def classify_url(self, domain, page, depth=0):
"""
Classify the documents after crawling them.
args:
domain - the domain part of the url
page - the other part of the url
depth - how deep to crawl
returns:
a list of predicted probabilities for each instance belonging to
each class
"""
# get the documents
documents, _ = crawl_page(domain, page, depth=0)
# parse the documents
string_data = []
for page, doc in documents.iteritems():
words = parse_html_simple(doc)
parsed = []
for word in words:
if (word in self.english_words
and word not in self.stop_words
and word in self.vocabulary):
parsed.append(word)
string_data.append(' '.join(parsed))
count_data = self.vectorizer.transform(string_data)
# classify the documents
probs = self.classifier.predict_proba(count_data)
return probs
def call_scraper(args):
domain = args[0]
link = args[1]
return crawl_page(domain, link, href_match='/wiki/', depth=1)
