def getClustersWithCat():
feedback = ''
str = ''
try:
category = request.args.get('category')
str += 'got parameter category: %s\n' % category
clusters = Cluster.query(Cluster.category == category).fetch()
str += 'fetched the clusters'
str = ''
i = 1
for c in clusters:
str += 'cluster %d:\n' % i
str += 'category: %s\n' % c.category
for np in c.listNews:
str += '\t%s\n' % np.title
str += '\n'
i+= 1
except Exception as inst:
feedback += 'Exception type: %s\n' % type(inst)
feedback += 'Exception: %s\n' % inst.message
str += feedback
return Response(str, mimetype='text/plain')
def getClustersServer():
category = request.args.get('category')
str = ''
clusters = Cluster.query(Cluster.category == category).fetch()
str += 'Got category %s\n' % category
if clusters is None:
str = 'None clusters :/'
else:
i = 1
str += 'Number of clusters %d\n' % len(clusters)
for c in clusters:
str += 'Cluster %d\n' % i
for np in c.listNews:
str += '\t%s\t%s\n' % (np.title, np.source_url)
i += 1
return Response(str, mimetype='text/plain')
def prepareClustering():
logging.debug('preparing for refreshing..')
ndb.delete_multi(Cluster.query().fetch(keys_only=True))
getClusters()
logging.debug('refreshed!')
return ''
def getMYClusters():
category = request.args.get('category')
clusters = Cluster.query(Cluster.category == category).fetch()
obj = {'listClusters' : [c.serialize() for c in clusters]}
result = json.dumps(obj, ensure_ascii=True)
return Response(result, mimetype='text/plain')
def getFilteredClustersDebug():
reqParam = request.args.get('wantedSources')
listWanted = []
#vrednosta na parametarot wantedSources kje bide od oblik id1,id2
if None != reqParam:
listWanted = reqParam.split(',')
# da se pretvorat vo integers
for i in range(0, len(listWanted)):
listWanted[i] = int(listWanted[i])
result = 'Wanted sources: %s\n' % listWanted
category = request.args.get('category')
clusters = Cluster.query(Cluster.category == category).fetch()
#tuka gi smestuvame samo klasterite shto se so filtrirani vesti
# (ne zemame vesti od izvori sto korisnikot ne gi saka)
clustersToReturn = []
for c in clusters:
listNews = []
for np in c.listNews:
if np.source_id in listWanted:
listNews.append(np)
#dokolku klasterot nema nitu edna vest posle filtriranjeto
#ne go vkluchuvame bidejkji e prazen :)
if len(listNews) > 0:
c.listNews = listNews
clustersToReturn.append(c)
i = 1
#debugging
result += 'Number of clusters %d\n' % len(clusters)
for c in clustersToReturn:
result += 'Cluster %d\n' % i
for np in c.listNews:
result += '\t%s\t%s\n' % (np.title, np.source_url)
i += 1
return Response(result, mimetype='text/plain')
def get_clusters(self):
clusters = []
with open(self.filename) as file:
data = json.load(file)
if data["clusters"]:
clusters_json = data["clusters"]
for cluster_json in clusters_json:
cluster = Cluster()
cluster.set_dimensions(self.get_dimensions())
if cluster_json["type"]:
type_found = False
for distribution in Distribution:
if distribution.name == cluster_json["type"]:
cluster.set_distribution(Distribution[distribution.name])
type_found = True
if not type_found:
print("Couldn't map distribution type. Defaulting to 'gaussian'")
else:
print("Couldn't find 'type' element for cluster")
if cluster_json["cardinality"]:
cardinality = int(cluster_json["cardinality"])
if 0 < cardinality <= 10:
cluster.set_cardinality(cardinality)
else:
print("Invalid cardinality. Has to be within 1 to 10. Defaulting to 5.")
else:
print("Couldn't find 'cardinality' element for cluster")
if cluster_json["density"]:
density = int(cluster_json["density"])
if 0 < density <= 10:
cluster.set_density(density)
else:
print("Invalid density. Has to be within 1 to 10. Defaulting to 5.")
else:
print("Couldn't find 'density' element for cluster")
clusters.append(cluster)
else:
print("Couldn't find 'clusters' element in input file")
return clusters
def getMYNews():
category = request.args.get('category')
news = NewsPost.query().fetch()
clusters = Cluster.query(Cluster.category == category).fetch()
newNews = []
for n in news:
newObject = NewsPostClient(url = n.url,host_page = n.host_page,title = n.title, description = n.description)
newNews.append(newObject)
#result = str(byteify(newNews[0].serialize()))
result = ''
return Response(result, mimetype='application/javascript')
def getFilteredClusters():
reqParam = request.args.get('wantedSources')
listWanted = []
if None != reqParam:
listWanted = reqParam.split(',')
for i in range(0, len(listWanted)):
listWanted[i] = int(listWanted[i])
category = request.args.get('category')
clusters = Cluster.query(Cluster.category == category).fetch()
clustersToReturn = []
for c in clusters:
listNews = []
for np in c.listNews:
if np.source_id in listWanted:
listNews.append(np)
if len(listNews) > 0:
c.listNews = listNews
clustersToReturn.append(c)
# WE NEED TO SORT THE CLUSTERS BY SOME PARAMETER (MAYBE THE SIZE OF THE CLUSTER)
obj = {'listClusters' : [c.serialize() for c in clustersToReturn]}
result = json.dumps(obj, ensure_ascii=True)
return Response(result, mimetype='text/plain')
def getClusters():
feedback = ''
str = ''
try:
newsPosts = crawler.take_all_news_posts()
# utility dicts for majority voting with naive bayes
fileToRead = open(naivebayes_classification.str_dict_word_in_cat)
dict_words = Unpickler(fileToRead).load()
fileToRead.close()
fileToRead = open(naivebayes_classification.str_dict_cat_count)
dict_cats = Unpickler(fileToRead).load()
fileToRead.close()
fileToRead = open(naivebayes_classification.str_dict_priors)
dict_priors = Unpickler(fileToRead).load()
fileToRead.close()
feedback += 'took the newsposts \n'
#return Response('%d' % counter, mimetype='text/plain')
clusters, innerfeedback = clustering.cluster_news(newsPosts)
feedback += '%s\n' % innerfeedback
feedback += 'done the clustering\n'
i = 0
feedback += 'num of clusters: %d\n' % len(clusters)
clusters = sorted(clusters, key=lambda x:-len(x.posts))
for c in clusters:
feedback += 'getting posts from cluster\n'
newsInCluster = c.posts
feedback += 'got the posts from cluster\n'
str += 'cluster %d\n' % i
#implementing the majority voting
votes_cat = {}
for np in newsInCluster:
str += ' \t %s\n' % np.title
category = test_classifications.get_NB_category(np.words,dict_words, dict_cats, dict_priors)
votes_cat[category] = 1 + votes_cat.get(category, 0)
maxVotes = 0
maxCat = ''
for cat in votes_cat:
if votes_cat[cat] > maxVotes:
maxVotes = votes_cat[cat]
maxCat = cat
feedback += '^^^ CLUSTER CATEGORY: %s with maxVotes: %d\n' % (maxCat, maxVotes)
listNews = []
feedback += ' number of posts in cluster %d\n' % len(c.posts)
for np in c.posts:
feedback += 'trying to create NewsPostClient\n'
feedback += 'title: %s \n' % np.title
feedback += 'numWords: %d\n' % np.numWords
feedback += 'url: %s\n' % np.url
newNews = NewsPostClient(url = np.url, host_page = np.host_page, title = np.title, numWords = np.numWords, source_id = np.source_id,
source_url = np.source_url,
img_url = np.img_url, description = np.description)
feedback += 'created NewsPostClient'
listNews.append(newNews)
feedback += 'appended newNews\n'
newCluster = Cluster(category = maxCat, listNews = listNews)
newCluster.put()
str += '\n'
i += 1
str += feedback
except Exception as inst:
feedback += 'Exception type: %s\n' % (type(inst))
feedback += 'Exception: %s\n' % (inst.message)
str += feedback
return Response(str, mimetype='text/plain')
def addCluster(self, r, x, y, v, theta, lambda0, color, is_point):
self.clusters_list.append(
Cluster(r, x, y, v, theta, lambda0, color, is_point))
