def processing_products(self, spider):
if self != spider or self.skus_parsed:
return
request = Request(self.start_urls[0], callback=self.yield_products, dont_filter=True)
self.crawler.engine.crawl(request, spider)
raise DontCloseSpider()
def spider_idle(self, spider):
for request in self._read_queue():
self.crawler.engine.crawl(request, self)
raise DontCloseSpider()
def spider_idle(cls, spider):
if cls.requests.get(spider):
spider.log("delayed requests pending, not closing spider")
raise DontCloseSpider()
def keep_spider_alive(self, spider):
raise DontCloseSpider("keeping it open")
def _spider_idle(self):
self._schedule_next_request()
raise DontCloseSpider()
def _logout(self, spider):
if spider != self: return
if self.logout_done: return
if not self.logout_url: return
self.crawler.engine.schedule(self.logout(), spider)
raise DontCloseSpider('logout scheduled')
def spider_idle(self, spider):
if spider != self: return
if self.items:
self.crawler.engine.schedule(Request('http://' + self.allowed_domains[0], callback=self.yield_product, dont_filter=True), spider)
raise DontCloseSpider('Found pending requests')
def spider_idle(self, spider):
raise DontCloseSpider('waiting for process')
def parse_article(self, response):
comment_item = CommentItem()
post_item = PostItem()
# define the next page url
try:
next_page = response.xpath(
"//div[@id='action-bar-container']//a[contains(text(), '上頁')]/@href"
)[0]
except:
next_page = None
logging.error('Cannot load the next page')
# see the number of articles
try:
article_list = response.css('.r-list-container > div')
except:
logging.error('There is not exist any article being loaded')
else:
logging.info('the number of articles is {}'.format(
len(article_list)))
# check all tags of div
while len(article_list) > 0:
div = article_list.pop(0)
try:
# get class name inside div tag
slot_name = div.xpath('@class')[0].extract()
# get url inside article respectively
canonicalUrl = response.urljoin(
div.css('.title a::attr(href)')[0].extract())
# get author inside article respectively
author_str = div.css('.author::text')[0].extract()
except:
logging.error('Fail to access url, author and classes')
else:
if slot_name == 'r-list-sep':
'''
Once receiving class='r-list-sep', we are going to crawl the next page.
'''
if next_page: # exists the next page
logging.warning('redirect to following {}'.format(
canonicalUrl))
yield scrapy.Request(canonicalUrl,
callback=self.parse_article,
dont_filter=True)
else: # there is no following page, we stop the spider and wait for a new request
raise CloseSpider('page exceeded')
break # Then we needn't search the following
else: # r-ent' or 'search-bar':
'''
load each article url and find information such as content, comment, author, etc..
'''
if slot_name != 'r-ent': continue # 'search-bar'
## try to access articles which match the period we defined
date_str = div.css('.date::text')[0].extract()
m_d = tuple(map(int, date_str.split('/')))
m_d = datetime(self.year, m_d[0], m_d[1])
in_period = self.start <= m_d <= self.end
logging.debug(
'the date is {}/{}, and reach or not {}'.format(
m_d.month, m_d.day, in_period))
if in_period:
# First get items in an article before loading url inside
try:
post_item['canonicalUrl'] = canonicalUrl
comment_item['canonicalUrl'] = canonicalUrl
post_item['authorId'] = author_str
post_item['title'] = div.css(
'.title a::text')[0].extract()
post_item['publishedTime'] = date_str
post_item['board'] = self.board
except IndexError:
logging.error('Fail to save object (postItem)')
# This part we try to get in url corresponding to article
try:
url = response.urljoin(
div.css('.title a::attr(href)')
[0].extract())
logging.info(
'load url inside every article:{}'.format(
url))
yield scrapy.Request(
url,
meta={
'post': post_item,
'comment': comment_item
},
callback=self.parse_comment,
dont_filter=True)
except IndexError:
logging.error('Cannot load article')
else:
self.maximum_missing_count -= 1
if len(article_list) == 0:
'''
Let's go to the next page since we have explored every article
'''
if next_page and self.maximum_missing_count > 0: # exists the next page
url = response.urljoin(next_page.extract())
logging.warning('redirect to following {}'.format(url))
yield scrapy.Request(url,
callback=self.parse_article,
dont_filter=True)
else: # there is no following page, we stop the spider and wait for a new request
logging.error('Without the next page')
# endless running
raise DontCloseSpider('page exceeded')
def _dont_close_me(self, spider):
raise DontCloseSpider("..I prefer live spiders.")
def spider_idle(self, spider):
"""当spider收到idle信号时, 去ZooKeeper上获取一个任务, 再抛出DontCloseSpider异常"""
self.request_new_job(spider)
raise DontCloseSpider()
def signal_dispatcher(self, signal):
if self.signals_callback:
if signal == signals.spider_idle or signal == signals.spider_error:
raise DontCloseSpider('I prefer live spiders')
elif signal == signals.spider_opened:
self.signals_callback(signal, spider=self)
def leftover_requests(self, spider):
if self.resume_iter:
self.defrost_in_batch(spider)
raise DontCloseSpider()
def spider_idle(self, spider):
if self.rountine_interval:
self.schedule_rountine_requests(spider)
raise DontCloseSpider()
