_ Git - youtube-dl/blob - youtube_dl/downloader/http.py

   1 import os
   2 import time
   3
   4 from .common import FileDownloader
   5 from ..utils import (
   6     compat_urllib_request,
   7     compat_urllib_error,
   8     ContentTooShortError,
   9
  10     encodeFilename,
  11     sanitize_open,
  12     format_bytes,
  13 )
  14
  15
  16 class HttpFD(FileDownloader):
  17     def real_download(self, filename, info_dict):
  18         url = info_dict['url']
  19         tmpfilename = self.temp_name(filename)
  20         stream = None
  21
  22         # Do not include the Accept-Encoding header
  23         headers = {'Youtubedl-no-compression': 'True'}
  24         if 'user_agent' in info_dict:
  25             headers['Youtubedl-user-agent'] = info_dict['user_agent']
  26         basic_request = compat_urllib_request.Request(url, None, headers)
  27         request = compat_urllib_request.Request(url, None, headers)
  28
  29         if self.params.get('test', False):
  30             request.add_header('Range','bytes=0-10240')
  31
  32         # Establish possible resume length
  33         if os.path.isfile(encodeFilename(tmpfilename)):
  34             resume_len = os.path.getsize(encodeFilename(tmpfilename))
  35         else:
  36             resume_len = 0
  37
  38         open_mode = 'wb'
  39         if resume_len != 0:
  40             if self.params.get('continuedl', False):
  41                 self.report_resuming_byte(resume_len)
  42                 request.add_header('Range','bytes=%d-' % resume_len)
  43                 open_mode = 'ab'
  44             else:
  45                 resume_len = 0
  46
  47         count = 0
  48         retries = self.params.get('retries', 0)
  49         while count <= retries:
  50             # Establish connection
  51             try:
  52                 data = compat_urllib_request.urlopen(request)
  53                 break
  54             except (compat_urllib_error.HTTPError, ) as err:
  55                 if (err.code < 500 or err.code >= 600) and err.code != 416:
  56                     # Unexpected HTTP error
  57                     raise
  58                 elif err.code == 416:
  59                     # Unable to resume (requested range not satisfiable)
  60                     try:
  61                         # Open the connection again without the range header
  62                         data = compat_urllib_request.urlopen(basic_request)
  63                         content_length = data.info()['Content-Length']
  64                     except (compat_urllib_error.HTTPError, ) as err:
  65                         if err.code < 500 or err.code >= 600:
  66                             raise
  67                     else:
  68                         # Examine the reported length
  69                         if (content_length is not None and
  70                                 (resume_len - 100 < int(content_length) < resume_len + 100)):
  71                             # The file had already been fully downloaded.
  72                             # Explanation to the above condition: in issue #175 it was revealed that
  73                             # YouTube sometimes adds or removes a few bytes from the end of the file,
  74                             # changing the file size slightly and causing problems for some users. So
  75                             # I decided to implement a suggested change and consider the file
  76                             # completely downloaded if the file size differs less than 100 bytes from
  77                             # the one in the hard drive.
  78                             self.report_file_already_downloaded(filename)
  79                             self.try_rename(tmpfilename, filename)
  80                             self._hook_progress({
  81                                 'filename': filename,
  82                                 'status': 'finished',
  83                             })
  84                             return True
  85                         else:
  86                             # The length does not match, we start the download over
  87                             self.report_unable_to_resume()
  88                             open_mode = 'wb'
  89                             break
  90             # Retry
  91             count += 1
  92             if count <= retries:
  93                 self.report_retry(count, retries)
  94
  95         if count > retries:
  96             self.report_error(u'giving up after %s retries' % retries)
  97             return False
  98
  99         data_len = data.info().get('Content-length', None)
 100         if data_len is not None:
 101             data_len = int(data_len) + resume_len
 102             min_data_len = self.params.get("min_filesize", None)
 103             max_data_len =  self.params.get("max_filesize", None)
 104             if min_data_len is not None and data_len < min_data_len:
 105                 self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
 106                 return False
 107             if max_data_len is not None and data_len > max_data_len:
 108                 self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
 109                 return False
 110
 111         data_len_str = format_bytes(data_len)
 112         byte_counter = 0 + resume_len
 113         block_size = self.params.get('buffersize', 1024)
 114         start = time.time()
 115         while True:
 116             # Download and write
 117             before = time.time()
 118             data_block = data.read(block_size)
 119             after = time.time()
 120             if len(data_block) == 0:
 121                 break
 122             byte_counter += len(data_block)
 123
 124             # Open file just in time
 125             if stream is None:
 126                 try:
 127                     (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
 128                     assert stream is not None
 129                     filename = self.undo_temp_name(tmpfilename)
 130                     self.report_destination(filename)
 131                 except (OSError, IOError) as err:
 132                     self.report_error(u'unable to open for writing: %s' % str(err))
 133                     return False
 134             try:
 135                 stream.write(data_block)
 136             except (IOError, OSError) as err:
 137                 self.to_stderr(u"\n")
 138                 self.report_error(u'unable to write data: %s' % str(err))
 139                 return False
 140             if not self.params.get('noresizebuffer', False):
 141                 block_size = self.best_block_size(after - before, len(data_block))
 142
 143             # Progress message
 144             speed = self.calc_speed(start, time.time(), byte_counter - resume_len)
 145             if data_len is None:
 146                 eta = percent = None
 147             else:
 148                 percent = self.calc_percent(byte_counter, data_len)
 149                 eta = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
 150             self.report_progress(percent, data_len_str, speed, eta)
 151
 152             self._hook_progress({
 153                 'downloaded_bytes': byte_counter,
 154                 'total_bytes': data_len,
 155                 'tmpfilename': tmpfilename,
 156                 'filename': filename,
 157                 'status': 'downloading',
 158                 'eta': eta,
 159                 'speed': speed,
 160             })
 161
 162             # Apply rate limit
 163             self.slow_down(start, byte_counter - resume_len)
 164
 165         if stream is None:
 166             self.to_stderr(u"\n")
 167             self.report_error(u'Did not get any data blocks')
 168             return False
 169         stream.close()
 170         self.report_finish(data_len_str, (time.time() - start))
 171         if data_len is not None and byte_counter != data_len:
 172             raise ContentTooShortError(byte_counter, int(data_len))
 173         self.try_rename(tmpfilename, filename)
 174
 175         # Update file modification time
 176         if self.params.get('updatetime', True):
 177             info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
 178
 179         self._hook_progress({
 180             'downloaded_bytes': byte_counter,
 181             'total_bytes': byte_counter,
 182             'filename': filename,
 183             'status': 'finished',
 184         })
 185
 186         return True