     if p.wait() != 0:
 239         raise Exception("js failed to execute: %d" % p.returncode)
 240
 241     return result_url
 242
 243 def get_best_video(player_config):
 244     js_url = player_config["assets"]["js"]
 245
 246     player_args = player_config["args"]
 247     player_response = json.loads(player_args["player_response"])
 248     formats = player_response["streamingData"]["formats"]
 249
 250     best_url = None
 251     best_quality = None
 252     best_extension = None
 253     for format_data in formats:
 254         mimetype = format_data["mimeType"].split(";")[0]
 255         quality = format_data["quality"]
 256
 257         if quality not in QUALITIES:
 258             continue
 259         if mimetype not in MIMETYPES:
 260             continue
 261
 262         extension = MIMETYPES[mimetype]
 263         quality = QUALITIES.get(quality, -1)
 264
 265         if best_quality is not None and quality < best_quality:
 266             continue
 267
 268         if "signatureCipher" in format_data:
 269             video_url = decode_cipher_url(js_url, format_data["signatureCipher"])
 270         else:
 271             video_url = format_data["url"]
 272
 273         best_url = video_url
 274         best_quality = quality
 275         best_extension = extension
 276
 277     return best_url, best_extension
 278
 279 def sanitize_filename(filename):
 280     return (
 281         re.sub("\s+", " ", filename.strip())
 282         .replace("\\", "-")
 283         .replace("/", "-")
 284         .replace("\0", " ")
 285     )
 286
 287 def get_video_url(page):
 288     player_config = get_player_config(page.scripts)
 289     if not player_config:
 290         raise VideoUnavailable(page.unavailable_message or "Could not find video URL")
 291
 292     video_url, extension = get_best_video(player_config)
 293     if not video_url:
 294         return None, None
 295
 296     title = player_config["args"].get("title", None)
 297     if not title:
 298         title = json.loads(player_config["args"]["player_response"])["videoDetails"]["title"]
 299     if not title:
 300         title = "Unknown title"
 301
 302     filename = sanitize_filename(title) + "." + extension
 303
 304     return video_url, filename
 305
 306 class YouTubeVideoPageParser(html.parser.HTMLParser):
 307     def __init__(self):
 308         super().__init__()
 309         self.unavailable_message = None
 310         self.scripts = []
 311
 312     def handle_starttag(self, tag, attrs):
 313         attrs = dict(attrs)
 314         self._handle_unavailable_message(tag, attrs)
 315         self._handle_script(tag, attrs)
 316
 317     def handle_endtag(self, tag):
 318         self.handle_data = self._ignore_data
 319
 320     def _ignore_data(self, _):
 321         pass
 322
 323     def _handle_unavailable_message(self, tag, attrs):
 324         if attrs.get("id", None) == "unavailable-message":
 325             self.handle_data = self._handle_unavailable_message_data
 326
 327     def _handle_unavailable_message_data(self, data):
 328         self.unavailable_message = data.strip()
 329
 330     def _handle_script(self, tag, attrs):
 331         if tag == "script":
 332             self.handle_data = self._handle_script_data
 333
 334     def _handle_script_data(self, data):
 335         if data:
 336             self.scripts.append(data)
 337
 338 def write_video(filename, video_data):
 339     quoted_filename = urllib.parse.quote(filename.encode("utf-8"))
 340     sys.stdout.buffer.write(
 341         b"Content-Disposition: attachment; filename*=UTF-8''{0}\r\n"
 342         .replace(b"{0}", quoted_filename.encode("utf-8"))
 343     )
 344     sys.stdout.buffer.write(
 345         b"Content-Length: {0}\r\n"
 346         .replace(b"{0}", video_data.getheader("Content-Length").encode("utf-8"))
 347     )
 348     sys.stdout.buffer.write(b"\r\n")
 349     shutil.copyfileobj(video_data, sys.stdout.buffer)
 350     video_data.close()
 351
 352 def cgimain():
 353     args = cgi.parse()
 354     try:
 355         url = args["url"][0]
 356     except:
 357         print_form(url="https://www.youtube.com/watch?v=FOOBAR")
 358         return
 359
 360     try:
 361         page = YouTubeVideoPageParser()
 362         validate_url(url)
 363         parse_url(url, page)
 364         video_url, filename = get_video_url(page)
 365         video_data = urlopen(video_url)
 366     except VideoUnavailable as e:
 367         print_form(
 368             url=url,
 369             msg="<p class='error'>Sorry, there was an error: %s</p>" % cgi.escape(e.args[0])
 370         )
 371     except NotYouTube:
 372         print_form(
 373             url=url,
 374             msg="<p class='error'>Sorry, that does not look like a YouTube page!</p>"
 375         )
 376     except Exception as e:
 377         print_form(
 378             url=url,
 379             msg="<p class='error'>Sorry, there was an unknown error.</p>"
 380         )
 381         return
 382
 383     write_video(filename, video_data)
 384
 385 def pp_size(size):
 386     suffixes = ["", "KiB", "MiB", "GiB"]
 387     for i, suffix in enumerate(suffixes):
 388         if size < 1024:
 389             break
 390         size /= 1024
 391     return "%.2f %s" % (size, suffix)
 392
 393 def copy_with_progress(content_length, infile, outfile):
 394     def print_status():
 395         rate = 0
 396         if now != last_ts:
 397             rate = last_bytes_read / (now - last_ts)
 398         sys.stdout.write("\33[2K\r")
 399         sys.stdout.write("%s / %s (%s/sec)" % (
 400             pp_size(bytes_read),
 401             pp_size(content_length),
 402             pp_size(rate),
 403         ))
 404         sys.stdout.flush()
 405
 406     last_ts = 0
 407     last_bytes_read = 0
 408     bytes_read = 0
 409     while True:
 410         now = time.time()
 411         if now - last_ts > 0.5:
 412             print_status()
 413             last_ts = now
 414             last_bytes_read = 0
 415
 416         buf = infile.read(32768)
 417         if not buf:
 418             break
 419         outfile.write(buf)
 420         last_bytes_read += len(buf)
 421         bytes_read += len(buf)
 422
 423     # Newline at the end
 424     print_status()
 425     print()
 426
 427 def main():
 428     try:
 429         url = sys.argv[1]
 430     except:
 431         print("Usage: %s https://youtube.com/watch?v=FOOBAR" % sys.argv[0], file=sys.stderr)
 432         sys.exit(1)
 433
 434     page = YouTubeVideoPageParser()
 435     parse_url(url, page)
 436     video_url, filename = get_video_url(page)
 437     print("Downloading", filename)
 438
 439     outfile = open(filename, "ab")
 440     offset = outfile.tell()
 441     if offset > 0:
 442         print("Resuming download from", pp_size(offset))
 443     total_size = None
 444
 445     while True:
 446         try:
 447             video_data = urlopen(video_url, offset)
 448         except urllib.error.HTTPError as e:
 449             if e.code == 416:
 450                 print("File is complete!")
 451                 break
 452             else:
 453                 raise
 454
 455         content_length = int(video_data.getheader("Content-Length"))
 456         if total_size is None:
 457             total_size = content_length
 458
 459         try:
 460             copy_with_progress(content_length, video_data, outfile)
 461         except IOError as e:
 462             print()
 463
 464         video_data.close()
 465         if outfile.tell() != total_size:
 466             old_offset = offset
 467             offset = outfile.tell()
 468             if old_offset == offset:
 469                 time.sleep(1)
 470             print("Restarting download from", pp_size(offset))
 471         else:
 472             break
 473
 474     outfile.close()
 475
 476
 477 if __name__ == "__main__":
 478     if "SCRIPT_NAME" in os.environ:
 479         cgimain()
 480     else:
 481         try:
 482             main()
 483         except KeyboardInterrupt:
 484             print("\nExiting...")
 485             sys.exit(1)
 486