def test_generate_html(self):
index = {
'http://www.davidcmoss.co.uk/':
Page(title=['David C Moss'],
links=set([
u'http://www.davidcmoss.co.uk/static/Curriculum Vitae.pdf'
]),
ex_links=set(['https://www.heroku.com']),
images=['/static/img/profile.jpeg']),
'http://www.davidcmoss.co.uk/static/Curriculum%20Vitae.pdf':
Page(title=[], links=set([]), ex_links=set([]), images=[])
}
html = generate_html(index)
assert "https://www.heroku.com" in html
def test_crawl_site(self):
index = crawl_site('http://www.davidcmoss.co.uk')
expected_response = {
u'http://www.davidcmoss.co.uk/':
Page(title=u'David C Moss',
links=set([
u'http://www.davidcmoss.co.uk/static/Curriculum Vitae.pdf'
]),
ex_links=['https://www.heroku.com'],
images=[
u'http://www.davidcmoss.co.uk/static/img/profile.jpeg'
]),
u'http://www.davidcmoss.co.uk/static/Curriculum Vitae.pdf':
Page(title=[], links=set([]), ex_links=[], images=[])
}
assert index == expected_response
def build_crawl_list(self):
"""
Build a list of all of the URLs based on the depth specified.
"""
current_depth = 1
page = requests.get(self.base_url).text
if self.children > 0:
self.urls = Page.get_urls(page)[:self.children]
else:
self.urls = Page.get_urls(page)
# Below list holds previously scanned URLs, to stop URLs being added twice
scanned_urls = []
while current_depth <= self.depth:
# Append the links for each page then search it for more
print 'Starting crawl depth', current_depth, 'with', len(self.urls), 'URLs to scan'
new_urls = []
for url in self.urls:
# If the url is not already scanned, and if it is not an image, xml etc. scan it.
if url not in scanned_urls:
if TasteDotCom.is_wanted_object(url):
print 'Looking for child URLs in ', url
markup = requests.get(url).text
scanned_urls.append(url)
if self.children > 0:
new_urls = Page.get_urls(markup)[:self.children]
else:
new_urls = Page.get_urls(markup)
print 'Found', len(new_urls), 'new pages'
# for url in new_urls:
# check_and_add(url)
self.urls += new_urls
current_depth += 1
print 'Finished crawling', self.base_url, 'found', len(self.urls), 'total URLs'
# def run(self):
# """
# Start Crawling the page specified
# """
# #todo Make use of this method
# print "Starting crawl session for", self.base_url
# page = requests.get(self.base_url).text
# child_urls = Page.get_urls(page)
# for url in child_urls:
# self.check_and_add(url)
# def check_and_add(url):
# pass
def test_get_urls(self):
# Use a known static page for testing
html = file("../miscellany/sausage_and_punpkin_curry.html").read()
urls = Page.get_urls(html)
self.assertEquals(len(urls), 387)
a = re.compile(r"http://www\.[//\a\w\.\+]+")
for url in urls:
# Check each URL matches a hyperlink pattern
self.assertTrue(a.match(url))
import network2 as nx
import pylab as plt
from crawler import Crawler, Page, Document, Corpus
if __name__ == '__main__':
start_page = Page('http://info.cern.ch/hypertext/WWW/TheProject.html')
crawler = Crawler(start_page)
crawler.crawl()
web_graph = nx.DiGraph()
edges = []
edges2 = []
for page in crawler.web:
for link in page.links:
edges.append((hash(page.address), hash(link)))
web_graph.add_edges_from(edges)
nx.draw(web_graph)
plt.show()
pageRanks = nx.pagerank(web_graph)
for page in crawler.web:
page.page_rank = pageRanks[hash(page.address)]
pages = sorted(crawler.web, key=lambda p: p.page_rank, reverse=True)
corpus = []
for page in pages:
corpus.append((page.address, page.text))
