def search(request):
profiles = SpiderProfile.objects.filter(active=True)
vars = {'profiles':profiles}
page_index = WhooshPageIndex()
query = request.GET.get('q', None)
if query:
query = query.replace('+', ' AND ').replace(' -', ' NOT ') #this is fairly lame
results = page_index.search(query, limit=100)
vars['query'] = query
vars['results'] = [ dict(title=r['title'], highlight=r.highlights('content'), url=r['url']) for r in results]
vars['num_results'] =len(results)
return render(request, 'search_results.html', vars)
def spider(profile, log=True):
depth = profile.depth
indexer = WhooshPageIndex()
pending_urls = Queue.Queue()
processed_responses = Queue.Queue()
finished = threading.Event()
visited = {}
scheduled = set()
thread_count = profile.threads or getattr(settings, 'SPIDER_THREADS', 4)
threads = [SpiderThread(pending_urls, processed_responses, finished, profile) for x in range(thread_count)]
pending_urls.put((profile.base_url, '', depth))
scheduled.add(profile.base_url)
extractor = profile.get_extractor_class()
[t.start() for t in threads]
processed_url = None
try:
while 1:
try:
# pull an item from the response queue
result_dict, urls, depth = processed_responses.get(timeout=profile.timeout)
except Queue.Empty:
#check to see if any of the workers are still doing anything
done = True
for t in threads:
if t.working:
print "Thread %s is still working, not exiting" % t
done = False
if done:
print "All threads done working"
finished.set()
break
else:
# save the result
processed_url = result_dict['url']
if log:
print "Adding page at url: %s, content length: %s to index" % (processed_url, len(result_dict['content']))
raw_content = result_dict['content']
e = extractor(raw_content)
title = e.get_title()
content = e.get_content()
headings = e.get_headings()
indexer.add_page(url=processed_url, title=title, content=content, site=profile.name, headings=headings)
# remove from the list of scheduled items
scheduled.remove(processed_url)
# store response status in the visited dictionary
visited[processed_url] = result_dict['response_status']
# enqueue any urls that need to be checked
if depth > 0:
for url in urls:
if url not in visited and url not in scheduled:
scheduled.add(url)
pending_urls.put((url, processed_url, depth - 1))
except KeyboardInterrupt:
pass
except Exception, e:
print "Got an exception while indexing page: %s\nWill exit." % processed_url
print "Headings: %s" % headings
raise
