def open_spider(self, spider):
self.enabled = self.is_enabled(spider)
self.spider = spider
for k, type_ in self._settings:
setattr(self, k, self._get_setting_value(spider, k, type_))
self._headers = self.crawler.settings.get('CRAWLERA_DEFAULT_HEADERS',
{}).items()
self.exp_backoff = exp_backoff(self.backoff_step, self.backoff_max)
if not self.enabled and not self.force_enable_on_http_codes:
return
if not self.apikey:
logging.warning("Crawlera can't be used without a APIKEY",
extra={'spider': spider})
return
self._proxyauth = self.get_proxyauth(spider)
logging.info(
"Using crawlera at %s (apikey: %s)" % (self.url, self.apikey[:7]),
extra={'spider': spider},
)
if not self.preserve_delay:
# Setting spider download delay to 0 to get maximum crawl rate
spider.download_delay = 0
logging.info(
"CrawleraMiddleware: disabling download delays on Scrapy side to optimize delays introduced by Crawlera. "
"To avoid this behaviour you can use the CRAWLERA_PRESERVE_DELAY setting but keep in mind that this may slow down the crawl significantly",
extra={'spider': spider},
)
def process_response(self, request, response, spider):
if not self._is_enabled_for_request(request):
return response
key = self._get_slot_key(request)
self._restore_original_delay(request)
if self._is_no_available_proxies(response):
self._set_custom_delay(request, next(self.exp_backoff))
else:
self.exp_backoff = exp_backoff(self.backoff_step, self.backoff_max)
if self._is_banned(response):
self._bans[key] += 1
if self._bans[key] > self.maxbans:
self.crawler.engine.close_spider(spider, 'banned')
else:
after = response.headers.get('retry-after')
if after:
self._set_custom_delay(request, float(after))
self.crawler.stats.inc_value('crawlera/response/banned')
else:
self._bans[key] = 0
# If placed behind `RedirectMiddleware`, it would not count 3xx responses
self.crawler.stats.inc_value('crawlera/response')
self.crawler.stats.inc_value('crawlera/response/status/%s' %
response.status)
crawlera_error = response.headers.get('X-Crawlera-Error')
if crawlera_error:
self.crawler.stats.inc_value('crawlera/response/error')
self.crawler.stats.inc_value('crawlera/response/error/%s' %
crawlera_error.decode('utf8'))
return response
def process_response(self, request, response, spider):
if not self._is_enabled_for_request(request):
return response
key = self._get_slot_key(request)
self._restore_original_delay(request)
if self._is_no_available_proxies(response):
self._set_custom_delay(request, next(self.exp_backoff))
else:
self.exp_backoff = exp_backoff(self.backoff_step, self.backoff_max)
if self._is_banned(response):
self._bans[key] += 1
if self._bans[key] > self.maxbans:
self.crawler.engine.close_spider(spider, 'banned')
else:
after = response.headers.get('retry-after')
if after:
self._set_custom_delay(request, float(after))
self.crawler.stats.inc_value('crawlera/response/banned')
else:
self._bans[key] = 0
# If placed behind `RedirectMiddleware`, it would not count 3xx responses
self.crawler.stats.inc_value('crawlera/response')
self.crawler.stats.inc_value('crawlera/response/status/%s' % response.status)
crawlera_error = response.headers.get('X-Crawlera-Error')
if crawlera_error:
self.crawler.stats.inc_value('crawlera/response/error')
self.crawler.stats.inc_value(
'crawlera/response/error/%s' % crawlera_error.decode('utf8'))
return response
def process_response(self, request, response, spider):
if not self._is_enabled_for_request(request):
return self._handle_not_enabled_response(request, response)
if not self._is_crawlera_response(response):
return response
key = self._get_slot_key(request)
self._restore_original_delay(request)
if self._is_no_available_proxies(response) or self._is_auth_error(response):
if self._is_no_available_proxies(response):
reason = 'noslaves'
else:
reason = 'autherror'
self._set_custom_delay(request, next(self.exp_backoff), reason=reason)
else:
self.crawler.stats.inc_value('crawlera/delay/reset_backoff')
self.exp_backoff = exp_backoff(self.backoff_step, self.backoff_max)
if self._is_auth_error(response):
# When crawlera has issues it might not be able to authenticate users
# we must retry
retries = request.meta.get('crawlera_auth_retry_times', 0)
if retries < self.max_auth_retry_times:
return self._retry_auth(response, request, spider)
else:
self.crawler.stats.inc_value('crawlera/retries/auth/max_reached')
logging.warning(
"Max retries for authentication issues reached, please check auth"
" information settings",
extra={'spider': self.spider},
)
if self._is_banned(response):
self._bans[key] += 1
if self._bans[key] > self.maxbans:
self.crawler.engine.close_spider(spider, 'banned')
else:
after = response.headers.get('retry-after')
if after:
self._set_custom_delay(request, float(after), reason='banned')
self.crawler.stats.inc_value('crawlera/response/banned')
else:
self._bans[key] = 0
# If placed behind `RedirectMiddleware`, it would not count 3xx responses
self.crawler.stats.inc_value('crawlera/response')
self.crawler.stats.inc_value('crawlera/response/status/%s' % response.status)
crawlera_error = response.headers.get('X-Crawlera-Error')
if crawlera_error:
self.crawler.stats.inc_value('crawlera/response/error')
self.crawler.stats.inc_value(
'crawlera/response/error/%s' % crawlera_error.decode('utf8'))
return response
def open_spider(self, spider):
self.enabled = self.is_enabled(spider)
if not self.enabled:
return
for k, type_ in self._settings:
setattr(self, k, self._get_setting_value(spider, k, type_))
self._proxyauth = self.get_proxyauth(spider)
logging.info("Using crawlera at %s (apikey: %s)" % (
self.url,
self.apikey[:7])
)
if not self.preserve_delay:
# Setting spider download delay to 0 to get maximum crawl rate
spider.download_delay = 0
logging.info(
"CrawleraMiddleware: disabling download delays on Scrapy side to optimize delays introduced by Crawlera. "
"To avoid this behaviour you can use the CRAWLERA_PRESERVE_DELAY setting but keep in mind that this may slow down the crawl significantly")
self._headers = self.crawler.settings.get('CRAWLERA_DEFAULT_HEADERS', {}).items()
self.exp_backoff = exp_backoff(self.backoff_step, self.backoff_max)
