def test_is_new(self, mc):
mc.return_value = mongomock.MongoClient()
sm = SeenManager("test", "host", 0, "db")
dmeta = DocumentMetadata("http://www.google.com")
dmeta.alternatives = [
"http://www.google.com",
"http://www.google2.com/",
"https://www.google3.com",
]
dmeta.dhash = 2413242
other_urls = [
"www.test.com",
"www.other.com",
]
# adding urls
sm.add(dmeta)
for u in dmeta.alternatives:
self.assertFalse(sm.is_new(canonize(u)))
for u in other_urls:
self.assertTrue(sm.is_new(canonize(u)))
def test_remove_seen(self, mc):
'''
Test delete some seen
'''
mc.return_value = mongomock.MongoClient()
qm = QueueManager("queues-names", START_DELAY, CONFIGURATION)
############################
# insert some urls into seen
dm_seen_1 = DocumentMetadata("http://www.randomurl1.it")
# this two are cosidered the same page with different urls
dm_seen_1.alternatives = [
"http://www.randomurl1.it",
"http://www.randomurl3.it",
]
dm_seen_1.dhash = 12345
dm_seen_2 = DocumentMetadata("http://www.randomurl4.it")
dm_seen_2.alternatives = [
"http://www.randomurl4.it",
]
dm_seen_2.dhash = 98765
already_seen_urls = set(dm_seen_1.alternatives).union(
dm_seen_2.alternatives)
qm.seen.add(dm_seen_1)
qm.seen.add(dm_seen_2)
self.assertTrue(qm.seen.get("www.randomurl1.it") is not None)
self.assertTrue(qm.seen.get("www.randomurl3.it") is not None)
self.assertTrue(qm.seen.get("www.randomurl4.it") is not None)
# deleting url should remove all alternatives
qm.remove_seen("www.randomurl1.it")
self.assertTrue(qm.seen.get("www.randomurl1.it") is None)
self.assertTrue(qm.seen.get("www.randomurl3.it") is None)
self.assertTrue(qm.seen.get("www.randomurl4.it") is not None)
def test_update(self, mc):
mc.return_value = mongomock.MongoClient()
sm = SeenManager("test", "host", 0, "db")
dmeta = DocumentMetadata("http://www.google.com?q=test")
dmeta.alternatives = [
"http://www.google.com?q=test",
"http://www.google2.com/",
"https://www.google3.com",
]
dmeta.dhash = 2413242
dmeta2 = DocumentMetadata("http://www.google2.com")
dmeta2.alternatives = [
"http://www.google2.com",
"https://www.google3.com",
]
dmeta2.dhash = 12121212
# adding urls
sm.add(dmeta)
sm.add(dmeta2)
output_alternatives = dmeta.alternatives + dmeta2.alternatives
output_alternatives = list(
set(canonize(i) for i in output_alternatives))
# checking presence and checking not double anonization
for i in output_alternatives:
self.assertTrue(i in sm.store)
for i, v in enumerate(sm.store.get(i)['alternatives']):
self.assertEqual(v, output_alternatives[i])
def setUp(self, mc):
super(TestNewsQueueRescheduling, self).setUp()
mc.return_value = mongomock.MongoClient()
self.qm = QueueManager("queues-names", START_DELAY, CONFIGURATION_NEWS)
############################
# insert some urls into seen
dm_seen_1 = DocumentMetadata("http://www.randomurl1.it")
# this two are cosidered the same page with different urls
dm_seen_1.alternatives = [
"http://www.randomurl1.it",
"http://www.randomurl3.it",
]
dm_seen_1.dhash = 12345
dm_seen_2 = DocumentMetadata("http://www.randomurl4.it")
dm_seen_2.alternatives = [
"http://www.randomurl4.it",
]
dm_seen_2.dhash = 98765
already_seen_urls = set(dm_seen_1.alternatives).union(
dm_seen_2.alternatives)
self.qm.seen.add(dm_seen_1)
self.qm.seen.add(dm_seen_2)
self.qm.seen.incr_n(
dm_seen_1.url) # increase counter to check it later
counter = self.qm.seen.get(dm_seen_1.url).get("count")
self.assertEqual(counter, 2)
def test_add_and_delete(self, mc):
mc.return_value = mongomock.MongoClient()
sm = SeenManager("test", "host", 0, "db")
dmeta = DocumentMetadata("http://www.google.com")
dmeta.alternatives = [
"http://www.google.com",
"http://www.google2.com/",
"https://www.google3.com",
]
dmeta.dhash = 2413242
other_urls = [
"www.prova.com",
"www.other.com",
]
# adding urls
sm.add(dmeta)
# tring removing not present urls
for o in other_urls:
sm.delete(o)
# checking presence
for i in dmeta.alternatives:
self.assertTrue(canonize(i) in sm.store)
self.assertEqual(len(dmeta.alternatives), len(sm.store))
# checking not precence
for u in other_urls:
self.assertFalse(canonize(u) in sm.store)
# checking correctness
for i in dmeta.alternatives:
data = sm.store.get(canonize(i))
self.assertEqual(data["count"], 1)
self.assertEqual(data["page_hash"], dmeta.dhash)
# deleting alternatives
sm.delete(dmeta.alternatives[0])
# checking empty db
for i in dmeta.alternatives:
self.assertFalse(canonize(i) in sm.store)
self.assertEqual(0, len(sm.store))
def test_incr_n(self, mc):
mc.return_value = mongomock.MongoClient()
sm = SeenManager("test", "host", 0, "db")
dmeta = DocumentMetadata("http://www.google.com")
dmeta.alternatives = [
"http://www.google.com",
"http://www.google2.com/",
"https://www.google3.com",
]
dmeta.dhash = 2413242
# adding urls
sm.add(dmeta)
# increase counters
sm.incr_n(dmeta.alternatives[0])
for i in dmeta.alternatives:
data = sm.store.get(canonize(i))
self.assertEqual(data["count"], 2)
def test_is_changed(self, mc):
mc.return_value = mongomock.MongoClient()
sm = SeenManager("test", "host", 0, "db")
dmeta = DocumentMetadata("http://www.google.com")
dmeta.alternatives = [
"http://www.google.com",
"http://www.google2.com/",
"https://www.google3.com",
]
dmeta.dhash = 2413242
# adding urls
sm.add(dmeta)
for u in dmeta.alternatives:
self.assertFalse(sm.is_changed(u, dmeta.dhash))
for u in dmeta.alternatives:
self.assertFalse(sm.is_changed(u, dmeta.dhash + 2))
for u in dmeta.alternatives:
self.assertTrue(sm.is_changed(u, dmeta.dhash + 3))
def test_add_normal_urls(self, mc):
'''testing insertion of new links'''
mc.return_value = mongomock.MongoClient()
qm = QueueManager("queues-names", START_DELAY, CONFIGURATION)
dm = DocumentMetadata("http://random-url.com")
dm.depth = 1
dm.links = [
"http://www.randomurl1.it",
"http://www.randomurl2.it",
"http://www.randomurl3.it",
"http://www.randomurl4.it",
"http://www.randomurl5.it",
"http://www.randomurl6.it",
"http://www.randomurl7.it",
]
all_url_lenght = len(dm.links)
# adding all the links found in the document to the normal list
qm.add_normal_urls(dm)
# checking if the urls are there:
stored = qm.normal_store.getall()
self.assertEqual(len(stored), all_url_lenght)
links_set = set(dm.links)
for s in stored:
self.assertTrue(s[1] in links_set)
# depth should be increased by 1
self.assertEqual(s[0], dm.depth + 1)
# checking that seen is still empty
for u in dm.links:
self.assertEqual(qm.seen.is_new(u), True)
# adding a duplicate with same depth
# entry should be replaced.
dm.links = ["http://www.randomurl1.it"]
qm.add_normal_urls(dm)
stored = qm.normal_store.getall()
self.assertEqual(len(stored), all_url_lenght)
# adding a duplicate with different depth
# we should have duplicate entries.
dm.depth = 3
dm.links = ["http://www.randomurl1.it"]
qm.add_normal_urls(dm)
stored = qm.normal_store.getall()
self.assertEqual(len(stored), all_url_lenght + 1)
def test_pop_ordering(self, mc):
'''
Test adding url to priority, normal and refetch
and checking the ordering of popping is correct
'''
mc.return_value = mongomock.MongoClient()
qm = QueueManager("queues-names", START_DELAY, CONFIGURATION)
# inserting a priority url.
urls = [
"www.daniele.com",
]
qm.init_priority_list(urls)
# inserting a normal url.
burls = [
{
"url": "www.daniele1.com",
"depth": "2"
},
]
qm.add_bootstrap_urls(burls)
# inserting a refetch url
dm = DocumentMetadata("http://www.randomurl8.it")
dm.depth = 1
dm.dhash = 121212
dm.source = Source.normal
dm.delay = 500
dm.alternatives = ["http://www.randomurl8.it"]
qm.add_seen_and_reschedule(dm)
# make sure all the inserted url are ready to be popped
with mock.patch("time.time", mock_time):
doc = qm.pop()
# first one from priority
self.assertEqual(doc.depth, 0)
self.assertEqual(doc.source, Source.priority)
doc = qm.pop()
# second one from normal
self.assertEqual(doc.source, Source.normal)
doc = qm.pop()
# third from refetching
self.assertEqual(doc.source, Source.refetch)
def test_fetcher(self, session):
for i in self.input_data():
url, content, status = i.split("\t")
status = int(status)
dm = DocumentMetadata(url)
mr = MockResponse(content, status)
session.return_value = mr
new_dm = fetcher.fetch({}, dm, 1)
self.assertEqual(dm.url, url)
if status == 200:
self.assertEqual(dm.status, 0)
self.assertEqual(new_dm.status, 0)
self.assertEqual(new_dm.response.content, content)
elif status == 301:
self.assertEqual(dm.status, fetcher.Status.SkipUrl)
elif status == 0:
self.assertEqual(dm.status, fetcher.Status.ConnectionError)
elif status >= 500 and status < 510:
self.assertEqual(dm.status, fetcher.Status.SkipUrl)
else:
self.assertEqual(dm.status, fetcher.Status.GenericError)
def test_reschedule_priority(self):
############################################
# inserting a priority url
dm = DocumentMetadata("http://www.randomurl8.it")
dm.depth = 1
dm.dhash = 121212
dm.source = Source.priority
dm.delay = 500
dm.alternatives = ["http://www.randomurl8.it"]
self.qm.add_seen_and_reschedule(dm)
# checking all the parameters
counter = self.qm.seen.get(dm.url).get("count")
dhash = self.qm.seen.get(dm.url).get("page_hash")
self.assertEqual(counter, 1)
self.assertEqual(dhash, dm.dhash)
with mock.patch("time.time", mock_time):
refetching_data = self.qm.pop()
self.assertEqual(refetching_data.delay, START_DELAY)
self.assertEqual(refetching_data.source, Source.priority)
def test_reschedule_different_content(self):
############################################
# this one is in seen with a different hash and
# not taken from priority queue
# I expect: halving the delay and set seen counter to 1
dm = DocumentMetadata("http://www.randomurl1.it")
dm.depth = 1
dm.dhash = 1936
dm.source = Source.normal
dm.delay = 20
dm.alternatives = ["http://www.randomurl1.it"]
self.qm.add_seen_and_reschedule(dm)
# checking all the parameters
counter = self.qm.seen.get(dm.url).get("count")
dhash = self.qm.seen.get(dm.url).get("page_hash")
self.assertEqual(counter, 1)
self.assertEqual(dhash, dm.dhash)
with mock.patch("time.time", mock_time):
refetching_data = self.qm.pop()
self.assertEqual(refetching_data.delay, dm.delay / 2)
self.assertEqual(refetching_data.source, Source.refetch)
def test_reschedule4(self):
############################################
# inserting a new url
dm = DocumentMetadata("http://www.randomurl8.it")
dm.depth = 1
dm.dhash = 121212
dm.source = Source.normal
dm.delay = 500
dm.alternatives = ["http://www.randomurl8.it"]
self.qm.add_seen_and_reschedule(dm)
# checking all the parameters
counter = self.qm.seen.get(dm.url).get("count")
dhash = self.qm.seen.get(dm.url).get("page_hash")
self.assertEqual(counter, 1)
self.assertEqual(dhash, dm.dhash)
with mock.patch("time.time", mock_time):
refetching_data = self.qm.pop()
self.assertEqual(refetching_data.delay,
CONFIGURATION["queues"]["refetching-delay"])
self.assertEqual(refetching_data.source, Source.refetch)
def test_reschedule3(self):
############################################
# as before but with a small delay.
# cheking delay not changing
dm = DocumentMetadata("http://www.randomurl1.it")
dm.depth = 1
dm.dhash = 121212
dm.source = Source.normal
dm.delay = 500
dm.alternatives = ["http://www.randomurl1.it"]
self.qm.add_seen_and_reschedule(dm)
# checking all the parameters
counter = self.qm.seen.get(dm.url).get("count")
dhash = self.qm.seen.get(dm.url).get("page_hash")
self.assertEqual(counter, 1)
self.assertEqual(dhash, dm.dhash)
with mock.patch("time.time", mock_time):
refetching_data = self.qm.pop()
self.assertEqual(refetching_data.delay, dm.delay)
self.assertEqual(refetching_data.source, Source.refetch)
def test_reschedule_samecontent_lastdelay(self):
############################################
# testing rescheduling some urls
# this one is in seen with the same hash and
# not taken from priority queue
# I expect: doublig the delay and set seen counter to 1
dm = DocumentMetadata("http://www.randomurl1.it")
dm.depth = 1
dm.dhash = 12345
dm.source = Source.normal
dm.delay = 40
# alternatives contains always at least one url.
dm.alternatives = ["http://www.randomurl1.it"]
# we want to check that former alternatives are also correctly
# updated even if new alternatives field is different.
alternatives = self.qm.seen.get(dm.url).get("alternatives")
self.assertNotEqual(len(dm.alternatives), len(alternatives))
self.qm.add_seen_and_reschedule(dm)
# check all the parameters
counter = self.qm.seen.get(dm.url).get("count")
dhash = self.qm.seen.get(dm.url).get("page_hash")
self.assertEqual(counter, 1)
# check updated all the alternatives
for urls in alternatives:
counter = self.qm.seen.get(urls).get("count")
dhash = self.qm.seen.get(dm.url).get("page_hash")
self.assertEqual(counter, 1)
self.assertEqual(dhash, dm.dhash)
with mock.patch("time.time", mock_time):
refetching_data = self.qm.pop()
self.assertEqual(refetching_data.url, "")
self.assertEqual(refetching_data.source, Source.unknown)
def pop(self):
"""Return the next document to fetch"""
document_metadata = DocumentMetadata()
item = self.priority_store.pop(int(time.time()))
if item:
logging.debug("Get priority:" + str(item[1]))
document_metadata.url = item[1]
document_metadata.depth = item[3]
document_metadata.delay = item[2]
document_metadata.source = Source.priority
else:
while not item:
item = self.normal_store.pop()
if not item:
break
# the following check is needed because urls are stored in seen
# after seeing them
# so we can have multiple identical url in normal list.
# and we do not want to have multiple same urls in refetching list
if not self.seen.is_new(item[1]):
item = None
if item:
# In case of network error I repush url on normal queue
# just to not loose them. So it is possible we have
# something already seen here.
# It is not a problem to refetch this cases
logging.debug("Get normal:" + str(item[1]))
document_metadata.url = item[1]
document_metadata.depth = item[0]
document_metadata.delay = 0
document_metadata.source = Source.normal
else:
item = self.refetch_store.pop(int(time.time()))
if item:
logging.debug("Get Refetch:" + str(item[1]))
document_metadata.url = item[1]
document_metadata.depth = item[3]
document_metadata.delay = item[2]
document_metadata.source = Source.refetch
return document_metadata
def test_add_normal_urls_some_seen(self, mc):
'''
Tesiting insertion of some new and old links.
'''
mc.return_value = mongomock.MongoClient()
qm = QueueManager("queues-names", START_DELAY, CONFIGURATION)
############################
# insert some urls into seen
dm_seen_1 = DocumentMetadata("http://www.randomurl1.it")
# this two are cosidered the same page with different urls
dm_seen_1.alternatives = [
"http://www.randomurl1.it",
"http://www.randomurl3.it",
]
dm_seen_1.dhash = 12345
dm_seen_2 = DocumentMetadata("http://www.randomurl4.it")
dm_seen_2.alternatives = [
"http://www.randomurl4.it",
]
dm_seen_2.dhash = 98765
already_seen_urls = set(dm_seen_1.alternatives).union(
dm_seen_2.alternatives)
qm.seen.add(dm_seen_1)
qm.seen.add(dm_seen_2)
############################################
# testing add normal url with some seen urls
dm = DocumentMetadata("http://random-url.com")
dm.depth = 1
dm.links = [
"http://www.randomurl1.it",
"http://www.randomurl2.it",
"http://www.randomurl3.it",
"http://www.randomurl4.it",
"http://www.randomurl5.it",
"http://www.randomurl6.it",
"http://www.randomurl7.it",
]
# adding all the links found in the document to the normal list
qm.add_normal_urls(dm)
# checking if the urls are there (all except 3 because already seen):
links_set = set(dm.links).difference(already_seen_urls)
stored = qm.normal_store.getall()
self.assertEqual(len(stored), len(links_set))
# count for this urls should be 1
for i in dm_seen_1.alternatives:
data = qm.seen.get(i)
# this is 3 becouse we inserted one in seen +
# we tried to insert in normal www.randomurl1.com +
# we tried to insert in normal www.randomurl3.com
self.assertEqual(data["count"], 3)
# adding a duplicate that is already in seen.
# should not be added in normal list, but counters of
# all alternatives shouls be updated (+1)
dm.links = [
dm.links[0],
]
qm.add_normal_urls(dm)
stored = qm.normal_store.getall()
for i in dm_seen_1.alternatives:
data = qm.seen.get(i)
self.assertEqual(data["count"], 4)
# adding again with different depth should not
# change the behaviour
dm.depth = 3
dm.links = [
dm.links[0],
]
qm.add_normal_urls(dm)
for i in dm_seen_1.alternatives:
# should be 2 because I inserted 2 times
data = qm.seen.get(i)
self.assertEqual(data["count"], 5)
