def process_status(status,proc_posts,movies10,music10,args):
i = args['i']
scores = args['scores']
movie_genres = args['movie']
music_genres = args['music']
picture = args['pic']
title=''
embed=''
description=''
score_now = 0
post_id = status['id']
if post_id not in proc_posts:
proc_posts.append(post_id)
else:
return proc_posts,movies10,music10
created = status['updated_time'][:10]
if 'message' in status:
message = status['message']
else:
message = ''
if 'comments' in status:
comments = status['comments']['data']
else:
comments = ''
f_score = 0 #final score
m_score = analysis.sentiment_analysis(message.encode('utf-8')) #message score
if m_score >=-0.1:
f_score +=1
else:
f_score -=1
for c in comments:
if c['from']['id'] == scores[i]['facebook_id']:
c_score = 0
c_score = analysis.sentiment_analysis(c['message'].encode('utf8')) #comment_score
if c_score >=-0.1:
f_score+=1
else:
f_score-=1
if f_score < 0:
return proc_posts,movies10,music10
message = re.sub('\n',' ', message)
message = re.split('[ .,!?]',message)
message = ' '.join(message[:8])
results = text_analysis_freebase.getResults(message.encode('utf-8'))
for entity in results:
ubuff = 0
if bool(entity):
if entity['type'] == 'movie' and entity['genres']:
for gen in entity['genres']:
if (gen in movie_genres.keys()):
ubuff += movie_genres[gen]
ubuff = float(ubuff)/len(entity['genres'])
score_now = float(scores[i]['movie_score']['movie_friend_score'])*ubuff
if score_now > 0:
title,url,description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in movies10:
if ((post_id == it["post_id"]) and (url == it["embed"])) or url == it["embed"]:
exist_flag = 1
if (exist_flag == 0):
movies10.append({"name":scores[i]['friend_name'],"f_id":scores[i]['facebook_id'],"post_id":post_id,"score":score_now,"message":message.encode('utf-8'),'picture':picture,'title':title,'embed':url,'description':description,'created':created,'genres':entity['genres'],'rated':0})
elif entity['type'] == 'music' and entity['genres']:
for key in entity['genres']:
if key in music_genres.keys():
ubuff += music_genres[key]
ubuff = float(ubuff)/len(entity['genres'])
score_now = float(scores[i]['music_score']['music_friend_score'])*ubuff
if score_now>0.0:
title,url,description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in music10:
if ((post_id == it["post_id"]) and (url == it["embed"])) or url == it["embed"]:
exist_flag = 1
if exist_flag == 0:
music10.append({"name":scores[i]['friend_name'],"f_id":scores[i]['facebook_id'],"post_id":post_id,"score":score_now,"message":message.encode('utf-8'),'picture':picture,'title':title,'embed':url,'description':description,'created':created,'genres':entity['genres'],'rated':0})
return proc_posts,movies10,music10
def process_status(status, proc_posts, movies10, music10, args):
i = args['i']
scores = args['scores']
movie_genres = args['movie']
music_genres = args['music']
picture = args['pic']
title = ''
embed = ''
description = ''
score_now = 0
post_id = status['id']
if post_id not in proc_posts:
proc_posts.append(post_id)
else:
return proc_posts, movies10, music10
created = status['updated_time'][:10]
if 'message' in status:
message = status['message']
else:
message = ''
if 'comments' in status:
comments = status['comments']['data']
else:
comments = ''
f_score = 0 #final score
m_score = analysis.sentiment_analysis(
message.encode('utf-8')) #message score
if m_score >= -0.1:
f_score += 1
else:
f_score -= 1
for c in comments:
if c['from']['id'] == scores[i]['facebook_id']:
c_score = 0
c_score = analysis.sentiment_analysis(
c['message'].encode('utf8')) #comment_score
if c_score >= -0.1:
f_score += 1
else:
f_score -= 1
if f_score < 0:
return proc_posts, movies10, music10
message = re.sub('\n', ' ', message)
message = re.split('[ .,!?]', message)
message = ' '.join(message[:8])
results = text_analysis_freebase.getResults(message.encode('utf-8'))
for entity in results:
ubuff = 0
if bool(entity):
if entity['type'] == 'movie' and entity['genres']:
for gen in entity['genres']:
if (gen in movie_genres.keys()):
ubuff += movie_genres[gen]
ubuff = float(ubuff) / len(entity['genres'])
score_now = float(
scores[i]['movie_score']['movie_friend_score']) * ubuff
if score_now > 0:
title, url, description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in movies10:
if ((post_id == it["post_id"]) and
(url == it["embed"])) or url == it["embed"]:
exist_flag = 1
if (exist_flag == 0):
movies10.append({
"name": scores[i]['friend_name'],
"f_id": scores[i]['facebook_id'],
"post_id": post_id,
"score": score_now,
"message": message.encode('utf-8'),
'picture': picture,
'title': title,
'embed': url,
'description': description,
'created': created,
'genres': entity['genres'],
'rated': 0
})
elif entity['type'] == 'music' and entity['genres']:
for key in entity['genres']:
if key in music_genres.keys():
ubuff += music_genres[key]
ubuff = float(ubuff) / len(entity['genres'])
score_now = float(
scores[i]['music_score']['music_friend_score']) * ubuff
if score_now > 0.0:
title, url, description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in music10:
if ((post_id == it["post_id"]) and
(url == it["embed"])) or url == it["embed"]:
exist_flag = 1
if exist_flag == 0:
music10.append({
"name": scores[i]['friend_name'],
"f_id": scores[i]['facebook_id'],
"post_id": post_id,
"score": score_now,
"message": message.encode('utf-8'),
'picture': picture,
'title': title,
'embed': url,
'description': description,
'created': created,
'genres': entity['genres'],
'rated': 0
})
return proc_posts, movies10, music10
def readStatusAndCreateLists(uid):
client = MongoClient('127.0.0.1')
db = client.recommendation_db
user_coll = db.users
movies10 = []
music10 = []
friends_list = []
query = user_coll.find_one({"id":uid})
user = query
acc_token = user['token']
graph = fb.GraphAPI(acc_token)
scores = user['scores']
lim = len(scores)
music_genres = user['music_genres']
movie_genres = user['movie_genres']
proc_posts = user['proc_posts'] #ore
ex = datetime.now() - timedelta(days=14)
fields = {"fields":'links.since('+str(ex)+').fields(comments.fields(id,from,message).limit(10),message,link,id,name,created_time,description),name,picture,statuses.since('+str(ex)+').fields(comments.fields(id,from,message).limit(10),message,id,updated_time),video.watches.fields(id,data,publish_time,message),music.listens.fields(id,data,publish_time,message)'}
for i in range(0,lim):
args = {}
links_filtered = []
statuses_filtered = []
videos = []
music = []
if not scores[i]['music_score'] and not scores[i]['movie_score']:
continue
try:
statuses = graph.get_object(scores[i]['facebook_id'],**fields)
except:
print str(scores[i]['friend_name'].encode('utf8')) + " - Graph error\n"
continue
if 'links' in statuses:
links_filtered = statuses['links']['data']
if 'statuses' in statuses:
statuses_filtered = statuses['statuses']['data']
if 'video.watches' in statuses:
videos = statuses['video.watches']['data']
if 'music.listens' in statuses:
music = statuses['music.listens']['data']
picture = statuses['picture']['data']['url']
for video in videos:
p_time = video['publish_time']
if 'movie' in video['data'] and p_time > str(ex):
entity = video['data']['movie']['title']
post_id = video['id']
if post_id not in proc_posts:
proc_posts.append(post_id)
else:
continue
check = text_analysis_freebase.search(entity,"movie")
entity = check
if bool(entity):
ubuff = 0
for gen in entity['genres']:
if (gen in movie_genres.keys()):
ubuff += movie_genres[gen]
ubuff = float(ubuff)/len(entity['genres'])
score_now = float(scores[i]['movie_score']['movie_friend_score'])*ubuff
title,url,description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in movies10: #TODO O markos to eixe valei se sxolia
if ((post_id == it["post_id"]) and (url == it["embed"])) or url == it["embed"]: #edw paizei na prepei na be "or" analoga me to ama 8eloume genika na uparxei to 1 vid 1 fora mono.
exist_flag = 1
if (exist_flag == 0):
movies10.append({"name":scores[i]['friend_name'],"f_id":scores[i]['facebook_id'],"post_id":post_id,"score":score_now,'picture':picture,'title':title,'embed':url,'description':description,'created':p_time[:10],'genres':entity['genres'],'rated':0})
for item in music:
p_time = item['publish_time']
if 'musician' in item['data'] and p_time > str(ex):
entity = item['data']['musician']['title']
post_id = item['id']
if post_id not in proc_posts:
proc_posts.append(post_id)
else:
continue
check = text_analysis_freebase.search(entity,"music")
entity = check
if bool(entity):
ubuff=0
for gen in entity['genres']:
if (gen in movie_genres.keys()):
ubuff += movie_genres[gen]
ubuff = float(ubuff)/len(entity['genres'])
score_now = float(scores[i]['movie_score']['movie_friend_score'])*ubuff
title,url,description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in music10:
if ((post_id == it["post_id"]) and (url == it["embed"])) or url == it["embed"]: #edw paizei na prepei na be "or" analoga me to ama 8eloume genika na uparxei to 1 vid 1 fora mono.
exist_flag = 1
if exist_flag == 0:
music10.append({"name":scores[i]['friend_name'],"f_id":scores[i]['facebook_id'],"post_id":post_id,"score":score_now,'picture':picture,'title':title,'embed':url,'description':description,'created':p_time[:10],'genres':entity['genres'],'rated':0})
args['i'] = i
args['scores'] = scores
args['movie'] = movie_genres
args['music'] = music_genres
args['pic'] = picture
for link in links_filtered:
if 'link' in link:
proc_posts, movies10, music10 = process_link(link,proc_posts,movies10,music10,args)
for status in statuses_filtered:
status_flag = 1
if 'message' in status:
lme = status['message']
try:#ore
lang = classify(lme)
if lang[0]!='en':
continue
except Exception as e:
print e
continue
split_buffer = status['message'].split(" ")
for word in split_buffer:
if 'http://' in word:
link_buffer = {}
link_buffer['id'] = status['id']
link_buffer['link'] = word
if 'updated_time' in status:
link_buffer['created_time'] = status['updated_time']
proc_posts, movies10, music10 = process_link(link_buffer,proc_posts,movies10,music10,args)
status_flag = 0
break
if status_flag == 1:
proc_posts, movies10, music10 = process_status(status,proc_posts,movies10,music10,args)
movies10 = sorted(movies10, key=lambda k:k['created'])
movies10 = movies10[::-1]
music10 = sorted(music10, key=lambda k:k['created'])
music10 = music10[::-1]
user_coll.update({"id":uid},{"$set":{"movies10":movies10,"music10":music10,"has_lists":1,"proc_posts":proc_posts}})
def readStatusAndCreateLists(uid):
client = MongoClient('127.0.0.1')
db = client.recommendation_db
user_coll = db.users
movies10 = []
music10 = []
friends_list = []
query = user_coll.find_one({"id": uid})
user = query
acc_token = user['token']
graph = fb.GraphAPI(acc_token)
scores = user['scores']
lim = len(scores)
music_genres = user['music_genres']
movie_genres = user['movie_genres']
proc_posts = user['proc_posts'] #ore
ex = datetime.now() - timedelta(days=14)
fields = {
"fields":
'links.since(' + str(ex) +
').fields(comments.fields(id,from,message).limit(10),message,link,id,name,created_time,description),name,picture,statuses.since('
+ str(ex) +
').fields(comments.fields(id,from,message).limit(10),message,id,updated_time),video.watches.fields(id,data,publish_time,message),music.listens.fields(id,data,publish_time,message)'
}
for i in range(0, lim):
args = {}
links_filtered = []
statuses_filtered = []
videos = []
music = []
if not scores[i]['music_score'] and not scores[i]['movie_score']:
continue
try:
statuses = graph.get_object(scores[i]['facebook_id'], **fields)
except:
print str(
scores[i]['friend_name'].encode('utf8')) + " - Graph error\n"
continue
if 'links' in statuses:
links_filtered = statuses['links']['data']
if 'statuses' in statuses:
statuses_filtered = statuses['statuses']['data']
if 'video.watches' in statuses:
videos = statuses['video.watches']['data']
if 'music.listens' in statuses:
music = statuses['music.listens']['data']
picture = statuses['picture']['data']['url']
for video in videos:
p_time = video['publish_time']
if 'movie' in video['data'] and p_time > str(ex):
entity = video['data']['movie']['title']
post_id = video['id']
if post_id not in proc_posts:
proc_posts.append(post_id)
else:
continue
check = text_analysis_freebase.search(entity, "movie")
entity = check
if bool(entity):
ubuff = 0
for gen in entity['genres']:
if (gen in movie_genres.keys()):
ubuff += movie_genres[gen]
ubuff = float(ubuff) / len(entity['genres'])
score_now = float(
scores[i]['movie_score']['movie_friend_score']) * ubuff
title, url, description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in movies10: #TODO O markos to eixe valei se sxolia
if ((post_id == it["post_id"]) and
(url == it["embed"])) or url == it[
"embed"]: #edw paizei na prepei na be "or" analoga me to ama 8eloume genika na uparxei to 1 vid 1 fora mono.
exist_flag = 1
if (exist_flag == 0):
movies10.append({
"name": scores[i]['friend_name'],
"f_id": scores[i]['facebook_id'],
"post_id": post_id,
"score": score_now,
'picture': picture,
'title': title,
'embed': url,
'description': description,
'created': p_time[:10],
'genres': entity['genres'],
'rated': 0
})
for item in music:
p_time = item['publish_time']
if 'musician' in item['data'] and p_time > str(ex):
entity = item['data']['musician']['title']
post_id = item['id']
if post_id not in proc_posts:
proc_posts.append(post_id)
else:
continue
check = text_analysis_freebase.search(entity, "music")
entity = check
if bool(entity):
ubuff = 0
for gen in entity['genres']:
if (gen in movie_genres.keys()):
ubuff += movie_genres[gen]
ubuff = float(ubuff) / len(entity['genres'])
score_now = float(
scores[i]['movie_score']['movie_friend_score']) * ubuff
title, url, description = youtubeAPI.getVideo(entity)
if title:
exist_flag = 0
for it in music10:
if ((post_id == it["post_id"]) and
(url == it["embed"])) or url == it[
"embed"]: #edw paizei na prepei na be "or" analoga me to ama 8eloume genika na uparxei to 1 vid 1 fora mono.
exist_flag = 1
if exist_flag == 0:
music10.append({
"name": scores[i]['friend_name'],
"f_id": scores[i]['facebook_id'],
"post_id": post_id,
"score": score_now,
'picture': picture,
'title': title,
'embed': url,
'description': description,
'created': p_time[:10],
'genres': entity['genres'],
'rated': 0
})
args['i'] = i
args['scores'] = scores
args['movie'] = movie_genres
args['music'] = music_genres
args['pic'] = picture
for link in links_filtered:
if 'link' in link:
proc_posts, movies10, music10 = process_link(
link, proc_posts, movies10, music10, args)
for status in statuses_filtered:
status_flag = 1
if 'message' in status:
lme = status['message']
try: #ore
lang = classify(lme)
if lang[0] != 'en':
continue
except Exception as e:
print e
continue
split_buffer = status['message'].split(" ")
for word in split_buffer:
if 'http://' in word:
link_buffer = {}
link_buffer['id'] = status['id']
link_buffer['link'] = word
if 'updated_time' in status:
link_buffer['created_time'] = status[
'updated_time']
proc_posts, movies10, music10 = process_link(
link_buffer, proc_posts, movies10, music10, args)
status_flag = 0
break
if status_flag == 1:
proc_posts, movies10, music10 = process_status(
status, proc_posts, movies10, music10, args)
movies10 = sorted(movies10, key=lambda k: k['created'])
movies10 = movies10[::-1]
music10 = sorted(music10, key=lambda k: k['created'])
music10 = music10[::-1]
user_coll.update({"id": uid}, {
"$set": {
"movies10": movies10,
"music10": music10,
"has_lists": 1,
"proc_posts": proc_posts
}
})