def parse_seed(self, response):
"""
Parse a webpage from the "seed" website.
"""
ld = self._load_webpage_item(response, is_seed=True)
if self.use_splash:
self._process_splash_response(response, ld)
yield ld.load_item()
this_domain = get_domain(response.url)
for link in self._get_links(response):
domain = get_domain(link.url)
if is_external_url(response.url, link.url):
yield self._offsite_request(
response, link,
count_key=(this_domain, domain),
max_count=self.max_external_links_per_seed_per_domain
)
else:
yield self._onsite_request(
response, link,
callback=self.parse,
max_depth=self.max_depth_seed,
count_key=domain,
max_count=self.max_internal_links_per_seed,
)
def parse_external(self, response):
"""
Parse a webpage from an external website.
"""
ld = self._load_webpage_item(response, is_seed=False)
if self.use_splash:
self._process_splash_response(response, ld)
yield ld.load_item()
for link in self._get_links(response):
domain = get_domain(link.url)
if is_external_url(response.url, link.url):
# total number of hops is limited by settings.DEPTH_LIMIT
yield self._offsite_request(
response, link,
count_key=domain,
max_count=self.max_external_links_per_domain
)
else:
yield self._onsite_request(
response, link,
callback=self.parse_external,
max_depth=self.max_depth_external,
count_key=domain,
max_count=self.max_external_links_per_domain
)
def _process_splash_response(self, response, splash_response, ld):
data = json.loads(splash_response.body, encoding='utf8')
screenshot_path = self._save_screenshot(get_domain(response.url), data)
ld.add_value('screenshot_path', screenshot_path)
if self.save_html:
ld.add_value('html_rendered', data['html'])
def _process_splash_response(self, response, ld):
screenshot_path = save_screenshot(
screenshot_dir=self.screenshot_dir,
prefix=get_domain(response.url),
png=base64.b64decode(response.meta['splash_response']['png']),
)
ld.add_value('screenshot_path', screenshot_path)
if self.save_html:
ld.add_value('html_rendered', response.meta['splash_response']['html'])
def _load_webpage_item(self, response, is_seed):
depth = response.meta.get('link_depth', 0)
ld = WebpageItemLoader(response=response)
ld.add_value('url', response.url)
ld.add_value('host', get_domain(response.url))
ld.add_xpath('title', '//title/text()')
ld.add_value('depth', depth)
ld.add_value('total_depth', response.meta.get('depth'))
ld.add_value('crawled_at', datetime.datetime.utcnow())
ld.add_value('is_seed', is_seed)
if self.save_html:
ld.add_value('html', response.body_as_unicode())
if 'link' in response.meta:
link = response.meta['link']
ld.add_value('link_text', link.text)
ld.add_value('link_url', link.url)
ld.add_value('referrer_url', response.meta['referrer_url'])
ld.add_value('referrer_depth', response.meta['referrer_depth'])
return ld
