def __init__ (self, RESTART_LIMIT=100, VISIT_MEMORY=100):
"""
RESTART_LIMIT: how many pages to visit before doing a random restart.
VISIT_MEMORY: max pages to keep in a tabu list, i.e. a list of
pages not to visit again. Prevents cycles.
"""
self.current_url = None
self.current_html = None
self.pages_crawled = 0
self.next_urls = None
self.visited = SetQueue(VISIT_MEMORY)
self.RESTART_LIMIT = RESTART_LIMIT
class Crawler (object):
__metaclass__ = abc.ABCMeta
def __init__ (self, RESTART_LIMIT=100, VISIT_MEMORY=100):
"""
RESTART_LIMIT: how many pages to visit before doing a random restart.
VISIT_MEMORY: max pages to keep in a tabu list, i.e. a list of
pages not to visit again. Prevents cycles.
"""
self.current_url = None
self.current_html = None
self.pages_crawled = 0
self.next_urls = None
self.visited = SetQueue(VISIT_MEMORY)
self.RESTART_LIMIT = RESTART_LIMIT
def next_page (self):
if self.current_url is not None and self.next_urls is None:
self._fetch_next_urls_from_current_page()
found_page = False
while not found_page:
if (self.current_url is None or
self._restart_required() or
self._no_new_pages_available_from_current()):
new_page = self._random_restart()
else:
new_page = self.next_urls.pop()
if new_page not in self.visited:
found_page = True
self.current_url = new_page
self.current_html = None
self.next_urls = None
self.visited.add(new_page)
self.pages_crawled += 1
@abc.abstractmethod
def _fetch_next_urls_from_current_page (self):
"""Scrape current page for possible URLs to crawl next.
Should form a list of possible URLs to be crawled next, using the
HTML tree of the current page. Then assign self.next_urls to hold this
list. Shuffle it as well, if randomness is desired.
Returns:
None
"""
return
@staticmethod
def _fetch_page_content (f):
"""Decorator - ensures HTML tree retrieval before function body."""
@wraps(f)
def wrapped (inst, *args, **kwargs):
if inst.current_html is None:
page = req.get(inst.current_url)
inst.current_html = html.fromstring(page.content)
return f(inst, *args, **kwargs)
return wrapped
@abc.abstractmethod
def _random_restart (self):
"""Return link to a 'random' page.
Returns:
Link to a new page to be crawled.
"""
return
@abc.abstractmethod
def scrape_page (self):
"""Scrape data from the current page.
Note: maybe shouldn't require this implementation?
Returns:
Data from the current page, in whatever format.
"""
return
def _restart_required (self):
return self.pages_crawled % self.RESTART_LIMIT == 0
def _no_new_pages_available_from_current (self):
return self.next_urls == []
