def get(self,
target_file,
start_info=None,
use_cache=False,
save_properties=False,
auth=None):
"""Download the URL and save it to `target_file`
Appropriate hook functions are automatically invoked
- start_info: message to print (LOG.info) when transfer begins, cache is
validated or or an error occurs; it can also be a callable.
- use_cache: if True, download will not actually happen if `target_file`
is already available *and* its ETag is not changed in the server. For
this to work, the previous call to `get` must have passed
`save_properties=True` (in order to save the ETag).
- save_properties: if True, a file named .$target_file.urlprops
containing request headers and other metadata will be saved along side
the target_file. (This is required for `use_cache` in future
downloads)
- auth: (username, password) -- optional http basic auth data
Return True only if the download actually happened.
"""
def invoke_start_info(status):
if not start_info:
return
if six.callable(start_info):
i = start_info(status)
else:
i = start_info.format(status=status)
if i:
LOG.info(i)
if not P.exists(target_file):
use_cache = False
urlprops = URLProperties(target_file)
props = urlprops.load()
if props:
# write back the new value of `last_attempt_utc` *now* so we don't
# have to deal with it when an exception arises later.
# last_attempt_utc is simpy the time of the last download attempt
props.custom['last_attempt_utc'] = BareDateTime.to_string(
datetime.utcnow())
urlprops.save(props.headers, props.custom)
req = None
if use_cache and props:
# Enable the cache header `If-None-Match`
etag = props.headers.get('ETag', props.headers.get('etag', None))
if etag:
req = six.moves.Request(self.url,
headers={'If-None-Match': etag})
else:
LOG.warn('no ETag in last headers: %s', props.headers)
req = six.moves.Request(self.url)
else:
req = six.moves.Request(self.url)
if auth:
username, password = auth
req.add_header('Authorization',
_create_http_basicauth_header(username, password))
# Set User-Agent
# XXX: (in 2.6) urllib2.py does not expose its default user-agent string
# so we copy-paste that code here (from urllib2.OpenerDirector)
urllib2_user_agent = "Python-urllib/%s" % six.moves.urllib_version
pypm_user_agent = get_user_agent(urllib2_user_agent)
assert pypm_user_agent
req.add_header('User-Agent', pypm_user_agent)
try:
u = six.moves.urlopen(req)
except six.moves.HTTPError as e:
if e.code == 304 and use_cache:
invoke_start_info('Hit')
return False # file not changed in server
else:
invoke_start_info('Get')
raise
else:
invoke_start_info('Get')
bs = 1024 * 8 # block size; from urllib.py:urlretrieve
headers = u.info()
# detect total size of the file to be downloaded
if 'Content-Length' in headers:
total_size = int(headers['Content-Length'])
assert total_size >= 0
else:
total_size = None
total_bytes_transferred = 0
chunk_size = 0
# Hook 1: initialize
self.hook_initialize(total_size)
with open(target_file + '.part', 'wb') as f:
while True:
if total_size:
assert total_bytes_transferred <= total_size, str(
(total_bytes_transferred, total_size))
# Hook 2: transferring
self.hook_transferring(chunk_size, total_bytes_transferred)
data = u.read(bs)
chunk_size = len(data)
if len(data) == 0:
break
f.write(data)
total_bytes_transferred += chunk_size
sh.mv(target_file + '.part', target_file)
# Hook 3: completed
self.hook_completed()
# save the new headers and other properties
if save_properties:
custom_dict = props and props.custom or {
} # reuse existing props.custom
custom_dict['last_attempt_utc'] = BareDateTime.to_string(
datetime.utcnow())
urlprops.save(dict(headers), custom_dict)
return True
def get(self,
target_file,
start_info=None,
use_cache=False,
save_properties=False,
auth=None):
"""Download the URL and save it to `target_file`
Appropriate hook functions are automatically invoked
- start_info: message to print (LOG.info) when transfer begins, cache is
validated or or an error occurs; it can also be a callable.
- use_cache: if True, download will not actually happen if `target_file`
is already available *and* its ETag is not changed in the server. For
this to work, the previous call to `get` must have passed
`save_properties=True` (in order to save the ETag).
- save_properties: if True, a file named .$target_file.urlprops
containing request headers and other metadata will be saved along side
the target_file. (This is required for `use_cache` in future
downloads)
- auth: (username, password) -- optional http basic auth data
Return True only if the download actually happened.
"""
def invoke_start_info(status):
if not start_info:
return
if six.callable(start_info):
i = start_info(status)
else:
i = start_info.format(status=status)
if i:
LOG.info(i)
if not P.exists(target_file):
use_cache = False
urlprops = URLProperties(target_file)
props = urlprops.load()
if props:
# write back the new value of `last_attempt_utc` *now* so we don't
# have to deal with it when an exception arises later.
# last_attempt_utc is simpy the time of the last download attempt
props.custom['last_attempt_utc'] = BareDateTime.to_string(datetime.utcnow())
urlprops.save(props.headers, props.custom)
req = None
if use_cache and props:
# Enable the cache header `If-None-Match`
etag = props.headers.get('ETag', props.headers.get('etag', None))
if etag:
req = six.moves.Request(self.url, headers={'If-None-Match': etag})
else:
LOG.warn('no ETag in last headers: %s', props.headers)
req = six.moves.Request(self.url)
else:
req = six.moves.Request(self.url)
if auth:
username, password = auth
req.add_header('Authorization', _create_http_basicauth_header(
username, password
))
# Set User-Agent
# XXX: (in 2.6) urllib2.py does not expose its default user-agent string
# so we copy-paste that code here (from urllib2.OpenerDirector)
urllib2_user_agent = "Python-urllib/%s" % six.moves.urllib_version
pypm_user_agent = get_user_agent(urllib2_user_agent)
assert pypm_user_agent
req.add_header('User-Agent', pypm_user_agent)
try:
u = six.moves.urlopen(req)
except six.moves.HTTPError as e:
if e.code == 304 and use_cache:
invoke_start_info('Hit')
return False # file not changed in server
else:
invoke_start_info('Get')
raise
else:
invoke_start_info('Get')
bs = 1024*8 # block size; from urllib.py:urlretrieve
headers = u.info()
# detect total size of the file to be downloaded
if 'Content-Length' in headers:
total_size = int(headers['Content-Length'])
assert total_size >= 0
else:
total_size = None
total_bytes_transferred = 0
chunk_size = 0
# Hook 1: initialize
self.hook_initialize(total_size)
with open(target_file + '.part', 'wb') as f:
while True:
if total_size:
assert total_bytes_transferred <= total_size, str((
total_bytes_transferred, total_size))
# Hook 2: transferring
self.hook_transferring(chunk_size, total_bytes_transferred)
data = u.read(bs)
chunk_size = len(data)
if len(data) == 0:
break
f.write(data)
total_bytes_transferred += chunk_size
sh.mv(target_file + '.part', target_file)
# Hook 3: completed
self.hook_completed()
# save the new headers and other properties
if save_properties:
custom_dict = props and props.custom or {} # reuse existing props.custom
custom_dict['last_attempt_utc'] = BareDateTime.to_string(datetime.utcnow())
urlprops.save(dict(headers), custom_dict)
return True
def _rollover_log(logfile, maxsize=(2<<20)):
"""Move $logfile to $logfile.old if its size exceeds `maxsize`"""
if exists(logfile):
filesize = os.stat(logfile)[stat.ST_SIZE]
if filesize >= maxsize:
sh.mv(logfile, logfile+'.old')
def _rollover_log(logfile, maxsize=(2 << 20)):
"""Move $logfile to $logfile.old if its size exceeds `maxsize`"""
if exists(logfile):
filesize = os.stat(logfile)[stat.ST_SIZE]
if filesize >= maxsize:
sh.mv(logfile, logfile + '.old')