class SpiderCore(object):
def __init__(self,
processor=None,
downloader=None,
use_proxy=False,
scheduler=None,
batch_size=None,
time_sleep=None,
test=False):
# 用于测试,爬取成功第一个以后结束
self.test = test
self._processor = processor
self._host_regex = self._get_host_regex()
self._spider_status = 'stopped'
# self._pipelines = {}
self._time_sleep = time_sleep
self._batch_size = 1
if time_sleep:
self._batch_size = 1
else:
if isinstance(downloader, SeleniumDownLoader):
self._batch_size = default_settings.DRIVER_POOL_SIZE - 1
else:
if batch_size:
self._batch_size = batch_size
else:
self._batch_size = 10
self._spider_id = processor.spider_id
self._process_count = 0
if not downloader:
self._downloader = RequestsDownLoader(use_proxy=use_proxy)
elif isinstance(downloader, SeleniumDownLoader):
self._downloader = downloader
self._batch_size = default_settings.DRIVER_POOL_SIZE - 1
else:
self._downloader = downloader
if not scheduler:
self._queue = PriorityQueue(self._processor)
else:
self._queue = scheduler
def create(self, processor):
self._processor = processor
return self
def set_scheduler(self, scheduler):
self._queue = scheduler
return self
def set_downloader(self, downloader):
self._downloader = downloader
if isinstance(downloader, SeleniumDownLoader):
self._batch_size = default_settings.DRIVER_POOL_SIZE - 1
return self
def stop(self):
if self._spider_status == 'stopped':
self._processor.logger.info("STOP %s SUCCESS" % self._spider_id)
return
elif self._spider_status == 'stopping':
while self._spider_status == 'stopping':
pass
elif self._spider_status == 'start':
self._spider_status = 'stopping'
while self._spider_status == 'stopping':
pass
def start(self):
try:
self._processor.logger.info("START %s SUCCESS" % self._spider_id)
self._spider_status = 'start'
self._queue = PriorityQueue(self._processor)
# print(self._processor.spider_start_requests)
if not self._processor.start_requests:
self._processor.init_start_requests()
print('START_REQUESTS COUNT ::: %s' %
str(len(self._processor.start_requests)))
for start_request in self._processor.start_requests:
if self._should_follow(start_request):
start_request.duplicate_remove = False
self._queue.push(start_request)
self._processor.logger.info("start request:>>>>>>" +
str(start_request))
# 去除所有添加到queue的start_request， 防止污染后续的processor
self._processor.start_requests.clear()
for batch in self._batch_requests():
if len(batch) > 0:
self._crawl(batch)
if self.test:
if self._process_count > 0:
return
if self._spider_status == 'stopping':
break
self._spider_status = 'stopped'
self._processor.logger.info("STOP %s SUCCESS" % self._spider_id)
except Exception:
self._processor.logger.info(
"%s -- Exception -- Stopped -- %s" %
(self._spider_id, traceback.format_exc()))
self._spider_status = 'stopped'
def restart(self):
self._queue = PriorityQueue(self._processor)
self._queue.clear()
self.start()
def _batch_requests(self):
batch = []
count = 0
retry = 0
try:
while True:
count += 1
if len(batch) >= self._batch_size or count >= self._batch_size:
batch.sort(key=_priority_compare_key, reverse=True)
yield batch
batch = []
count = 0
temp_request = self._queue.pop()
queue_count = len(self._queue)
if temp_request is not None:
if not temp_request.callback:
temp_request.callback = self._processor.process
batch.append(temp_request)
elif len(batch) == 0 and queue_count == 0: # 等待几次，看队列中是否还有
time.sleep(self._time_sleep)
retry += 1
if retry >= 3:
return []
except KeyboardInterrupt:
pass
def _crawl(self, batch):
responses = self._downloader.download(batch)
if self._time_sleep:
time.sleep(self._time_sleep)
for response in responses:
self._processor.logger.info(response)
callback = response.request.callback(response)
if isinstance(callback, types.GeneratorType):
pipe = self._queue.get_pipe()
for item in callback:
if isinstance(item, Request):
# logger.info("push request to queue..." + str(item))
if self._should_follow(item):
self._queue.push_pipe(item, pipe)
elif isinstance(item, PipeItem):
# 如果返回对象是pipeItem，则用对应的pipeline处理
self._process_count += 1
for pipe_name in item.pipe_names:
queue_job(PIPELINE_TASK,
PipelineArgs(pipe_name, item.result),
queue=PIPELINE)
if self.test:
if self._process_count > 0:
return
elif isinstance(
item,
Violet): # 如果返回的是tuple，即详情页的processor和详情页的请求信息
queue_job(CRAWLER_TASK,
CrawlArgs(item.processor, item.request),
queue=CRAWLER)
else:
raise Exception('not return correct value!!!')
pipe.execute()
elif isinstance(callback, Request):
# logger.info("push request to queue..." + str(back))
if self._should_follow(callback):
self._queue.push(callback)
elif isinstance(callback, PipeItem):
# 如果返回对象是pipeItem，则用对应的pipeline处理
self._process_count += 1
for pipe_name in callback.pipe_names:
queue_job(PIPELINE_TASK,
PipelineArgs(pipe_name, callback.result),
queue=PIPELINE)
elif isinstance(callback,
Violet): # 如果返回的是tuple，即详情页的processor和详情页的请求信息
queue_job(CRAWLER_TASK,
CrawlArgs(item.processor, item.request),
queue=CRAWLER)
else:
# # 如果返回对象不是pipeItem，则默认用每个pipeline处理
raise Exception('not return correct value!!!')
def _should_follow(self, request):
regex = self._host_regex
# hostname can be None for wrong urls (like javascript links)
host = urlparse_cached(request).hostname or ''
return bool(regex.search(host))
def _get_host_regex(self):
"""Override this method to implement a different offsite policy"""
allowed_domains = getattr(self._processor, 'allowed_domains', None)
if not allowed_domains:
return re.compile('') # allow all by default
regex = r'^(.*\.)?(%s)$' % '|'.join(
re.escape(d) for d in allowed_domains if d is not None)
return re.compile(regex)
class SpiderCore(object):
def __init__(self,
processor=None,
downloader=None,
use_proxy=False,
scheduler=None,
batch_size=None,
time_sleep=None,
test=False):
# 用于测试,爬取成功第一个以后结束
self.test = test
self._processor = processor
FetchManLogger.init_logger(processor.spider_id)
self._host_regex = self._get_host_regex()
self._spider_status = 'stopped'
self._pipelines = {}
self._time_sleep = time_sleep
if time_sleep:
self._batch_size = 0
else:
if isinstance(downloader, SeleniumDownLoader):
self._batch_size = default_settings.DRIVER_POOL_SIZE - 1
else:
if batch_size:
self._batch_size = batch_size - 1
else:
self._batch_size = 9
self._spider_id = processor.spider_id
self._process_count = 0
if not downloader:
self._downloader = RequestsDownLoader(use_proxy=use_proxy)
elif isinstance(downloader, SeleniumDownLoader):
self._downloader = downloader
self._batch_size = default_settings.DRIVER_POOL_SIZE - 1
else:
self._downloader = downloader
if not scheduler:
self._queue = PriorityQueue(self._processor)
else:
self._queue = scheduler
def create(self, processor):
self._processor = processor
return self
def set_scheduler(self, scheduler):
self._queue = scheduler
return self
def set_downloader(self, downloader):
self._downloader = downloader
if isinstance(downloader, SeleniumDownLoader):
self._batch_size = default_settings.DRIVER_POOL_SIZE - 1
return self
def set_pipeline(
self,
pipeline=None,
pipeline_name=None,
):
if not pipeline_name:
pipeline_name = str(uuid.uuid1())
self._pipelines[pipeline_name] = pipeline
return self
def stop(self):
if self._spider_status == 'stopped':
FetchManLogger.logger.info("STOP %s SUCCESS" % self._spider_id)
return
elif self._spider_status == 'stopping':
while self._spider_status == 'stopping':
pass
elif self._spider_status == 'start':
self._spider_status = 'stopping'
while self._spider_status == 'stopping':
pass
def start(self):
try:
FetchManLogger.logger.info("START %s SUCCESS" % self._spider_id)
self._spider_status = 'start'
# 启动爬虫记录日志
self._queue = PriorityQueue(self._processor)
if not self._processor.start_requests:
self._processor.init_start_requests()
for start_request in self._processor.start_requests:
if self._should_follow(start_request):
start_request.duplicate_remove = False
self._queue.push(start_request)
FetchManLogger.logger.info("start request:" +
str(start_request))
for batch in self._batch_requests():
print(batch)
if len(batch) > 0:
self._crawl(batch)
if self.test:
if self._process_count > 0:
return
if self._spider_status == 'stopping':
break
# 循环结束，记录停止日志
self._process_count = 0
self._spider_status = 'stopped'
FetchManLogger.logger.info("STOP %s SUCCESS" % self._spider_id)
except Exception:
FetchManLogger.logger.info(
"%s -- Exception -- Stopped -- %s" %
(self._spider_id, traceback.format_exc()))
self._spider_status = 'stopped'
def restart(self):
self._queue = PriorityQueue(self._processor)
self._queue.clear()
self.start()
def _batch_requests(self):
batch = []
count = 0
while True:
count += 1
temp_request = self._queue.pop()
if temp_request:
if not temp_request.callback:
temp_request.callback = self._processor.process
batch.append(temp_request)
if len(batch) > self._batch_size or count > self._batch_size:
if sys.version_info < (3, 0):
batch.sort(_priority_compare)
else:
batch.sort(key=_priority_compare_key, reverse=True)
if len(batch) > 0:
yield batch
else:
break
batch = []
count = 0
def _crawl(self, batch):
responses = self._downloader.download(batch)
if self._time_sleep:
time.sleep(self._time_sleep)
for response in responses:
callback = response.request.callback(response)
if isinstance(callback, types.GeneratorType):
pipe = self._queue.get_pipe()
for item in callback:
if isinstance(item, Request):
# logger.info("push request to queue..." + str(item))
if self._should_follow(item):
self._queue.push_pipe(item, pipe)
else:
if isinstance(item, pipeItem):
# 如果返回对象是pipeItem，则用对应的pipeline处理
self._process_count += 1
for pipename in item.pipenames:
if pipename in self._pipelines:
self._pipelines[pipename].process_item(
item.result)
if self.test:
if self._process_count > 0:
return
else:
# 如果返回对象不是pipeItem，则默认用每个pipeline处理
self._process_count += 1
for pipeline in self._pipelines.itervalues():
pipeline.process_item(item)
if self.test:
if self._process_count > 0:
return
pipe.execute()
elif isinstance(callback, Request):
# logger.info("push request to queue..." + str(back))
if self._should_follow(callback):
self._queue.push(callback)
elif isinstance(callback, pipeItem):
# 如果返回对象是pipeItem，则用对应的pipeline处理
self._process_count += 1
for pipename in callback.pipenames:
if pipename in self._pipelines:
self._pipelines[pipename].process_item(callback.result)
else:
# 如果返回对象不是pipeItem，则默认用每个pipeline处理
self._process_count += 1
for pipeline in self._pipelines.itervalues:
pipeline.process_item(item)
if self.test:
if self._process_count > 0:
return
def _should_follow(self, request):
regex = self._host_regex
# hostname can be None for wrong urls (like javascript links)
host = urlparse_cached(request).hostname or ''
return bool(regex.search(host))
def _get_host_regex(self):
"""Override this method to implement a different offsite policy"""
allowed_domains = getattr(self._processor, 'allowed_domains', None)
if not allowed_domains:
return re.compile('') # allow all by default
regex = r'^(.*\.)?(%s)$' % '|'.join(
re.escape(d) for d in allowed_domains if d is not None)
return re.compile(regex)