def create_info_page_task(username, elaboration_id, url):
elaboration = Elaboration(elaboration_id=elaboration_id)
redis = RedisCache()
twitter = get_twitter_from_username(username)
info = twitter.show_user(screen_name=username)
user = User.objects.get(username=username)
###########################################################################
# PERSONAL INFORMATION
###########################################################################
try:
user_info = UserInfo.objects.get(user=user)
except UserInfo.DoesNotExist:
user_info = UserInfo()
user_info.user = user
user_info.full_name = info['name']
user_info.nick = info['screen_name']
user_info.bio = info['description']
user_info.image = info['profile_image_url'].replace('_normal', '')
user_info.save()
###########################################################################
# CLUSTER & TWEETS
###########################################################################
k = 50
if NUMBER_OF_TWEETS <= 20: # debug
k = 10
redis.set('{}:step'.format(elaboration_id), 1)
# clusterify = KMeansClusterify(TwitterProfileReader(username), k)
clusterify = AffinityPropagationClusterify(TwitterProfileReader(username), k)
# clusterify = SpectralClusterify(TwitterProfileReader(username), k)
# clusterify = StarClusterify(TwitterProfileReader(username))
redis.set('{}:step'.format(elaboration_id), 2)
elaboration.tweets = clusterify.annotate()
redis.set('{}:step'.format(elaboration_id), 3)
elaboration.cluster = clusterify.do_cluster()
logger.info(elaboration.cluster)
redis.set('{}:step'.format(elaboration_id), 4) # exit code
elaboration.user = user
elaboration.save()
subject, to = "Work done!", user.email
text_content = "Checkout the final elaboration here: {}".format(url)
html_content = '<p>Checkout the final elaboration <a href="{}">' \
'here</a></p>'.format(url)
msg = EmailMultiAlternatives(subject, text_content, EMAIL_HOST_USER, [to])
msg.attach_alternative(html_content, "text/html")
msg.send()
def start_elaboration(request):
user = request.user
redis = RedisCache()
elaboration_id = sha1(user.username + str(datetime.now())).hexdigest()
redis.set('{}:step'.format(elaboration_id), 0)
create_info_page_task.delay(
user.username, elaboration_id, request.build_absolute_uri(
reverse('show_info_page', args=[elaboration_id])
)
)
return HttpResponseRedirect(
reverse('elaboration', args=[elaboration_id])
)
class InterWikiRecon(object):
def __init__(self):
self.cache = RedisCache()
self.requests = requests.session()
def get_inter_wikilinks(self, page):
"""
>>> a = InterWikiRecon()
>>> a.get_inter_wikilinks('http://it.wikipedia.org/wiki/Mozilla')
{u'EN': u'http://en.wikipedia.org/wiki/Mozilla', \
u'IT': u'http://it.wikipedia.org/wiki/Mozilla'}
>>> a.get_inter_wikilinks('http://en.wikipedia.org/wiki/Mozilla')
{u'EN': u'http://en.wikipedia.org/wiki/Mozilla', \
u'IT': u'http://it.wikipedia.org/wiki/Mozilla'}
"""
key = "{}:interwikirecon".format(page)
if not self.cache.has(key):
given_lang = 'IT' if '://it.' in page else 'EN'
new_lang = 'IT' if given_lang == 'EN' else 'EN'
url = 'http://interwikirecon.spaziodati.eu/reconcile?' \
'queries={"q0":{"query": "%s", "type": "Wikipedia %s"}}' \
'' % (page, new_lang)
try:
new_page = self.requests.get(url).json()['q0']['result'][0]['id']
except IndexError:
new_page = None
self.cache.set(key, {
given_lang: page,
new_lang: new_page,
})
return self.cache.get(key)
class DataTXT(object):
def __init__(self):
self.interWikiRecon = InterWikiRecon()
self.cache = RedisCache()
self.requests = requests.session()
self.datatxt = datatxt.DataTXT(
app_id=DATATXT_APP_ID,
app_key=DATATXT_APP_KEY,
)
def nex(self, *args):
try:
annotated = self.datatxt.nex(
args[0],
min_confidence=.6,
parse_hashtag=True,
epsilon=.5,
)
id_ = sha1(annotated.lang + str(annotated.annotations)).hexdigest()
return {
'id': id_,
'lang': annotated.lang,
'annotations': {
a.uri: a for a in annotated.annotations
}
}
except DandelionException:
logger.info("dandelion exception: " + args[0])
def _rel_request(self, lang, topic1, topic2):
if all(x is None for x in topic1) or all(x is None for x in topic2):
return {}
url = 'http://api.dandelion.eu/datatxt/rel/v1'
response = self.requests.get(url, params={
'lang': lang,
'topic1': topic1,
'topic2': topic2,
'$app_id': DATATXT_APP_ID,
'$app_key': DATATXT_APP_KEY,
})
if response.ok:
return response.json()
print response.json()
return {}
def rel(self, topics1, topics2, enable_cache=True):
rel = zeros((len(topics1), len(topics2)))
response = None
for i, topic1 in enumerate(topics1):
for j, topic2 in enumerate(topics2):
cache_key = "{}-{}:relatedness".format(topic1, topic2)
if enable_cache and self.cache.has(cache_key):
value = self.cache.get(cache_key)
else:
if response is None:
response = self._rel(topics1, topics2)
value = response[i][j]
self.cache.set(cache_key, value)
rel[i][j] = value
return rel
def _rel(self, topics1, topics2):
def lang_topic(topics, wanted_lang):
for topic in topics:
lang = 'it' if '://it.' in topic else 'en'
if lang == wanted_lang:
yield topic
else:
pages = self.interWikiRecon.get_inter_wikilinks(topic)
yield pages.get(wanted_lang.upper())
it_topics1 = list(lang_topic(topics1, 'it'))
it_topics2 = list(lang_topic(topics2, 'it'))
en_topics1 = list(lang_topic(topics1, 'en'))
en_topics2 = list(lang_topic(topics2, 'en'))
it_response = {
tuple(sorted([x['topic1']['topic']['uri'], x['topic2']['topic']['uri']])): x['weight']
for x in self._rel_request('it', it_topics1, it_topics2).get('relatedness', [])
if not x.get('error', False)
}
en_response = {
tuple(sorted([x['topic1']['topic']['uri'], x['topic2']['topic']['uri']])): x['weight']
for x in self._rel_request('en', en_topics1, en_topics2).get('relatedness', [])
if not x.get('error', False)
}
response = zeros((len(topics1), len(topics2)))
for i, (it_topic1, en_topic1) in enumerate(zip(it_topics1, en_topics1)):
for j, (it_topic2, en_topic2) in enumerate(zip(it_topics2, en_topics2)):
value_it = it_response.get(
tuple(sorted([it_topic1, it_topic2])), 0.
)
value_en = en_response.get(
tuple(sorted([en_topic1, en_topic2])), 0.
)
response[i][j] = max(value_it, value_en)
return response