def cluster(url):
"""
Read URLs from sitemaps and return clusters
url is either a website (and we detect sitemaps) or a sitemap
"""
data = {}
if url[:4] != "http":
url = "http://" + url
if re.search(r"https?://[^/?#]+[/?#].+", url):
sitemaps = [url] # sitemap URL given
else:
robots = url.strip("/") + "/robots.txt"
sitemaps = sitemaps_from_robots(robots)
if not sitemaps:
# assume sitemap.xml
sitemaps = [url.strip("/") + "/sitemap.xml"]
if sitemaps:
try:
urls = read_sitemaps(sitemaps)
if not urls:
data["error"] = "No URLs found in sitemap"
else:
data["count"] = len(urls)
urls = [x.strip() for x in urls]
# cluster URLs
c = urlclustering.cluster(urls)
tmp = deepcopy(c["clusters"])
try:
improve_patterns(c["clusters"])
except:
c["clusters"] = tmp
pass
# prepare HTML
html = "<pre>CLUSTERS:"
keys = sorted(c["clusters"], key=lambda k: len(c["clusters"][k]), reverse=True)
for key in keys:
urls = c["clusters"][key]
html += "\n" + key[1] + " [%s URLs]<br/>" % len(urls)
html += "\t" + "\n\t".join(urls[:5])
html += "\n\t...%s more" % (len(urls) - 5)
html += "\n\nUNCLUSTERED:\n"
html += "\t" + "\n\t".join(c["unclustered"])
html += "</pre>"
data["html"] = html
except:
logging.debug(traceback.format_exc())
data["error"] = "An error happened while fetching sitemaps"
else:
data["error"] = "No sitemaps found"
return json.dumps(data)
def test_other(self):
x_urls = ['http://s.com/blah/%d' % x for x in range(1, 20)]
y_urls = ['http://s.com/a/b/aa%dbb' % x for x in range(1, 20)]
z_urls = ['http://b.com/ab/aa%dbb' % x for x in range(1, 50)]
c = cluster_urls(x_urls + y_urls + z_urls, 10)
improve_patterns(c['clusters'])
self.assertEqual(c['unclustered'], [])
self.assertEqual(
sorted(c['clusters'].keys()),
sorted([
('http://b.com/ab/aa([^/]+)bb', 'http://b.com/ab/aa[...]bb'),
('http://s.com/blah/(\\d+)', 'http://s.com/blah/[NUMBER]'),
('http://s.com/a/b/aa([^/]+)bb', 'http://s.com/a/b/aa[...]bb')
]))
def test_other(self):
x_urls = ['http://s.com/blah/%d' % x for x in range(1, 20)]
y_urls = ['http://s.com/a/b/aa%dbb' % x for x in range(1, 20)]
z_urls = ['http://b.com/ab/aa%dbb' % x for x in range(1, 50)]
c = cluster_urls(x_urls + y_urls + z_urls, 10)
improve_patterns(c['clusters'])
self.assertEqual(c['unclustered'], [])
self.assertEqual(
sorted(c['clusters'].keys()),
sorted([('http://b.com/ab/aa([^/]+)bb',
'http://b.com/ab/aa[...]bb'),
('http://s.com/blah/(\\d+)',
'http://s.com/blah/[NUMBER]'),
('http://s.com/a/b/aa([^/]+)bb',
'http://s.com/a/b/aa[...]bb')]))
