def load_proxylist(self, source, source_type=None, proxy_type='http',
auto_init=True, auto_change=True,
**kwargs):
self.proxylist = ProxyList()
if isinstance(source, BaseProxySource):
self.proxylist.set_source(source)
elif isinstance(source, six.string_types):
if source_type == 'text_file':
self.proxylist.load_file(source, proxy_type=proxy_type)
elif source_type == 'url':
self.proxylist.load_url(source, proxy_type=proxy_type)
else:
raise SpiderMisuseError('Method `load_proxylist` received '
'invalid `source_type` argument: %s'
% source_type)
else:
raise SpiderMisuseError('Method `load_proxylist` received '
'invalid `source` argument: %s'
% source)
self.proxylist_enabled = True
self.proxy = None
if not auto_change and auto_init:
self.proxy = self.proxylist.get_random_proxy()
self.proxy_auto_change = auto_change
def clone(self, **kwargs):
"""
Clone Task instance.
Reset network_try_count, increase task_try_count.
"""
# First, create exact copy of the current Task object
attr_copy = self.__dict__.copy()
if attr_copy.get('grab_config') is not None:
del attr_copy['url']
task = Task(**attr_copy)
# Reset some task properties if they have not
# been set explicitly in kwargs
if not 'network_try_count' in kwargs:
task.network_try_count = 0
if not 'task_try_count' in kwargs:
task.task_try_count = self.task_try_count + 1
if not 'refresh_cache' in kwargs:
task.refresh_cache = False
if not 'disable_cache' in kwargs:
task.disable_cache = False
if kwargs.get('url') is not None and kwargs.get('grab') is not None:
raise SpiderMisuseError('Options url and grab could not be '
'used together')
if kwargs.get('url') is not None and kwargs.get(
'grab_config') is not None:
raise SpiderMisuseError('Options url and grab_config could not '
'be used together')
if kwargs.get('grab') is not None and kwargs.get(
'grab_config') is not None:
raise SpiderMisuseError('Options grab and grab_config could not '
'be used together')
if kwargs.get('grab'):
task.setup_grab_config(kwargs['grab'].dump_config())
del kwargs['grab']
elif kwargs.get('grab_config'):
task.setup_grab_config(kwargs['grab_config'])
del kwargs['grab_config']
elif kwargs.get('url'):
task.url = kwargs['url']
if task.grab_config:
task.grab_config['url'] = kwargs['url']
del kwargs['url']
for key, value in kwargs.items():
setattr(task, key, value)
# WTF?
# The `Task` object can't has `delay` attribute
# I think in next line the `process_delay_option` method
# always gets None as input argument
task.process_delay_option(task.get('delay', None))
return task
def clone(self, **kwargs):
"""
Clone Task instance.
Reset network_try_count, increase task_try_count.
Reset priority attribute if it was not set explicitly.
"""
# First, create exact copy of the current Task object
attr_copy = self.__dict__.copy()
if attr_copy.get('grab_config') is not None:
del attr_copy['url']
if not attr_copy['priority_set_explicitly']:
attr_copy['priority'] = None
task = Task(**attr_copy)
# Reset some task properties if they have not
# been set explicitly in kwargs
if 'network_try_count' not in kwargs:
task.network_try_count = 0
if 'task_try_count' not in kwargs:
task.task_try_count = self.task_try_count + 1
if 'refresh_cache' not in kwargs:
task.refresh_cache = False
if 'disable_cache' not in kwargs:
task.disable_cache = False
if kwargs.get('url') is not None and kwargs.get('grab') is not None:
raise SpiderMisuseError('Options url and grab could not be '
'used together')
if (kwargs.get('url') is not None
and kwargs.get('grab_config') is not None):
raise SpiderMisuseError('Options url and grab_config could not '
'be used together')
if (kwargs.get('grab') is not None
and kwargs.get('grab_config') is not None):
raise SpiderMisuseError('Options grab and grab_config could not '
'be used together')
if kwargs.get('grab'):
task.setup_grab_config(kwargs['grab'].dump_config())
del kwargs['grab']
elif kwargs.get('grab_config'):
task.setup_grab_config(kwargs['grab_config'])
del kwargs['grab_config']
elif kwargs.get('url'):
task.url = kwargs['url']
if task.grab_config:
task.grab_config['url'] = kwargs['url']
del kwargs['url']
for key, value in kwargs.items():
setattr(task, key, value)
task.process_delay_option(None)
return task
def add_task(self, task, queue=None, raise_error=False):
"""
Add task to the task queue.
"""
if queue is None:
queue = self.task_queue
if queue is None:
raise SpiderMisuseError('You should configure task queue before '
'adding tasks. Use `setup_queue` method.')
if task.priority is None or not task.priority_set_explicitly:
task.priority = self.generate_task_priority()
task.priority_set_explicitly = False
else:
task.priority_set_explicitly = True
if not task.url.startswith(('http://', 'https://', 'ftp://',
'file://', 'feed://')):
self.stat.collect('task-with-invalid-url', task.url)
msg = 'Invalid task URL: %s' % task.url
if raise_error:
raise SpiderError(msg)
else:
logger.error(
'%s\nTraceback:\n%s', msg, ''.join(format_stack()),
)
return False
else:
# TODO: keep original task priority if it was set explicitly
# WTF the previous comment means?
queue.put(
task, priority=task.priority, schedule_time=task.schedule_time
)
return True
def setup_cache(self,
backend='mongo',
database=None,
use_compression=True,
**kwargs):
"""
Setup cache.
:param backend: Backend name
Should be one of the following: 'mongo', 'mysql' or 'postgresql'.
:param database: Database name.
:param kwargs: Additional credentials for backend.
"""
if database is None:
raise SpiderMisuseError('setup_cache method requires database '
'option')
self.cache_enabled = True
mod = __import__('grab.spider.cache_backend.%s' % backend, globals(),
locals(), ['foo'])
cache = mod.CacheBackend(database=database,
use_compression=use_compression,
spider=self,
**kwargs)
self.cache_pipeline = CachePipeline(self, cache)
def setup_cache(self, backend='mongodb', database=None,
**kwargs):
"""
Setup cache.
:param backend: Backend name
Should be one of the following: 'mongo', 'mysql' or 'postgresql'.
:param database: Database name.
:param kwargs: Additional credentials for backend.
"""
if database is None:
raise SpiderMisuseError('setup_cache method requires database '
'option')
if backend == 'mongo':
warn('Backend name "mongo" is deprecated. Use "mongodb" instead.')
backend = 'mongodb'
mod = __import__('grab.spider.cache_backend.%s' % backend,
globals(), locals(), ['foo'])
backend = mod.CacheBackend(
database=database, spider=self, **kwargs
)
self.cache_reader_service = CacheReaderService(self, backend)
backend = mod.CacheBackend(
database=database, spider=self, **kwargs
)
self.cache_writer_service = CacheWriterService(self, backend)
def add_task(self, task):
"""
Add task to the task queue.
"""
if self.taskq is None:
raise SpiderMisuseError('You should configure task queue before '
'adding tasks. Use `setup_queue` method.')
if task.priority is None or not task.priority_is_custom:
task.priority = self.generate_task_priority()
task.priority_is_custom = False
else:
task.priority_is_custom = True
if not isinstance(task, NullTask):
if not task.url.startswith(
('http://', 'https://', 'ftp://', 'file://', 'feed://')):
if self.base_url is None:
msg = 'Could not resolve relative URL because base_url ' \
'is not specified. Task: %s, URL: %s'\
% (task.name, task.url)
logger.error(msg)
self.add_item('task-with-invalid-url', task.url)
return False
else:
task.url = urljoin(self.base_url, task.url)
# If task has grab_config object then update it too
if task.grab_config:
task.grab_config['url'] = task.url
# TODO: keep original task priority if it was set explicitly
self.taskq.put(task, task.priority, schedule_time=task.schedule_time)
is_valid = True
return is_valid
def setup_cache(self, backend='mongo', database=None, use_compression=True, **kwargs):
if database is None:
raise SpiderMisuseError('setup_cache method requires database option')
self.cache_enabled = True
mod = __import__('grab.spider.cache_backend.%s' % backend,
globals(), locals(), ['foo'])
self.cache = mod.CacheBackend(database=database, use_compression=use_compression,
spider=self, **kwargs)
def put(self, task, priority, schedule_time=None):
if schedule_time is not None:
raise SpiderMisuseError('Mongo task queue does not support delayed task')
item = {
'task': Binary(pickle.dumps(task)),
'priority': priority,
}
self.collection.save(item)
def load_proxylist(self,
source,
source_type=None,
proxy_type='http',
auto_init=True,
auto_change=True):
"""
Load proxy list.
:param source: Proxy source.
Accepts string (file path, url) or ``BaseProxySource`` instance.
:param source_type: The type of the specified source.
Should be one of the following: 'text_file' or 'url'.
:param proxy_type:
Should be one of the following: 'socks4', 'socks5' or'http'.
:param auto_change:
If set to `True` then automatical random proxy rotation
will be used.
Proxy source format should be one of the following (for each line):
- ip:port
- ip:port:login:password
"""
self.proxylist = ProxyList()
if isinstance(source, BaseProxySource):
self.proxylist.set_source(source)
elif isinstance(source, six.string_types):
if source_type == 'text_file':
self.proxylist.load_file(source, proxy_type=proxy_type)
elif source_type == 'url':
self.proxylist.load_url(source, proxy_type=proxy_type)
else:
raise SpiderMisuseError('Method `load_proxylist` received '
'invalid `source_type` argument: %s' %
source_type)
else:
raise SpiderMisuseError('Method `load_proxylist` received '
'invalid `source` argument: %s' % source)
self.proxylist_enabled = True
self.proxy = None
if not auto_change and auto_init:
self.proxy = self.proxylist.get_random_proxy()
self.proxy_auto_change = auto_change
def add_interface(self, backend=None, **kwargs):
if backend == 'redis':
iface = RedisCommandInterface(self.spider.get_name(), **kwargs)
self.ifaces[backend] = iface
self.enabled = True
return iface
else:
raise SpiderMisuseError('Unknown command interface: %s' % backend)
def put(self, task, priority, schedule_time=None):
if schedule_time is not None:
raise SpiderMisuseError('Redis task queue does not support '
'delayed task')
# Add attribute with random value
# This is required because qr library
# does not allow to store multiple values with same hash
# in the PriorityQueue
task.redis_qr_rnd = random.random()
self.queue_object.push(task, priority)
def add_task(self, task, raise_error=False):
"""
Add task to the task queue.
"""
# MP:
# ***
if self.parser_mode:
self.parser_result_queue.put((task, None))
return
if self.task_queue is None:
raise SpiderMisuseError('You should configure task queue before '
'adding tasks. Use `setup_queue` method.')
if task.priority is None or not task.priority_is_custom:
task.priority = self.generate_task_priority()
task.priority_is_custom = False
else:
task.priority_is_custom = True
try:
if not task.url.startswith(
('http://', 'https://', 'ftp://', 'file://', 'feed://')):
if self.base_url is None:
msg = 'Could not resolve relative URL because base_url ' \
'is not specified. Task: %s, URL: %s'\
% (task.name, task.url)
raise SpiderError(msg)
else:
warn('Class attribute `Spider::base_url` is deprecated. '
'Use Task objects with absolute URLs')
task.url = urljoin(self.base_url, task.url)
# If task has grab_config object then update it too
if task.grab_config:
task.grab_config['url'] = task.url
except Exception as ex:
self.stat.collect('task-with-invalid-url', task.url)
if raise_error:
raise
else:
logger.error('', exc_info=ex)
return False
# TODO: keep original task priority if it was set explicitly
self.task_queue.put(task,
task.priority,
schedule_time=task.schedule_time)
return True
def add_task(self, task, raise_error=False):
"""
Add task to the task queue.
"""
# MP:
# ***
if self.parser_mode:
self.parser_result_queue.put((task, None))
return
if self.task_queue is None:
raise SpiderMisuseError('You should configure task queue before '
'adding tasks. Use `setup_queue` method.')
if task.priority is None or not task.priority_set_explicitly:
task.priority = self.generate_task_priority()
task.priority_set_explicitly = False
else:
task.priority_set_explicitly = True
if not task.url.startswith(('http://', 'https://', 'ftp://',
'file://', 'feed://')):
self.stat.collect('task-with-invalid-url', task.url)
msg = ('It is not allowed to build Task object with '
'relative URL: %s' % task.url)
ex = SpiderError(msg)
if raise_error:
raise ex
else:
# Just want to print traceback
# Do this to avoid the error
# http://bugs.python.org/issue23003
# FIXME: use something less awkward
try:
raise ex
except SpiderError as ex:
logger.error('', exc_info=ex)
return False
# TODO: keep original task priority if it was set explicitly
# WTF the previous comment means?
self.task_queue.put(task, task.priority,
schedule_time=task.schedule_time)
return True
def add_task(self, task):
"""
Add task to the task queue.
Abort the task which was restarted too many times.
"""
if self.taskq is None:
raise SpiderMisuseError('You should configure task queue before adding tasks. Use `setup_queue` method.')
if task.priority is None or not task.priority_is_custom:
task.priority = self.generate_task_priority()
task.priority_is_custom = False
else:
task.priority_is_custom = True
if not isinstance(task, NullTask):
if not task.url.startswith(('http://', 'https://', 'ftp://', 'file://')):
if self.base_url is None:
#raise SpiderMisuseError('Could not resolve relative URL because base_url is not specified. Task: %s, URL: %s' % (task.name, task.url))
msg = 'Could not resolve relative URL because base_url is not specified. Task: %s, URL: %s' % (task.name, task.url)
logger.error(msg)
self.add_item('task-with-invalid-url', task.url)
return False
else:
task.url = urljoin(self.base_url, task.url)
# If task has grab_config object then update it too
if task.grab_config:
task.grab_config['url'] = task.url
if self.config.get('GRAB_TASK_REFRESH_CACHE', {}).get(task.name, False):
task.refresh_cache = True
if not self.config.get('TASK_ENABLED', {}).get(task.name, True):
logger.debug('Task %s disabled via config' % task.name)
self.inc_count('task-disabled-via-config')
is_valid = False
else:
# TODO: keep original task priority if it was set explicitly
self.taskq.put(task, task.priority, schedule_time=task.schedule_time)
is_valid = True
return is_valid
def __init__(
self,
thread_number=3,
network_try_limit=10,
task_try_limit=10,
request_pause=NULL,
priority_mode='random',
meta=None,
only_cache=False,
config=None,
slave=False,
max_task_generator_chunk=None,
args=None,
# New options start here
waiting_shutdown_event=None,
taskq=None,
result_queue=None,
network_response_queue=None,
shutdown_event=None,
generator_done_event=None,
ng=False,
):
"""
Arguments:
* thread-number - Number of concurrent network streams
* network_try_limit - How many times try to send request
again if network error was occurred, use 0 to disable
* network_try_limit - Limit of tries to execute some task
this is not the same as network_try_limit
network try limit limits the number of tries which
are performed automatically in case of network timeout
of some other physical error
but task_try_limit limits the number of attempts which
are scheduled manually in the spider business logic
* priority_mode - could be "random" or "const"
* meta - arbitrary user data
* retry_rebuild_user_agent - generate new random user-agent for each
network request which is performed again due to network error
* args - command line arguments parsed with `setup_arg_parser` method
New options:
* waiting_shutdown_event=None,
* taskq=None,
* result_queue=None,
* newtork_response_queue=None,
* shutdown_event=None,
* generator_done_event=None):
"""
# New options starts
self.waiting_shutdown_event = waiting_shutdown_event
self.taskq = taskq
self.result_queue = result_queue
self.shutdown_event = shutdown_event
self.generator_done_event = generator_done_event
self.network_response_queue = network_response_queue
self.ng = ng
# New options ends
if args is None:
self.args = {}
else:
self.args = args
self.slave = slave
self.max_task_generator_chunk = max_task_generator_chunk
self.timers = {
'network-name-lookup': 0,
'network-connect': 0,
'network-total': 0,
}
self.time_points = {}
self.start_timer('total')
if config is not None:
self.config = config
else:
# Fix circular import error
from grab.util.config import Config
self.config = Config()
if meta:
self.meta = meta
else:
self.meta = {}
self.task_generator_enabled = False
self.only_cache = only_cache
self.thread_number = thread_number
self.counters = defaultdict(int)
self._grab_config = {}
self.items = {}
self.task_try_limit = task_try_limit
self.network_try_limit = network_try_limit
if priority_mode not in ['random', 'const']:
raise SpiderMisuseError('Value of priority_mode option should be '
'"random" or "const"')
else:
self.priority_mode = priority_mode
try:
signal.signal(signal.SIGUSR1, self.sigusr1_handler)
except (ValueError, AttributeError):
pass
try:
signal.signal(signal.SIGUSR2, self.sigusr2_handler)
except (ValueError, AttributeError):
pass
# Initial cache-subsystem values
self.cache_enabled = False
self.cache = None
self.work_allowed = True
if request_pause is not NULL:
logger.error('Option `request_pause` is deprecated and is not '
'supported anymore')
self.proxylist_enabled = None
self.proxylist = None
self.proxy = None
self.proxy_auto_change = False
# FIXIT: REMOVE
self.dump_spider_stats = None
self.controller = CommandController(self)
# snapshots contains information about spider's state
# for each 10 seconds interval
self.snapshots = {}
self.last_snapshot_values = {
'timestamp': 0,
'download-size': 0,
'upload-size': 0,
'download-size-with-cache': 0,
'request-count': 0,
}
self.snapshot_timestamps = []
self.snapshot_interval = self.config.get('GRAB_SNAPSHOT_CONFIG',
{}).get('interval', 10)
self.snapshot_file = self.config.get('GRAB_SNAPSHOT_CONFIG',
{}).get('file', None)
if self.snapshot_file:
open(self.snapshot_file, 'w').write('')
def __init__(
self,
thread_number=None,
network_try_limit=None, task_try_limit=None,
request_pause=NULL,
priority_mode='random',
meta=None,
config=None,
args=None,
parser_requests_per_process=10000,
parser_pool_size=1,
http_api_port=None,
network_service='threaded',
grab_transport='pycurl',
# Deprecated
transport=None,
only_cache=False,
):
"""
Arguments:
* thread-number - Number of concurrent network streams
* network_try_limit - How many times try to send request
again if network error was occurred, use 0 to disable
* task_try_limit - Limit of tries to execute some task
this is not the same as network_try_limit
network try limit limits the number of tries which
are performed automatically in case of network timeout
of some other physical error
but task_try_limit limits the number of attempts which
are scheduled manually in the spider business logic
* priority_mode - could be "random" or "const"
* meta - arbitrary user data
* retry_rebuild_user_agent - generate new random user-agent for each
network request which is performed again due to network error
* args - command line arguments parsed with `setup_arg_parser` method
"""
self.fatal_error_queue = Queue()
self.task_queue_parameters = None
self.http_api_port = http_api_port
self._started = None
assert grab_transport in ('pycurl', 'urllib3')
self.grab_transport_name = grab_transport
self.parser_requests_per_process = parser_requests_per_process
self.stat = Stat()
self.task_queue = None
if args is None:
self.args = {}
else:
self.args = args
if config is not None:
self.config = config
else:
self.config = {}
if meta:
self.meta = meta
else:
self.meta = {}
self.thread_number = (
thread_number or
int(self.config.get('thread_number',
DEFAULT_NETWORK_STREAM_NUMBER)))
self.task_try_limit = (
task_try_limit or
int(self.config.get('task_try_limit', DEFAULT_TASK_TRY_LIMIT)))
self.network_try_limit = (
network_try_limit or
int(self.config.get('network_try_limit',
DEFAULT_NETWORK_TRY_LIMIT)))
self._grab_config = {}
if priority_mode not in ['random', 'const']:
raise SpiderMisuseError('Value of priority_mode option should be '
'"random" or "const"')
else:
self.priority_mode = priority_mode
if only_cache:
raise_feature_is_deprecated('Cache feature')
self.work_allowed = True
if request_pause is not NULL:
warn('Option `request_pause` is deprecated and is not '
'supported anymore')
self.proxylist_enabled = None
self.proxylist = None
self.proxy = None
self.proxy_auto_change = False
self.interrupted = False
self.parser_pool_size = parser_pool_size
self.parser_service = ParserService(
spider=self,
pool_size=self.parser_pool_size,
)
if transport is not None:
warn('The "transport" argument of Spider constructor is'
' deprecated. Use "network_service" argument.')
network_service = transport
assert network_service in ('threaded',)
if network_service == 'threaded':
# pylint: disable=no-name-in-module, import-error
from grab.spider.network_service.threaded import (
NetworkServiceThreaded
)
self.network_service = NetworkServiceThreaded(
self, self.thread_number
)
self.task_dispatcher = TaskDispatcherService(self)
if self.http_api_port:
self.http_api_service = HttpApiService(self)
else:
self.http_api_service = None
self.task_generator_service = TaskGeneratorService(
self.task_generator(), self,
)
def __init__(self, thread_number=None,
network_try_limit=None, task_try_limit=None,
request_pause=NULL,
priority_mode='random',
meta=None,
only_cache=False,
config=None,
slave=None,
args=None,
# New options start here
taskq=None,
# MP:
network_result_queue=None,
parser_result_queue=None,
is_parser_idle=None,
shutdown_event=None,
mp_mode=False,
parser_pool_size=None,
parser_mode=False,
parser_requests_per_process=10000,
# http api
http_api_port=None,
transport='multicurl',
grab_transport='pycurl',
):
"""
Arguments:
* thread-number - Number of concurrent network streams
* network_try_limit - How many times try to send request
again if network error was occurred, use 0 to disable
* network_try_limit - Limit of tries to execute some task
this is not the same as network_try_limit
network try limit limits the number of tries which
are performed automatically in case of network timeout
of some other physical error
but task_try_limit limits the number of attempts which
are scheduled manually in the spider business logic
* priority_mode - could be "random" or "const"
* meta - arbitrary user data
* retry_rebuild_user_agent - generate new random user-agent for each
network request which is performed again due to network error
* args - command line arguments parsed with `setup_arg_parser` method
New options:
* taskq=None,
* newtork_response_queue=None,
"""
if slave is not None:
raise SpiderConfigurtionError(
'Slave mode is not supported anymore. '
'Use `mp_mode=True` option to run multiple HTML'
' parser processes.')
# API:
self.http_api_port = http_api_port
assert transport in ('multicurl', 'threaded')
self.transport_name = transport
assert grab_transport in ('pycurl', 'urllib3')
self.grab_transport_name = grab_transport
# MP:
self.mp_mode = mp_mode
if self.mp_mode:
from multiprocessing import Process, Event, Queue
else:
from multiprocessing.dummy import Process, Event, Queue
if network_result_queue is not None:
self.network_result_queue = network_result_queue
else:
self.network_result_queue = Queue()
self.parser_result_queue = parser_result_queue
self.is_parser_idle = is_parser_idle
if shutdown_event is not None:
self.shutdown_event = shutdown_event
else:
self.shutdown_event = Event()
if not self.mp_mode and parser_pool_size and parser_pool_size > 1:
raise SpiderConfigurationError(
'Parser pool size could be only 1 in '
'non-multiprocess mode')
self.parser_pool_size = parser_pool_size
self.parser_mode = parser_mode
self.parser_requests_per_process = parser_requests_per_process
self.stat = Stat()
self.timer = Timer()
self.task_queue = taskq
if args is None:
self.args = {}
else:
self.args = args
if config is not None:
self.config = config
else:
self.config = {}
if meta:
self.meta = meta
else:
self.meta = {}
self.thread_number = (
thread_number or
int(self.config.get('thread_number',
DEFAULT_NETWORK_STREAM_NUMBER)))
self.task_try_limit = (
task_try_limit or
int(self.config.get('task_try_limit', DEFAULT_TASK_TRY_LIMIT)))
self.network_try_limit = (
network_try_limit or
int(self.config.get('network_try_limit',
DEFAULT_NETWORK_TRY_LIMIT)))
self._grab_config = {}
if priority_mode not in ['random', 'const']:
raise SpiderMisuseError('Value of priority_mode option should be '
'"random" or "const"')
else:
self.priority_mode = priority_mode
self.only_cache = only_cache
self.cache_pipeline = None
self.work_allowed = True
if request_pause is not NULL:
warn('Option `request_pause` is deprecated and is not '
'supported anymore')
self.proxylist_enabled = None
self.proxylist = None
self.proxy = None
self.proxy_auto_change = False
self.interrupted = False
def __init__(
self,
thread_number=None,
network_try_limit=None,
task_try_limit=None,
request_pause=NULL,
priority_mode='random',
meta=None,
only_cache=False,
config=None,
slave=False,
max_task_generator_chunk=None,
args=None,
# New options start here
taskq=None,
):
"""
Arguments:
* thread-number - Number of concurrent network streams
* network_try_limit - How many times try to send request
again if network error was occurred, use 0 to disable
* network_try_limit - Limit of tries to execute some task
this is not the same as network_try_limit
network try limit limits the number of tries which
are performed automatically in case of network timeout
of some other physical error
but task_try_limit limits the number of attempts which
are scheduled manually in the spider business logic
* priority_mode - could be "random" or "const"
* meta - arbitrary user data
* retry_rebuild_user_agent - generate new random user-agent for each
network request which is performed again due to network error
* args - command line arguments parsed with `setup_arg_parser` method
New options:
* taskq=None,
* newtork_response_queue=None,
"""
self.taskq = taskq
if args is None:
self.args = {}
else:
self.args = args
self.slave = slave
self.max_task_generator_chunk = max_task_generator_chunk
self.timers = {
'network-name-lookup': 0,
'network-connect': 0,
'network-total': 0,
}
self.time_points = {}
self.start_timer('total')
if config is not None:
self.config = config
else:
self.config = {}
if meta:
self.meta = meta
else:
self.meta = {}
self.task_generator_enabled = False
self.only_cache = only_cache
self.thread_number = thread_number or\
int(self.config.get('thread_number', 3))
self.task_try_limit = task_try_limit or\
int(self.config.get('task_try_limit', 10))
self.network_try_limit = network_try_limit or \
int(self.config.get('network_try_limit', 10))
self.counters = defaultdict(int)
self._grab_config = {}
self.items = {}
if priority_mode not in ['random', 'const']:
raise SpiderMisuseError('Value of priority_mode option should be '
'"random" or "const"')
else:
self.priority_mode = priority_mode
try:
signal.signal(signal.SIGUSR1, self.sigusr1_handler)
except (ValueError, AttributeError):
pass
try:
signal.signal(signal.SIGUSR2, self.sigusr2_handler)
except (ValueError, AttributeError):
pass
# Initial cache-subsystem values
self.cache_enabled = False
self.cache = None
self.work_allowed = True
if request_pause is not NULL:
logger.error('Option `request_pause` is deprecated and is not '
'supported anymore')
self.proxylist_enabled = None
self.proxylist = None
self.proxy = None
self.proxy_auto_change = False
self.interrupted = False
def __init__(
self,
name=None,
url=None,
grab=None,
grab_config=None,
priority=None,
priority_set_explicitly=True,
network_try_count=0,
task_try_count=1,
valid_status=None,
use_proxylist=True,
delay=None,
raw=False,
callback=None,
fallback_name=None,
# deprecated
disable_cache=False,
refresh_cache=False,
cache_timeout=None,
# kwargs
**kwargs):
"""
Create `Task` object.
If more than one of url, grab and grab_config options are non-empty
then they processed in following order:
* grab overwrite grab_config
* grab_config overwrite url
Args:
:param name: name of the task. After successful network operation
task's result will be passed to `task_<name>` method.
:param url: URL of network document. Any task requires `url` or
`grab` option to be specified.
:param grab: configured `Grab` instance. You can use that option in
case when `url` option is not enough. Do not forget to
configure `url` option of `Grab` instance because in this case
the `url` option of `Task` constructor will be overwritten
with `grab.config['url']`.
:param priority: - priority of the Task. Tasks with lower priority
will be processed earlier. By default each new task is assigned
with random priority from (80, 100) range.
:param priority_set_explicitly: - internal flag which tells if that
task priority was assigned manually or generated by spider
according to priority generation rules.
:param network_try_count: you'll probably will not need to use it.
It is used internally to control how many times this task was
restarted due to network errors. The `Spider` instance has
`network_try_limit` option. When `network_try_count` attribute
of the task exceeds the `network_try_limit` attribute then
processing of the task is abandoned.
:param task_try_count: the as `network_try_count` but it increased
only then you use `clone` method. Also you can set it manually.
It is useful if you want to restart the task after it was
cancelled due to multiple network errors. As you might guessed
there is `task_try_limit` option in `Spider` instance. Both
options `network_try_count` and `network_try_limit` guarantee
you that you'll not get infinite loop of restarting some task.
:param valid_status: extra status codes which counts as valid
:param use_proxylist: it means to use proxylist which was
configured via `setup_proxylist` method of spider
:param delay: if specified tells the spider to schedule the task
and execute it after `delay` seconds
:param raw: if `raw` is True then the network response is
forwarding to the corresponding handler without any check of
HTTP status code of network error, if `raw` is False (by
default) then failed response is putting back to task queue or
if tries limit is reached then the processing of this request
is finished.
:param callback: if you pass some function in `callback` option
then the network response will be passed to this callback and
the usual 'task_*' handler will be ignored and no error will be
raised if such 'task_*' handler does not exist.
:param fallback_name: the name of method that is called when spider
gives up to do the task (due to multiple network errors)
Any non-standard named arguments passed to `Task` constructor will
be saved as attributes of the object. You can get their values
later as attributes or with `get` method which allows to use
default value if attribute does not exist.
"""
if disable_cache or refresh_cache or cache_timeout:
raise_feature_is_deprecated('Cache feature')
if name == 'generator':
# The name "generator" is restricted because
# `task_generator` handler could not be created because
# this name is already used for special method which
# generates new tasks
raise SpiderMisuseError('Task name could not be "generator"')
self.name = name
if url is None and grab is None and grab_config is None:
raise SpiderMisuseError('Either url, grab or grab_config argument '
'of Task constructor should not be None')
if url is not None and grab is not None:
raise SpiderMisuseError('Options url and grab could not be used '
'together')
if url is not None and grab_config is not None:
raise SpiderMisuseError('Options url and grab_config could not be '
'used together')
if grab is not None and grab_config is not None:
raise SpiderMisuseError(
'Options grab and grab_config could not be used together')
if grab:
self.setup_grab_config(grab.dump_config())
elif grab_config:
self.setup_grab_config(grab_config)
else:
self.grab_config = None
self.url = url
if valid_status is None:
self.valid_status = []
else:
self.valid_status = valid_status
self.process_delay_option(delay)
self.fallback_name = fallback_name
self.priority_set_explicitly = priority_set_explicitly
self.priority = priority
self.network_try_count = network_try_count
self.task_try_count = task_try_count
self.use_proxylist = use_proxylist
self.raw = raw
self.callback = callback
self.coroutines_stack = []
for key, value in kwargs.items():
setattr(self, key, value)