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

   1 import os
   2 import re
   3 import subprocess
   4 import sys
   5 import time
   6
   7 from .common import FileDownloader
   8 from ..utils import (
   9     compat_urllib_request,
  10     compat_urllib_error,
  11     ContentTooShortError,
  12
  13     encodeFilename,
  14     sanitize_open,
  15     format_bytes,
  16 )
  17
  18
  19 class HttpFD(FileDownloader):
  20     def real_download(self, filename, info_dict):
  21         url = info_dict['url']
  22         tmpfilename = self.temp_name(filename)
  23         stream = None
  24
  25         # Do not include the Accept-Encoding header
  26         headers = {'Youtubedl-no-compression': 'True'}
  27         if 'user_agent' in info_dict:
  28             headers['Youtubedl-user-agent'] = info_dict['user_agent']
  29         basic_request = compat_urllib_request.Request(url, None, headers)
  30         request = compat_urllib_request.Request(url, None, headers)
  31
  32         if self.params.get('test', False):
  33             request.add_header('Range','bytes=0-10240')
  34
  35         # Establish possible resume length
  36         if os.path.isfile(encodeFilename(tmpfilename)):
  37             resume_len = os.path.getsize(encodeFilename(tmpfilename))
  38         else:
  39             resume_len = 0
  40
  41         open_mode = 'wb'
  42         if resume_len != 0:
  43             if self.params.get('continuedl', False):
  44                 self.report_resuming_byte(resume_len)
  45                 request.add_header('Range','bytes=%d-' % resume_len)
  46                 open_mode = 'ab'
  47             else:
  48                 resume_len = 0
  49
  50         count = 0
  51         retries = self.params.get('retries', 0)
  52         while count <= retries:
  53             # Establish connection
  54             try:
  55                 data = compat_urllib_request.urlopen(request)
  56                 break
  57             except (compat_urllib_error.HTTPError, ) as err:
  58                 if (err.code < 500 or err.code >= 600) and err.code != 416:
  59                     # Unexpected HTTP error
  60                     raise
  61                 elif err.code == 416:
  62                     # Unable to resume (requested range not satisfiable)
  63                     try:
  64                         # Open the connection again without the range header
  65                         data = compat_urllib_request.urlopen(basic_request)
  66                         content_length = data.info()['Content-Length']
  67                     except (compat_urllib_error.HTTPError, ) as err:
  68                         if err.code < 500 or err.code >= 600:
  69                             raise
  70                     else:
  71                         # Examine the reported length
  72                         if (content_length is not None and
  73                                 (resume_len - 100 < int(content_length) < resume_len + 100)):
  74                             # The file had already been fully downloaded.
  75                             # Explanation to the above condition: in issue #175 it was revealed that
  76                             # YouTube sometimes adds or removes a few bytes from the end of the file,
  77                             # changing the file size slightly and causing problems for some users. So
  78                             # I decided to implement a suggested change and consider the file
  79                             # completely downloaded if the file size differs less than 100 bytes from
  80                             # the one in the hard drive.
  81                             self.report_file_already_downloaded(filename)
  82                             self.try_rename(tmpfilename, filename)
  83                             self._hook_progress({
  84                                 'filename': filename,
  85                                 'status': 'finished',
  86                             })
  87                             return True
  88                         else:
  89                             # The length does not match, we start the download over
  90                             self.report_unable_to_resume()
  91                             open_mode = 'wb'
  92                             break
  93             # Retry
  94             count += 1
  95             if count <= retries:
  96                 self.report_retry(count, retries)
  97
  98         if count > retries:
  99             self.report_error(u'giving up after %s retries' % retries)
 100             return False
 101
 102         data_len = data.info().get('Content-length', None)
 103         if data_len is not None:
 104             data_len = int(data_len) + resume_len
 105             min_data_len = self.params.get("min_filesize", None)
 106             max_data_len =  self.params.get("max_filesize", None)
 107             if min_data_len is not None and data_len < min_data_len:
 108                 self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
 109                 return False
 110             if max_data_len is not None and data_len > max_data_len:
 111                 self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
 112                 return False
 113
 114         data_len_str = format_bytes(data_len)
 115         byte_counter = 0 + resume_len
 116         block_size = self.params.get('buffersize', 1024)
 117         start = time.time()
 118         while True:
 119             # Download and write
 120             before = time.time()
 121             data_block = data.read(block_size)
 122             after = time.time()
 123             if len(data_block) == 0:
 124                 break
 125             byte_counter += len(data_block)
 126
 127             # Open file just in time
 128             if stream is None:
 129                 try:
 130                     (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
 131                     assert stream is not None
 132                     filename = self.undo_temp_name(tmpfilename)
 133                     self.report_destination(filename)
 134                 except (OSError, IOError) as err:
 135                     self.report_error(u'unable to open for writing: %s' % str(err))
 136                     return False
 137             try:
 138                 stream.write(data_block)
 139             except (IOError, OSError):
 140                 self.to_stderr(u"\n")
 141                 self.report_error(u'unable to write data: %s' % str(err))
 142                 return False
 143             if not self.params.get('noresizebuffer', False):
 144                 block_size = self.best_block_size(after - before, len(data_block))
 145
 146             # Progress message
 147             speed = self.calc_speed(start, time.time(), byte_counter - resume_len)
 148             if data_len is None:
 149                 eta = percent = None
 150             else:
 151                 percent = self.calc_percent(byte_counter, data_len)
 152                 eta = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
 153             self.report_progress(percent, data_len_str, speed, eta)
 154
 155             self._hook_progress({
 156                 'downloaded_bytes': byte_counter,
 157                 'total_bytes': data_len,
 158                 'tmpfilename': tmpfilename,
 159                 'filename': filename,
 160                 'status': 'downloading',
 161                 'eta': eta,
 162                 'speed': speed,
 163             })
 164
 165             # Apply rate limit
 166             self.slow_down(start, byte_counter - resume_len)
 167
 168         if stream is None:
 169             self.to_stderr(u"\n")
 170             self.report_error(u'Did not get any data blocks')
 171             return False
 172         stream.close()
 173         self.report_finish(data_len_str, (time.time() - start))
 174         if data_len is not None and byte_counter != data_len:
 175             raise ContentTooShortError(byte_counter, int(data_len))
 176         self.try_rename(tmpfilename, filename)
 177
 178         # Update file modification time
 179         if self.params.get('updatetime', True):
 180             info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
 181
 182         self._hook_progress({
 183             'downloaded_bytes': byte_counter,
 184             'total_bytes': byte_counter,
 185             'filename': filename,
 186             'status': 'finished',
 187         })
 188
 189         return True