-const_video_url_str = 'http://www.youtube.com/watch?v=%s'
-const_video_url_re = re.compile(r'^((?:http://)?(?:\w+\.)?youtube\.com/(?:v/|(?:watch(?:\.php)?)?\?(?:.+&)?v=))?([0-9A-Za-z_-]+)(?(1)[&/].*)?$')
-const_video_url_best_quality_suffix='&fmt=18'
-const_login_url_str = 'http://www.youtube.com/login?next=/watch%%3Fv%%3D%s'
-const_login_post_str = 'current_form=loginForm&next=%%2Fwatch%%3Fv%%3D%s&username=%s&password=%s&action_login=Log+In'
-const_age_url_str = 'http://www.youtube.com/verify_age?next_url=/watch%%3Fv%%3D%s'
-const_age_post_str = 'next_url=%%2Fwatch%%3Fv%%3D%s&action_confirm=Confirm'
-const_url_t_param_re = re.compile(r', "t": "([^"]+)"')
-const_video_url_real_str = 'http://www.youtube.com/get_video?video_id=%s&t=%s'
-const_video_title_re = re.compile(r'<title>YouTube - ([^<]*)</title>', re.M | re.I)
-
-# Print error message, followed by standard advice information, and then exit
-def error_advice_exit(error_text):
- sys.stderr.write('Error: %s.\n' % error_text)
- sys.stderr.write('Try again several times. It may be a temporary problem.\n')
- sys.stderr.write('Other typical problems:\n\n')
- sys.stderr.write('* Video no longer exists.\n')
- sys.stderr.write('* Video requires age confirmation but you did not provide an account.\n')
- sys.stderr.write('* You provided the account data, but it is not valid.\n')
- sys.stderr.write('* The connection was cut suddenly for some reason.\n')
- sys.stderr.write('* YouTube changed their system, and the program no longer works.\n')
- sys.stderr.write('\nTry to confirm you are able to view the video using a web browser.\n')
- sys.stderr.write('Use the same video URL and account information, if needed, with this program.\n')
- sys.stderr.write('When using a proxy, make sure http_proxy has http://host:port format.\n')
- sys.stderr.write('Try again several times and contact me if the problem persists.\n')
- sys.exit('\n')
-
-# Wrapper to create custom requests with typical headers
-def request_create(url, extra_headers, post_data):
- retval = urllib2.Request(url)
- if post_data is not None:
- retval.add_data(post_data)
- # Try to mimic Firefox, at least a little bit
- retval.add_header('User-Agent', 'Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.11) Gecko/20071127 Firefox/2.0.0.11')
- retval.add_header('Accept-Charset', 'ISO-8859-1,utf-8;q=0.7,*;q=0.7')
- 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')
- retval.add_header('Accept-Language', 'en-us,en;q=0.5')
- if extra_headers is not None:
- for header in extra_headers:
- retval.add_header(header[0], header[1])
- return retval
-
-# Perform a request, process headers and return response
-def perform_request(url, headers=None, data=None):
- request = request_create(url, headers, data)
- response = urllib2.urlopen(request)
- return response
-
-# Conditional print
-def cond_print(str):
- global cmdl_opts
- if not (cmdl_opts.quiet or cmdl_opts.get_url):
- sys.stdout.write(str)
- sys.stdout.flush()
-
-# Title string normalization
-def title_string_norm(title):
- title = ''.join((x in string.ascii_letters or x in string.digits) and x or ' ' for x in title)
- title = '_'.join(title.split())
- title = title.lower()
- return title
-
-# Generic download step
-def download_step(return_data_flag, step_title, step_error, url, post_data=None):
- try:
- cond_print('%s... ' % step_title)
- data = perform_request(url, data=post_data).read()
- cond_print('done.\n')
- if return_data_flag:
- return data
- return None
-
- except (urllib2.URLError, ValueError, httplib.HTTPException, TypeError, socket.error):
- cond_print('failed.\n')
- error_advice_exit(step_error)
-
- except KeyboardInterrupt:
- sys.exit('\n')
-
-# Generic extract step
-def extract_step(step_title, step_error, regexp, data):
- try:
- cond_print('%s... ' % step_title)
- match = regexp.search(data)
-
- if match is None:
- cond_print('failed.\n')
- error_advice_exit(step_error)
-
- extracted_data = match.group(1)
- cond_print('done.\n')
- return extracted_data
-
- except KeyboardInterrupt:
- sys.exit('\n')
-
-# Calculate new block size based on previous block size
-def new_block_size(before, after, bytes):
- new_min = max(bytes / 2.0, 1.0)
- new_max = max(bytes * 2.0, 1.0)
- dif = after - before
- if dif < const_epsilon:
- return int(new_max)
- rate = bytes / dif
- if rate > new_max:
- return int(new_max)
- if rate < new_min:
- return int(new_min)
- return int(rate)
-
-# Get optimum 1k exponent to represent a number of bytes
-def optimum_k_exp(num_bytes):
- global const_1k
- if num_bytes == 0:
- return 0
- return long(math.log(num_bytes, const_1k))
-
-# Get optimum representation of number of bytes
-def format_bytes(num_bytes):
- global const_1k
- try:
- exp = optimum_k_exp(num_bytes)
- suffix = 'bkMGTPEZY'[exp]
- if exp == 0:
- return '%s%s' % (num_bytes, suffix)
- converted = float(num_bytes) / float(const_1k**exp)
- return '%.2f%s' % (converted, suffix)
- except IndexError:
- sys.exit('Error: internal error formatting number of bytes.')
-
-# Calculate ETA and return it in string format as MM:SS
-def calc_eta(start, now, total, current):
- dif = now - start
- if current == 0 or dif < const_epsilon:
- return '--:--'
- rate = float(current) / dif
- eta = long((total - current) / rate)
- (eta_mins, eta_secs) = divmod(eta, 60)
- if eta_mins > 99:
- return '--:--'
- return '%02d:%02d' % (eta_mins, eta_secs)
-
-# Calculate speed and return it in string format
-def calc_speed(start, now, bytes):
- dif = now - start
- if bytes == 0 or dif < const_epsilon:
- return 'N/A b'
- return format_bytes(float(bytes) / dif)
-
-
-# Title string minimal transformation
-def title_string_touch(title):
- return title.replace(os.sep, '%')
-
-# Create the command line options parser and parse command line
-cmdl_usage = 'usage: %prog [options] video_url'
-cmdl_version = '2008.03.08'
-cmdl_parser = optparse.OptionParser(usage=cmdl_usage, version=cmdl_version, conflict_handler='resolve')
-cmdl_parser.add_option('-h', '--help', action='help', help='print this help text and exit')
-cmdl_parser.add_option('-v', '--version', action='version', help='print program version and exit')
-cmdl_parser.add_option('-u', '--username', dest='username', metavar='USERNAME', help='account username')
-cmdl_parser.add_option('-p', '--password', dest='password', metavar='PASSWORD', help='account password')
-cmdl_parser.add_option('-o', '--output', dest='outfile', metavar='FILE', help='output video file name')
-cmdl_parser.add_option('-q', '--quiet', action='store_true', dest='quiet', help='activates quiet mode')
-cmdl_parser.add_option('-s', '--simulate', action='store_true', dest='simulate', help='do not download video')
-cmdl_parser.add_option('-t', '--title', action='store_true', dest='use_title', help='use title in file name')
-cmdl_parser.add_option('-l', '--literal', action='store_true', dest='use_literal', help='use literal title in file name')
-cmdl_parser.add_option('-n', '--netrc', action='store_true', dest='use_netrc', help='use .netrc authentication data')
-cmdl_parser.add_option('-g', '--get-url', action='store_true', dest='get_url', help='print final video URL only')
-cmdl_parser.add_option('-2', '--title-too', action='store_true', dest='get_title', help='used with -g, print title too')
-cmdl_parser.add_option('-b', '--best-quality', action='store_true', dest='best_quality', help='try to download the best quality version')
-(cmdl_opts, cmdl_args) = cmdl_parser.parse_args()
-
-# Set socket timeout
-socket.setdefaulttimeout(const_timeout)
-
-# Get video URL
-if len(cmdl_args) != 1:
- cmdl_parser.print_help()
- sys.exit('\n')
-video_url_cmdl = cmdl_args[0]
-
-# Verify video URL format and convert to "standard" format
-video_url_mo = const_video_url_re.match(video_url_cmdl)
-if video_url_mo is None:
- sys.exit('Error: URL does not seem to be a youtube video URL. If it is, report a bug.')
-video_url_id = video_url_mo.group(2)
-video_url = const_video_url_str % video_url_id
-if cmdl_opts.best_quality:
- video_url = '%s%s' % (video_url, const_video_url_best_quality_suffix)
- video_extension = '.mp4'
-else:
- video_extension = '.flv'
-
-# Check conflicting options
-if cmdl_opts.outfile is not None and (cmdl_opts.simulate or cmdl_opts.get_url):
- sys.stderr.write('Warning: video file name given but will not be used.\n')
-
-if cmdl_opts.outfile is not None and (cmdl_opts.use_title or cmdl_opts.use_literal):
- sys.exit('Error: using the video title conflicts with using a given file name.')
-
-if cmdl_opts.use_title and cmdl_opts.use_literal:
- sys.exit('Error: cannot use title and literal title at the same time.')
-
-if cmdl_opts.quiet and cmdl_opts.get_url:
- sys.exit('Error: cannot be quiet and print final URL at the same time.')
-
-# Incorrect option formatting
-if cmdl_opts.username is None and cmdl_opts.password is not None:
- sys.exit('Error: password give but username is missing.')
-
-if cmdl_opts.use_netrc and (cmdl_opts.username is not None or cmdl_opts.password is not None):
- sys.exit('Error: cannot use netrc and username/password at the same time.')
-
-if cmdl_opts.get_url is None and cmdl_opts.get_title is not None:
- sys.exit('Error: getting title requires getting URL.')
-
-# Get account information if any
-account_username = None
-account_password = None
-
-if cmdl_opts.use_netrc:
- try:
- info = netrc.netrc().authenticators('youtube')
- if info is None:
- sys.exit('Error: no authenticators for machine youtube.')
- account_username = info[0]
- account_password = info[2]
- except IOError:
- sys.exit('Error: unable to read .netrc file.')
- except netrc.NetrcParseError:
- sys.exit('Error: unable to parse .netrc file.')
-else:
- account_username = cmdl_opts.username
- if account_username is not None:
- if cmdl_opts.password is None:
- account_password = getpass.getpass('Type YouTube password and press return: ')
- else:
- account_password = cmdl_opts.password
-
-# Get output file name
-if cmdl_opts.outfile is None:
- video_filename = '%s%s' % (video_url_id, video_extension)
-else:
- video_filename = cmdl_opts.outfile
-
-# Install cookie and proxy handlers
-urllib2.install_opener(urllib2.build_opener(urllib2.ProxyHandler()))
-urllib2.install_opener(urllib2.build_opener(urllib2.HTTPCookieProcessor()))
-
-# Log in and confirm age if needed
-if account_username is not None:
- url = const_login_url_str % video_url_id
- post = const_login_post_str % (video_url_id, account_username, account_password)
- download_step(False, 'Logging in', 'unable to log in', url, post)
-
- url = const_age_url_str % video_url_id
- post = const_age_post_str % video_url_id
- download_step(False, 'Confirming age', 'unable to confirm age', url, post)
-
-# Retrieve video webpage
-video_webpage = download_step(True, 'Retrieving video webpage', 'unable to retrieve video webpage', video_url)
-
-# Extract video title if needed
-if cmdl_opts.use_title or cmdl_opts.use_literal or cmdl_opts.get_title:
- video_title = extract_step('Extracting video title', 'unable to extract video title', const_video_title_re, video_webpage)
-
-# Extract needed video URL parameters
-video_url_t_param = extract_step('Extracting URL "t" parameter', 'unable to extract URL "t" parameter', const_url_t_param_re, video_webpage)
-video_url_real = const_video_url_real_str % (video_url_id, video_url_t_param)
-if cmdl_opts.best_quality:
- video_url_real = '%s%s' % (video_url_real, const_video_url_best_quality_suffix)
-
-# Rebuild filename if needed
-if cmdl_opts.use_title or cmdl_opts.use_literal:
- if cmdl_opts.use_title:
- prefix = title_string_norm(video_title)
- else:
- prefix = title_string_touch(video_title)
- video_filename = '%s-%s%s' % (prefix, video_url_id, video_extension)
-
-# Check name
-if not video_filename.lower().endswith(video_extension):
- sys.stderr.write('Warning: video file name does not end in %s\n' % video_extension)
-
-# Retrieve video data
Rapsys Git / youtubedl / blobdiff