Add a number of retries with tweaked patch, originally from Neil Channen

This commit is contained in:
Ricardo Garcia 2010-05-30 18:34:56 +02:00
parent e616ec0ca6
commit 7031008c98

View File

@ -193,6 +193,7 @@ class FileDownloader(object):
ignoreerrors: Do not stop on download errors.
ratelimit: Download speed limit, in bytes/sec.
nooverwrites: Prevent overwriting files.
retries: Number of times to retry for HTTP error 503
continuedl: Try to continue downloads if possible.
noprogress: Do not print the progress bar.
"""
@ -364,6 +365,10 @@ class FileDownloader(object):
"""Report attemtp to resume at given byte."""
self.to_stdout(u'[download] Resuming download at byte %s' % resume_len)
def report_retry(self, count, retries):
"""Report retry in case of HTTP error 503"""
self.to_stdout(u'[download] Got HTTP error 503. Retrying (attempt %d of %d)...' % (count, retries))
def report_file_already_downloaded(self, file_name):
"""Report file has already been fully downloaded."""
try:
@ -527,25 +532,34 @@ class FileDownloader(object):
request.add_header('Range','bytes=%d-' % resume_len)
open_mode = 'ab'
# Establish connection
try:
data = urllib2.urlopen(request)
except (urllib2.HTTPError, ), err:
if err.code != 416: # 416 is 'Requested range not satisfiable'
raise
# Unable to resume
data = urllib2.urlopen(basic_request)
content_length = data.info()['Content-Length']
count = 0
retries = self.params.get('retries', 0)
while True:
# Establish connection
try:
data = urllib2.urlopen(request)
break
except (urllib2.HTTPError, ), err:
if err.code == 503:
# Retry in case of HTTP error 503
count += 1
if count <= retries:
self.report_retry(count, retries)
continue
if err.code != 416: # 416 is 'Requested range not satisfiable'
raise
# Unable to resume
data = urllib2.urlopen(basic_request)
content_length = data.info()['Content-Length']
if content_length is not None and long(content_length) == resume_len:
# Because the file had already been fully downloaded
self.report_file_already_downloaded(filename)
self._num_downloads += 1
return True
else:
# Because the server didn't let us
self.report_unable_to_resume()
open_mode = 'wb'
if content_length is not None and long(content_length) == resume_len:
# Because the file had already been fully downloaded
self.report_file_already_downloaded(filename)
return True
else:
# Because the server didn't let us
self.report_unable_to_resume()
open_mode = 'wb'
data_len = data.info().get('Content-length', None)
data_len_str = self.format_bytes(data_len)
@ -1985,6 +1999,8 @@ if __name__ == '__main__':
action='store_true', dest='ignoreerrors', help='continue on download errors', default=False)
parser.add_option('-r', '--rate-limit',
dest='ratelimit', metavar='L', help='download rate limit (e.g. 50k or 44.6m)')
parser.add_option('-R', '--retries',
dest='retries', metavar='T', help='number of retries (default is 10)', default=10)
authentication = optparse.OptionGroup(parser, 'Authentication Options')
authentication.add_option('-u', '--username',
@ -2073,6 +2089,11 @@ if __name__ == '__main__':
if numeric_limit is None:
parser.error(u'invalid rate limit specified')
opts.ratelimit = numeric_limit
if opts.retries is not None:
try:
opts.retries = long(opts.retries)
except (TypeError, ValueError), err:
parser.error(u'invalid retry count specified')
# Information extractors
youtube_ie = YoutubeIE()
@ -2109,6 +2130,7 @@ if __name__ == '__main__':
'ignoreerrors': opts.ignoreerrors,
'ratelimit': opts.ratelimit,
'nooverwrites': opts.nooverwrites,
'retries': opts.retries,
'continuedl': opts.continue_dl,
'noprogress': opts.noprogress,
})