def searchPage():
query = request.GET.get('query', '').strip()
print query
if query:
db = db_lib.sql3()
results = db.doSearch(query)
else:
results = None
s = template('main', results=results)
return s
def searchPage():
query = request.GET.get('query', '').strip()
print query
if query:
db = db_lib.sql3()
results = db.doSearch(query)
else:
results = None
s = template('main', results=results)
return s
def setUp(self):
self.db = db_lib.sql3({'clean': True, 'location': 'test.db'})
def setUp(self):
self.db = db_lib.sql3({"clean": True, "location": "test.db"})
def __init__(self, url_file):
"""Initialize the crawler with a connection to the database to populate
and with the file containing the list of seed URLs to begin indexing."""
#Do direct db store instead of memory cache.
#Reasons..
#Not guaranteed enough memory for very large indexes
#Have to rebuild cache every time crawler is run.. Crawler should be updated, not rewritten
#End result, will be slightly slower due to INSERT IF NOT EXISTS mysql commands.
self._url_queue = [ ]
# functions to call when entering and exiting specific tags
self._enter = defaultdict(lambda *a, **ka: self._visit_ignore)
self._exit = defaultdict(lambda *a, **ka: self._visit_ignore)
# add a link to our graph, and indexing info to the related page
self._enter['a'] = self._visit_a
# record the currently indexed document's title an increase
# the font size
def visit_title(*args, **kargs):
self._visit_title(*args, **kargs)
self._increase_font_factor(7)(*args, **kargs)
# increase the font size when we enter these tags
self._enter['b'] = self._increase_font_factor(2)
self._enter['strong'] = self._increase_font_factor(2)
self._enter['i'] = self._increase_font_factor(1)
self._enter['em'] = self._increase_font_factor(1)
self._enter['h1'] = self._increase_font_factor(7)
self._enter['h2'] = self._increase_font_factor(6)
self._enter['h3'] = self._increase_font_factor(5)
self._enter['h4'] = self._increase_font_factor(4)
self._enter['h5'] = self._increase_font_factor(3)
self._enter['title'] = visit_title
# decrease the font size when we exit these tags
self._exit['b'] = self._increase_font_factor(-2)
self._exit['strong'] = self._increase_font_factor(-2)
self._exit['i'] = self._increase_font_factor(-1)
self._exit['em'] = self._increase_font_factor(-1)
self._exit['h1'] = self._increase_font_factor(-7)
self._exit['h2'] = self._increase_font_factor(-6)
self._exit['h3'] = self._increase_font_factor(-5)
self._exit['h4'] = self._increase_font_factor(-4)
self._exit['h5'] = self._increase_font_factor(-3)
self._exit['title'] = self._increase_font_factor(-7)
# never go in and parse these tags
self._ignored_tags = set([
'meta', 'script', 'link', 'meta', 'embed', 'iframe', 'frame',
'noscript', 'object', 'svg', 'canvas', 'applet', 'frameset',
'textarea', 'style', 'area', 'map', 'base', 'basefont', 'param',
])
# set of words to ignore
self._ignored_words = set([
'', 'the', 'of', 'at', 'on', 'in', 'is', 'it',
'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j',
'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't',
'u', 'v', 'w', 'x', 'y', 'z', 'and', 'or',
])
# keep track of some info about the page we are currently parsing
self._curr_depth = 0
self._curr_url = ""
self._font_size = 0
# get all urls into the queue
try:
with open(url_file, 'r') as f:
for line in f:
self._url_queue.append((self._fix_url(line.strip(), ""), 0))
except IOError:
pass
self.db = db_lib.sql3({'clean': True, 'location':'test.db'})
incoming_link_sets = defaultdict(set)
incoming_links = defaultdict(lambda: np.array([]))
damping_factor = 0.85
# collect the number of outbound links and the set of all incoming documents
# for every document
for (from_id,to_id) in links:
num_outgoing_links[int(from_id)] += 1.0
incoming_link_sets[to_id].add(int(from_id))
# convert each set of incoming links into a numpy array
for doc_id in incoming_link_sets:
incoming_links[doc_id] = np.array([from_doc_id for from_doc_id in incoming_link_sets[doc_id]])
num_documents = float(len(num_outgoing_links))
lead = (1.0 - damping_factor) / num_documents
partial_PR = np.vectorize(lambda doc_id: page_rank[doc_id] / num_outgoing_links[doc_id])
for _ in xrange(num_iterations):
for doc_id in num_outgoing_links:
tail = 0.0
if len(incoming_links[doc_id]):
tail = damping_factor * partial_PR(incoming_links[doc_id]).sum()
page_rank[doc_id] = lead + tail
return page_rank
if __name__ == "__main__":
db = db_lib.sql3()
ranks = page_rank(db.)
