Use a temporary filename to download files

This commit is contained in:
Ricardo Garcia 2010-12-04 10:38:53 +01:00
parent a1f03c7b06
commit 62cf7aaf9a

View file

@ -235,6 +235,11 @@ class FileDownloader(object):
if not os.path.exists(dir): if not os.path.exists(dir):
os.mkdir(dir) os.mkdir(dir)
@staticmethod
def temp_name(filename):
"""Returns a temporary filename for the given filename."""
return filename + '.part'
@staticmethod @staticmethod
def format_bytes(bytes): def format_bytes(bytes):
if bytes is None: if bytes is None:
@ -353,6 +358,12 @@ class FileDownloader(object):
speed = float(byte_counter) / elapsed speed = float(byte_counter) / elapsed
if speed > rate_limit: if speed > rate_limit:
time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit) time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit)
def try_rename(self, old_filename, new_filename):
try:
os.rename(old_filename, new_filename)
except (IOError, OSError), err:
self.trouble(u'ERROR: unable to rename file')
def report_destination(self, filename): def report_destination(self, filename):
"""Report destination filename.""" """Report destination filename."""
@ -484,6 +495,7 @@ class FileDownloader(object):
def _download_with_rtmpdump(self, filename, url, player_url): def _download_with_rtmpdump(self, filename, url, player_url):
self.report_destination(filename) self.report_destination(filename)
tmpfilename = self.temp_name(filename)
# Check for rtmpdump first # Check for rtmpdump first
try: try:
@ -495,36 +507,43 @@ class FileDownloader(object):
# Download using rtmpdump. rtmpdump returns exit code 2 when # Download using rtmpdump. rtmpdump returns exit code 2 when
# the connection was interrumpted and resuming appears to be # the connection was interrumpted and resuming appears to be
# possible. This is part of rtmpdump's normal usage, AFAIK. # possible. This is part of rtmpdump's normal usage, AFAIK.
basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', filename] basic_args = ['rtmpdump', '-q'] + [[], ['-W', player_url]][player_url is not None] + ['-r', url, '-o', tmpfilename]
retval = subprocess.call(basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)]) retval = subprocess.call(basic_args + [[], ['-e', '-k', '1']][self.params.get('continuedl', False)])
while retval == 2 or retval == 1: while retval == 2 or retval == 1:
prevsize = os.path.getsize(filename) prevsize = os.path.getsize(tmpfilename)
self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True) self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True)
time.sleep(5.0) # This seems to be needed time.sleep(5.0) # This seems to be needed
retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1]) retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1])
cursize = os.path.getsize(filename) cursize = os.path.getsize(tmpfilename)
if prevsize == cursize and retval == 1: if prevsize == cursize and retval == 1:
break break
if retval == 0: if retval == 0:
self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(filename)) self.to_screen(u'\r[rtmpdump] %s bytes' % os.path.getsize(tmpfilename))
self.try_rename(tmpfilename, filename)
return True return True
else: else:
self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval) self.trouble(u'\nERROR: rtmpdump exited with code %d' % retval)
return False return False
def _do_download(self, filename, url, player_url): def _do_download(self, filename, url, player_url):
# Check file already present
if self.params.get('continuedl', False) and os.path.isfile(filename):
self.report_file_already_downloaded(filename)
return True
# Attempt to download using rtmpdump # Attempt to download using rtmpdump
if url.startswith('rtmp'): if url.startswith('rtmp'):
return self._download_with_rtmpdump(filename, url, player_url) return self._download_with_rtmpdump(filename, url, player_url)
tmpfilename = self.temp_name(filename)
stream = None stream = None
open_mode = 'wb' open_mode = 'wb'
basic_request = urllib2.Request(url, None, std_headers) basic_request = urllib2.Request(url, None, std_headers)
request = urllib2.Request(url, None, std_headers) request = urllib2.Request(url, None, std_headers)
# Establish possible resume length # Establish possible resume length
if os.path.isfile(filename): if os.path.isfile(tmpfilename):
resume_len = os.path.getsize(filename) resume_len = os.path.getsize(tmpfilename)
else: else:
resume_len = 0 resume_len = 0
@ -566,6 +585,7 @@ class FileDownloader(object):
# completely downloaded if the file size differs less than 100 bytes from # completely downloaded if the file size differs less than 100 bytes from
# the one in the hard drive. # the one in the hard drive.
self.report_file_already_downloaded(filename) self.report_file_already_downloaded(filename)
self.try_rename(tmpfilename, filename)
return True return True
else: else:
# The length does not match, we start the download over # The length does not match, we start the download over
@ -599,7 +619,7 @@ class FileDownloader(object):
# Open file just in time # Open file just in time
if stream is None: if stream is None:
try: try:
(stream, filename) = sanitize_open(filename, open_mode) (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
self.report_destination(filename) self.report_destination(filename)
except (OSError, IOError), err: except (OSError, IOError), err:
self.trouble(u'ERROR: unable to open for writing: %s' % str(err)) self.trouble(u'ERROR: unable to open for writing: %s' % str(err))
@ -623,6 +643,7 @@ class FileDownloader(object):
self.report_finish() self.report_finish()
if data_len is not None and str(byte_counter) != data_len: if data_len is not None and str(byte_counter) != data_len:
raise ContentTooShortError(byte_counter, long(data_len)) raise ContentTooShortError(byte_counter, long(data_len))
self.try_rename(tmpfilename, filename)
return True return True
class InfoExtractor(object): class InfoExtractor(object):