Raphaël G. Git Repositories - youtubedl/blob - youtube-dl

   1 #!/usr/bin/env python
   2 #
   3 # Copyright (c) 2006-2007 Ricardo Garcia Gonzalez
   4 #
   5 # Permission is hereby granted, free of charge, to any person obtaining a
   6 # copy of this software and associated documentation files (the "Software"),
   7 # to deal in the Software without restriction, including without limitation
   8 # the rights to use, copy, modify, merge, publish, distribute, sublicense,
   9 # and/or sell copies of the Software, and to permit persons to whom the
  10 # Software is furnished to do so, subject to the following conditions:
  11 #
  12 # The above copyright notice and this permission notice shall be included
  13 # in all copies or substantial portions of the Software.
  14 #
  15 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  16 # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  17 # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
  18 # THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
  19 # OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
  20 # ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  21 # OTHER DEALINGS IN THE SOFTWARE.
  22 #
  23 # Except as contained in this notice, the name(s) of the above copyright
  24 # holders shall not be used in advertising or otherwise to promote the
  25 # sale, use or other dealings in this Software without prior written
  26 # authorization.
  27 #
  28 import getpass
  29 import httplib
  30 import math
  31 import netrc
  32 import optparse
  33 import os
  34 import re
  35 import socket
  36 import string
  37 import sys
  38 import time
  39 import urllib2
  40
  41 # Global constants
  42 const_video_url_str = 'http://www.youtube.com/watch?v=%s'
  43 const_video_url_re = re.compile(r'^((?:http://)?(?:\w+\.)?youtube\.com/(?:v/|(?:watch(?:\.php)?)?\?(?:.+&)?v=))?([0-9A-Za-z_-]+)(?(1)[&/].*)?$')
  44 const_login_url_str = 'http://www.youtube.com/login?next=/watch%%3Fv%%3D%s'
  45 const_login_post_str = 'current_form=loginForm&next=%%2Fwatch%%3Fv%%3D%s&username=%s&password=%s&action_login=Log+In'
  46 const_age_url_str = 'http://www.youtube.com/verify_age?next_url=/watch%%3Fv%%3D%s'
  47 const_age_post_str = 'next_url=%%2Fwatch%%3Fv%%3D%s&action_confirm=Confirm'
  48 const_url_t_param_re = re.compile(r'[,{]t:\'([^\']*)\'')
  49 const_video_url_real_str = 'http://www.youtube.com/get_video?video_id=%s&t=%s'
  50 const_video_title_re = re.compile(r'<title>YouTube - ([^<]*)</title>', re.M | re.I)
  51 const_1k = 1024
  52 const_initial_block_size = 10 * const_1k
  53
  54 # Print error message, followed by standard advice information, and then exit
  55 def error_advice_exit(error_text):
  56         sys.stderr.write('Error: %s.\n' % error_text)
  57         sys.stderr.write('Try again several times. It may be a temporary problem.\n')
  58         sys.stderr.write('Other typical problems:\n\n')
  59         sys.stderr.write('* Video no longer exists.\n')
  60         sys.stderr.write('* Video requires age confirmation but you did not provide an account.\n')
  61         sys.stderr.write('* You provided the account data, but it is not valid.\n')
  62         sys.stderr.write('* The connection was cut suddenly for some reason.\n')
  63         sys.stderr.write('* YouTube changed their system, and the program no longer works.\n')
  64         sys.stderr.write('\nTry to confirm you are able to view the video using a web browser.\n')
  65         sys.stderr.write('Use the same video URL and account information, if needed, with this program.\n')
  66         sys.stderr.write('When using a proxy, make sure http_proxy has http://host:port format.\n')
  67         sys.stderr.write('Try again several times and contact me if the problem persists.\n')
  68         sys.exit('\n')
  69
  70 # Wrapper to create custom requests with typical headers
  71 def request_create(url, data=None):
  72         retval = urllib2.Request(url)
  73         if data is not None:
  74                 retval.add_data(data)
  75         # Try to mimic Firefox, at least a little bit
  76         retval.add_header('User-Agent', 'Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1) Gecko/20061010 Firefox/2.0')
  77         retval.add_header('Accept-Charset', 'ISO-8859-1,utf-8;q=0.7,*;q=0.7')
  78         retval.add_header('Accept', 'text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5')
  79         retval.add_header('Accept-Language', 'en-us,en;q=0.5')
  80         return retval
  81
  82 # Perform a request, process headers and return response
  83 def perform_request(url, data=None):
  84         request = request_create(url, data)
  85         response = urllib2.urlopen(request)
  86         return response
  87
  88 # Conditional print
  89 def cond_print(str):
  90         global cmdl_opts
  91         if not (cmdl_opts.quiet or cmdl_opts.get_url):
  92                 sys.stdout.write(str)
  93                 sys.stdout.flush()
  94
  95 # Title string normalization
  96 def title_string_norm(title):
  97         title = ''.join((x in string.ascii_letters or x in string.digits) and x or ' ' for x in title)
  98         title = '_'.join(title.split())
  99         title = title.lower()
 100         return title
 101
 102 # Title string minimal transformation
 103 def title_string_touch(title):
 104         return title.replace(os.sep, '%')
 105
 106 # Generic download step
 107 def download_step(return_data_flag, step_title, step_error, url, post_data=None):
 108         try:
 109                 cond_print('%s... ' % step_title)
 110                 data = perform_request(url, post_data).read()
 111                 cond_print('done.\n')
 112                 if return_data_flag:
 113                         return data
 114                 return None
 115
 116         except (urllib2.URLError, ValueError, httplib.HTTPException, TypeError, socket.error):
 117                 cond_print('failed.\n')
 118                 error_advice_exit(step_error)
 119
 120         except KeyboardInterrupt:
 121                 sys.exit('\n')
 122
 123 # Generic extract step
 124 def extract_step(step_title, step_error, regexp, data):
 125         try:
 126                 cond_print('%s... ' % step_title)
 127                 match = regexp.search(data)
 128
 129                 if match is None:
 130                         cond_print('failed.\n')
 131                         error_advice_exit(step_error)
 132
 133                 extracted_data = match.group(1)
 134                 cond_print('done.\n')
 135                 return extracted_data
 136
 137         except KeyboardInterrupt:
 138                 sys.exit('\n')
 139
 140 # Calculate new block size based on previous block size
 141 def new_block_size(before, after, bytes):
 142         new_min = max(bytes / 2.0, 1.0)
 143         new_max = max(bytes * 2.0, 1.0)
 144         dif = after - before
 145         if dif < 0.0001:
 146                 return int(new_max)
 147         rate = bytes / dif
 148         if rate > new_max:
 149                 return int(new_max)
 150         if rate < new_min:
 151                 return int(new_min)
 152         return int(rate)
 153
 154 # Get optimum 1k exponent to represent a number of bytes
 155 def optimum_k_exp(num_bytes):
 156         global const_1k
 157         if num_bytes == 0:
 158                 return 0
 159         return long(math.log(num_bytes, const_1k))
 160
 161 # Get optimum representation of number of bytes
 162 def format_bytes(num_bytes):
 163         global const_1k
 164         try:
 165                 exp = optimum_k_exp(num_bytes)
 166                 suffix = 'bkMGTPEZY'[exp]
 167                 if exp == 0:
 168                         return '%s%s' % (num_bytes, suffix)
 169                 converted = float(num_bytes) / float(const_1k**exp)
 170                 return '%.2f%s' % (converted, suffix)
 171         except IndexError:
 172                 sys.exit('Error: internal error formatting number of bytes.')
 173
 174 # Calculate ETA and return it in string format as MM:SS
 175 def calc_eta(start, now, total, current):
 176         if current == 0:
 177                 return '--:--'
 178         rate = float(current) / (now - start)
 179         eta = long((total - current) / rate)
 180         eta_mins = eta / 60
 181         eta_secs = eta % 60
 182         if eta_mins > 99:
 183                 return '--:--'
 184         return '%02d:%02d' % (eta_mins, eta_secs)
 185
 186 # Calculate speed and return it in string format
 187 def calc_speed(start, now, bytes):
 188         if bytes == 0:
 189                 return 'N/A b'
 190         return format_bytes(float(bytes) / (now - start))
 191
 192 # Create the command line options parser and parse command line
 193 cmdl_usage = 'usage: %prog [options] video_url'
 194 cmdl_version = '2007.08.24'
 195 cmdl_parser = optparse.OptionParser(usage=cmdl_usage, version=cmdl_version, conflict_handler='resolve')
 196 cmdl_parser.add_option('-h', '--help', action='help', help='print this help text and exit')
 197 cmdl_parser.add_option('-v', '--version', action='version', help='print program version and exit')
 198 cmdl_parser.add_option('-u', '--username', dest='username', metavar='USERNAME', help='account username')
 199 cmdl_parser.add_option('-p', '--password', dest='password', metavar='PASSWORD', help='account password')
 200 cmdl_parser.add_option('-o', '--output', dest='outfile', metavar='FILE', help='output video file name')
 201 cmdl_parser.add_option('-q', '--quiet', action='store_true', dest='quiet', help='activates quiet mode')
 202 cmdl_parser.add_option('-s', '--simulate', action='store_true', dest='simulate', help='do not download video')
 203 cmdl_parser.add_option('-t', '--title', action='store_true', dest='use_title', help='use title in file name')
 204 cmdl_parser.add_option('-l', '--literal', action='store_true', dest='use_literal', help='use literal title in file name')
 205 cmdl_parser.add_option('-n', '--netrc', action='store_true', dest='use_netrc', help='use .netrc authentication data')
 206 cmdl_parser.add_option('-g', '--get-url', action='store_true', dest='get_url', help='print final video URL only')
 207 cmdl_parser.add_option('-2', '--title-too', action='store_true', dest='get_title', help='used with -g, print title too')
 208 (cmdl_opts, cmdl_args) = cmdl_parser.parse_args()
 209
 210 # Get video URL
 211 if len(cmdl_args) != 1:
 212         cmdl_parser.print_help()
 213         sys.exit('\n')
 214 video_url_cmdl = cmdl_args[0]
 215
 216 # Verify video URL format and convert to "standard" format
 217 video_url_mo = const_video_url_re.match(video_url_cmdl)
 218 if video_url_mo is None:
 219         sys.exit('Error: URL does not seem to be a youtube video URL. If it is, report a bug.')
 220 video_url_id = video_url_mo.group(2)
 221 video_url = const_video_url_str % video_url_id
 222
 223 # Check conflicting options
 224 if cmdl_opts.outfile is not None and (cmdl_opts.simulate or cmdl_opts.get_url):
 225         sys.stderr.write('Warning: video file name given but will not be used.\n')
 226
 227 if cmdl_opts.outfile is not None and (cmdl_opts.use_title or cmdl_opts.use_literal):
 228         sys.exit('Error: using the video title conflicts with using a given file name.')
 229
 230 if cmdl_opts.use_netrc and cmdl_opts.password is not None:
 231         sys.exit('Error: using netrc conflicts with giving command line password.')
 232
 233 if cmdl_opts.use_title and cmdl_opts.use_literal:
 234         sys.exit('Error: cannot use title and literal title at the same time.')
 235
 236 if cmdl_opts.quiet and cmdl_opts.get_url:
 237         sys.exit('Error: cannot be quiet and print final URL at the same time.')
 238
 239 # Incorrect option formatting
 240 if cmdl_opts.username is None and cmdl_opts.password is not None:
 241         sys.exit('Error: password give but username is missing.')
 242
 243 if cmdl_opts.get_url is None and cmdl_opts.get_title is not None:
 244         sys.exit('Error: getting title requires getting URL.')
 245
 246 # Get account information if any
 247 account_username = None
 248 account_password = None
 249
 250 if cmdl_opts.use_netrc:
 251         try:
 252                 info = netrc.netrc().authenticators('youtube')
 253                 if info is None:
 254                         sys.exit('Error: no authenticators for machine youtube.')
 255                 netrc_username = info[0]
 256                 netrc_password = info[2]
 257         except IOError:
 258                 sys.exit('Error: unable to read .netrc file.')
 259         except netrc.NetrcParseError:
 260                 sys.exit('Error: unable to parse .netrc file.')
 261
 262 if cmdl_opts.password is not None:
 263         account_username = cmdl_opts.username
 264         account_password = cmdl_opts.password
 265 else:
 266         if cmdl_opts.username is not None and cmdl_opts.use_netrc:
 267                 if cmdl_opts.username != netrc_username:
 268                         sys.exit('Error: conflicting username from .netrc and command line options.')
 269                 account_username = cmdl_opts.username
 270                 account_password = netrc_password
 271         elif cmdl_opts.username is not None:
 272                 account_username = cmdl_opts.username
 273                 account_password = getpass.getpass('Type YouTube password and press return: ')
 274         elif cmdl_opts.use_netrc:
 275                 if len(netrc_username) == 0:
 276                         sys.exit('Error: empty username in .netrc file.')
 277                 account_username = netrc_username
 278                 account_password = netrc_password
 279
 280 # Get output file name
 281 if cmdl_opts.outfile is None:
 282         video_filename = '%s.flv' % video_url_id
 283 else:
 284         video_filename = cmdl_opts.outfile
 285
 286 # Check name
 287 if not video_filename.lower().endswith('.flv'):
 288         sys.stderr.write('Warning: video file name does not end in .flv\n')
 289
 290 # Test writable file
 291 if not (cmdl_opts.simulate or cmdl_opts.get_url):
 292         try:
 293                 disk_test = open(video_filename, 'wb')
 294                 disk_test.close()
 295
 296         except (OSError, IOError):
 297                 sys.exit('Error: unable to open %s for writing.' % video_filename)
 298
 299 # Install cookie and proxy handlers
 300 urllib2.install_opener(urllib2.build_opener(urllib2.ProxyHandler()))
 301 urllib2.install_opener(urllib2.build_opener(urllib2.HTTPCookieProcessor()))
 302
 303 # Log in and confirm age if needed
 304 if account_username is not None:
 305         url = const_login_url_str % video_url_id
 306         post = const_login_post_str % (video_url_id, account_username, account_password)
 307         download_step(False, 'Logging in', 'unable to log in', url, post)
 308
 309         url = const_age_url_str % video_url_id
 310         post = const_age_post_str % video_url_id
 311         download_step(False, 'Confirming age', 'unable to confirm age', url, post)
 312
 313 # Retrieve video webpage
 314 video_webpage = download_step(True, 'Retrieving video webpage', 'unable to retrieve video webpage', video_url)
 315
 316 # Extract video title if needed
 317 if cmdl_opts.use_title or cmdl_opts.use_literal or cmdl_opts.get_title:
 318         video_title = extract_step('Extracting video title', 'unable to extract video title', const_video_title_re, video_webpage)
 319
 320 # Extract needed video URL parameters
 321 video_url_t_param = extract_step('Extracting URL "t" parameter', 'unable to extract URL "t" parameter', const_url_t_param_re, video_webpage)
 322 video_url_real = const_video_url_real_str % (video_url_id, video_url_t_param)
 323
 324 # Retrieve video data
 325 try:
 326         video_data = perform_request(video_url_real)
 327         cond_print('Video data found at %s\n' % video_data.geturl())
 328
 329         if cmdl_opts.get_title:
 330                 print video_title
 331
 332         if cmdl_opts.get_url:
 333                 print video_data.geturl()
 334
 335         if cmdl_opts.simulate or cmdl_opts.get_url:
 336                 sys.exit()
 337
 338         video_file = open(video_filename, 'wb')
 339         try:
 340                 video_len = long(video_data.info()['Content-length'])
 341                 video_len_str = format_bytes(video_len)
 342         except KeyError:
 343                 video_len = None
 344                 video_len_str = 'N/A'
 345
 346         byte_counter = 0
 347         block_size = const_initial_block_size
 348         start_time = time.time()
 349         while True:
 350                 if video_len is not None:
 351                         percent = float(byte_counter) / float(video_len) * 100.0
 352                         percent_str = '%.1f' % percent
 353                         eta_str = calc_eta(start_time, time.time(), video_len, byte_counter)
 354                 else:
 355                         percent_str = '---.-'
 356                         eta_str = '--:--'
 357                 counter = format_bytes(byte_counter)
 358                 speed_str = calc_speed(start_time, time.time(), byte_counter)
 359                 cond_print('\rRetrieving video data: %5s%% (%8s of %s) at %8s/s ETA %s ' % (percent_str, counter, video_len_str, speed_str, eta_str))
 360
 361                 before = time.time()
 362                 video_block = video_data.read(block_size)
 363                 after = time.time()
 364                 dl_bytes = len(video_block)
 365                 if dl_bytes == 0:
 366                         break
 367                 byte_counter += dl_bytes
 368                 video_file.write(video_block)
 369                 block_size = new_block_size(before, after, dl_bytes)
 370
 371         if video_len is not None and byte_counter != video_len:
 372                 error_advice_exit('server did not send the expected ammount of data')
 373
 374         video_file.close()
 375         cond_print('done.\n')
 376         cond_print('Video data saved to %s\n' % video_filename)
 377
 378 except (urllib2.URLError, ValueError, httplib.HTTPException, TypeError, socket.error):
 379         cond_print('failed.\n')
 380         error_advice_exit('unable to download video data')
 381
 382 except KeyboardInterrupt:
 383         sys.exit('\n')
 384
 385 # Rename video file if needed
 386 if cmdl_opts.use_title or cmdl_opts.use_literal:
 387         try:
 388                 if cmdl_opts.use_title:
 389                         prefix = title_string_norm(video_title)
 390                 else:
 391                         prefix = title_string_touch(video_title)
 392                 final_filename = '%s-%s.flv' % (prefix, video_url_id)
 393                 os.rename(video_filename, final_filename)
 394                 cond_print('Video file renamed to %s\n' % final_filename)
 395
 396         except OSError:
 397                 sys.stderr.write('Warning: unable to rename file.\n')
 398
 399         except KeyboardInterrupt:
 400                 sys.exit('\n')
 401
 402 # Finish
 403 sys.exit()