def recommend():
redirectPage = 'recommendations.html'
data = {}
productid2name = {}
userRatings= {}
db = connect_to_db()
cur = db.cursor(cursor_factory=psycopg2.extras.DictCursor)
try:
cur.execute(movieRatingQuery)
results = cur.fetchall()
except Exception as e:
print("Error: SEARCH in 'movie_ratings table: %s" % e)
for row in results:
user = row['id']
movie = row['movie_id']
rating = float(row['rating'])
if user in data:
currentRatings = data[user]
else:
currentRatings = {}
currentRatings[movie] = rating
data[user] = currentRatings
try:
cur.execute(movieIDQuery)
results = cur.fetchall()
except Exception as e:
print("Error: SEARCH in 'movie_titles' table: %s" % e)
cur.close()
db.close()
movieLens = recommender(5, 15) #Manhattan Distance 5 Nearest Neighbors
def coffee_table_seeker():
image_url = str(request.form['image_url'])
description = str(unicode(request.form['description']).encode('ascii', 'ignore'))
price_limit = str(request.form['price_limit'])
category = 'coffee_table'
cat = 'coffee table'
form_action = "/" + category + "_seeker#portfolio"
if len(image_url) > 0:
try:
image = imread(image_url)
except IOError:
return render_template('input.html', cat=cat, msg='Oops! That was not a image url. ', form_action=form_action)
else:
image = None
if len(price_limit) > 0:
try:
price_limit = int(price_limit)
except ValueError:
return render_template('input.html', cat=cat, msg='Oops! Please enter a number in the price limit box. ', form_action=form_action)
else:
price_limit = None
final_df = recommender(image=image, text=description, category=category, pca_scaler_dict=pca_scaler_dict,
pca_model_dict=pca_model_dict, tfidf_dict=tfidf_dict, all_info_df_dict=all_info_df_dict,
color=False, price_limit=price_limit)
if final_df is None:
return render_template('input.html', cat=cat, msg='Oops! No items found! Please increase the price limit. ', form_action=form_action)
base_path = '../static/img/wayfair/' + category + '/'
return render_template('seeker.html', cat=cat, df=final_df, base_path=base_path, form_action=form_action)
def main():
print("This is the demo for the recommender!")
# because I don't have the keywords for'Office for Students with Disabilities'
dummy_services = SERVICES
# shall be the review csv file, use the dummy one I created for testing
review_database = REVIEW_DATABASE
# points to the folder containing the keywords for each service
service_folder = SERVICE_KEYWORD_FOLDER
print("Services are: {}".format(dummy_services))
# init the recommender
rr = recommender.recommender(n_service=N_SERVICE,
service_list=dummy_services,
n_fixed_feature=10,
top_k=TOP_K,
cut_off=CUTOFF,
review_file=review_database,
service_kw_folder=service_folder,
n_keyword=7,
learning_rate=0.0015)
# randomly init weights!
rr.init_weights()
print("recommender model inited!")
while (True):
print("Key in user profile:\n")
user_profile = dict()
for f in U_FEATURES:
user_profile[f] = input("\t{}: ".format(f))
print(user_profile)
print(
"Preset user's goal: I want spend time talking iwth someone. I need advice."
)
user_goal = [['time', 0.5], ['talk', 0.5], ['friendly', 0.5],
['advice', 0.5]]
user_rating = {
'Group Therapy': (3, 5),
'Vent Over Tea': (5, 5),
'7 Cups': (4, 5)
}
print("Preset user's rating: user_rating")
# get the service index of top_k choices, start from 0
choices, service_names = rr.get_recommendation(user_profile,
user_goal,
verbose=False)
# print("returned indices (start from 0): {}".format(choices))
choices_for_chatbot = [c + 1 for c in choices]
print(
"service indicies (start from 1): {}".format(choices_for_chatbot))
print("returned names:{}".format(service_names))
# update the weights (ML-ish learning)
user_scores = rr.process_user_rating(user_rating)
rr.update_weights(user_profile, user_scores, choices, verbose=False)
def results():
if request.method == 'POST':
#get the data from the form
className = request.form.get('className', None)
classNumber = request.form.get('classNumber', None)
instructor = request.form.get('instructor', None).upper()
classInput = className.upper() + "-" + classNumber
#retrive data from "server"
result = parse.getClass(classInput, instructor, mainDictionary)
#check if the input matched a class
if not result:
return render_template('Error.html',
results = classInput,
instructor = instructor)
else:
final = recommender.recommender(classInput, instructor, result)
return render_template('results.html',
title = 'Results',
results = final)
elif request.method == 'GET':
return render_template('results.html')
def searchAdaptation(user, timestamp, event, idView):
#global mutex
#while mutex:# getSemaforo(session):
# pass
#mutex = 1#setSemaforo(session)
#atualiza o sumario
#global sumario
##aqui arquivo
sumarioFile = open("sumario.csv", "r")
dataSumario = sumarioFile.readlines()
sumarioAux = list()
for i in dataSumario:
sumarioAux.append(i[0:-1])
sumarioFile.close()
sumario = LoadSummarizerByUser(user, timestamp, event, idView, sumarioAux)
sumarioAux = open("sumario.csv", "w")
sumarioAux.close
sumarioAux = open("sumario.csv", "a+")
for i in sumario:
sumarioAux.write(i + "\n")
sumarioAux.close
#sumarioL = sumario
recomendationUser = analytics(user, sumario, idView)
print "sumario"
print sumario
dt = datetime.now()
strTimeDate = str(dt.day) + "/" + str(dt.month) + "/" + str(
dt.year) + "-" + str(dt.hour) + ":" + str(dt.minute) + ":" + str(
dt.second) + "-" + user + ";" + str(sumario) + "\n"
logService = open("SERVICE_log.csv", "a+")
logService.write(strTimeDate)
logService.close()
#print ""
#print "recomendation"
#print recomendation
#print ""
idQuestion = idView.split(":")
feedback = []
if recomendationUser == True:
feedback = recommender(user, recomendation, int(idQuestion[1]),
int(timestamp))
else:
feedback = lastRecommendation(user, recomendation, int(idQuestion[1]),
int(timestamp))
mutex = 0 #releaseSemaforo
print "feedback recomendation\n\n"
print feedback[0]
print ""
return feedback
def index(self, path=flickr_test_set_path[9]):
print "index page, sample path with", path
img = caffe.io.load_image(path)
print "start predict"
res = self.our_model.predict(img)
res["recommended"] = recom.recommender(res)
print res
print "end predict"
return json.dumps(res)
def get_recommendations(id, n_recommendations):
# query = str(request.args['Query'])
# return query
d = {}
# d['Query'] = str(request.args['Query'])
rec = recommender()
d['category'] = rec.make_recommendations(id, n_recommendations)
print(d['category'])
return jsonify(d)
def auto_rec_add(username, add_tracks='max', clientID = '<>', clientSecret = '<>'):
pl_uri = raw_input('Please enter playlist URI: ')
uri = pl_uri.split(':')[-1]
username = username.split(':')[-1]
test_class = recommender.recommender(username=username, clientID=clientID, clientSecret=clientSecret)
print('Token Retrieved')
test_class.create_user_instance()
print('User Instance Created')
test_class.user_get_playlist_tracks(uri)
print('Playlist Tracks Retrieved')
test_class.user_get_playlist_track_audio_feat()
print('Track features Retrieved')
test_class.user_playlist_recommend_tracks()
print('Tracks recommended')
test_class.user_playlist_add_tracks(playlist_id=uri, num_tracks_add=add_tracks)
def recommend():
redirectPage = 'recommendations.html'
data = {}
productid2name = {}
userRatings= {}
db = connect_to_db()
cur = db.cursor(cursor_factory=psycopg2.extras.DictCursor)
cur.execute(movieRatingQuery)
results = cur.fetchall()
for row in results:
user = row['id']
movie = row['movie_id']
rating = float(row['rating'])
if user in data:
currentRatings = data[user]
else:
currentRatings = {}
currentRatings[movie] = rating
data[user] = currentRatings
cur.execute(movieIDQuery)
results = cur.fetchall()
for row in results:
movieID = row['id']
title = row['movie_title']
productid2name[movieID] = title
cur.close()
db.close()
movieLens = recommender(5, 15) #Manhattan Distance 5 Nearest Neighbors
movieLens.data = data
print(movieLens.recommend(1))
# movieLens.computeSlopeOneDeviations()
# print("Did I get ere")
# print(movieLens.slopeOneRecommendations(data['1']))
return render_template(redirectPage)
def recommend(test):
print("Do I get here?")
redirectPage = 'recommendations.html'
data = {}
productid2name = {}
userRatings= {}
db = connect_to_db()
cur = db.cursor(cursor_factory=psycopg2.extras.DictCursor)
try:
cur.execute(movieRatingQuery)
results = cur.fetchall()
except Exception as e:
print("Error: SEARCH in 'movie_ratings table: %s" % e)
for row in results:
user = row['id']
movie = row['movie_id']
rating = float(row['rating'])
if user in data:
currentRatings = data[user]
else:
currentRatings = {}
currentRatings[movie] = rating
data[user] = currentRatings
try:
cur.execute(movieIDQuery)
results = cur.fetchall()
except Exception as e:
print("Error: SEARCH in 'movie_titles' table: %s" % e)
cur.close()
db.close()
movieLens = recommender(5, 15) #Manhattan Distance 5 Nearest Neighbors
movieLens.data = data
results = movieLens.recommend(session['id'])
print(results)
queryResults = []
for i,movie in results:
queryResults.append({'text': movie[0]})
print(queryResults)
emit('recommendationResults', queryResults)
def recommend(test):
data = {}
db = connect_to_db()
cur = db.cursor(cursor_factory=psycopg2.extras.DictCursor)
error = ''
try:
cur.execute(movieRatingQuery)
results = cur.fetchall()
except Exception as e:
print("Error: SEARCH in 'movie_ratings table: %s" % e)
if len(results) == 0:
error += 'You have not rated any movies yet!'
emit('error', {'exists' : len(error) != 0, 'message' : error, 'type' : 'recommend'})
else:
for row in results:
user = row['id']
movie = row['movie_id']
rating = float(row['rating'])
if user in data:
currentRatings = data[user]
else:
currentRatings = {}
currentRatings[movie] = rating
data[user] = currentRatings
try:
cur.execute(movieIDQuery)
results = cur.fetchall()
except Exception as e:
print("Error: SEARCH in 'movie_titles' table: %s" % e)
cur.close()
db.close()
movieLens = recommender(5, 15) #Manhattan Distance 5 Nearest Neighbors
movieLens.data = data
results = movieLens.recommend(session['id'])
print(results)
queryResults = []
for i,movie in enumerate(results):
queryResults.append({'text': movie[0], 'rank': str(i+1)})
emit('recommendationResults', {'results' : queryResults})
def main():
print("This is the demo for the recommender!")
# because I don't have the keywords for'Office for Students with Disabilities'
dummy_services = [
'Peer Support Center', 'Therapist Assisted Online', 'Group Therapy',
'PhD Support Group', 'Vent Over Tea', '7 Cups',
'McGill Student\'s Nightline', 'Project 10 Listening Line',
'SACOMSS Support Group', 'The Buddy Programm'
]
# shall be the review csv file, use the dummy one I created for testing
review_database = REVIEW_DATABASE
# points to the folder containing the keywords for each service
service_folder = SERVICE_KEYWORD_FOLDER
# init the recommender
rr = recommender.recommender(n_service=10,
service_list=dummy_services,
n_fixed_feature=4,
top_k=TOP_K,
cut_off=CUTOFF,
review_file=review_database,
service_kw_folder=service_folder,
n_keyword=7,
learning_rate=0.0015)
# randomly init weights!
rr.init_weights()
print("recommender model inited!")
while (True):
print("Key in user profile:\n")
user_profile = dict()
user_profile["ufeature1"] = input("Gender [f/m]:")
user_profile["ufeature2"] = input("Language [fr/en]:")
user_profile["ufeature3"] = input("Year of Study, e.g. U3:")
user_profile["ufeature4"] = input("cultural background [ca/us/asian]:")
print(user_profile)
print(
"Preset user's goal: I want spend time talking iwth someone. I need advice."
)
user_goal = [['time', 0.5], ['talk', 0.5], ['friendly', 0.5],
['advice', 0.5]]
print(
"Preset user's rating: I want spend time talking iwth someone. I need advice."
)
user_rating = {
'Group Therapy': (3, 5),
'Vent Over Tea': (5, 5),
'7 Cups': (4, 5)
}
# get the service index of top_k choices, start from 0
choices, service_names = rr.get_recommendation(user_profile,
user_goal,
verbose=False)
# print("returned indices (start from 0): {}".format(choices))
choices_for_chatbot = [c + 1 for c in choices]
print(
"service indicies (start from 1): {}".format(choices_for_chatbot))
print("returned names:{}".format(service_names))
# update the weights (ML-ish learning)
user_scores = rr.process_user_rating(user_rating)
rr.update_weights(user_profile, user_scores, choices, verbose=False)
from parsing_text import parse_train
from parse_test import parse_test
from recommender2 import recommender2
from reco import reco
from recommender import recommender
from recommender_item import recommender_item
from output import output
user_train,movie_train=parse_train('train.txt')
user_test5=parse_test('test5.txt')
user_test10=parse_test('test10.txt')
user_test20=parse_test('test20.txt')
user_test1=parse_test('test1.txt')
k=4
reco_dict=recommender(user_train,user_test20,movie_train,k)
#reco_dict1=recommender(user_train,user_test20,movie_train,k)
#reco_dict2=recommender2(user_train,user_test20,movie_train,k)
#reco_dict=reco(reco_dict1,reco_dict2)
#reco_item=recommender_item(user_train,user_test20,movie_train,k)
output(reco_dict,'test20.txt','result20.txt')
#output(reco_item,'test20.txt','result20.txt')
import psycopg2 as pg import os from recommender import recommender # get heroku database URL DATABASE_URL = os.environ['DATABASE_URL'] # Postgres connection sql_conn = pg.connect(DATABASE_URL, sslmode='require') # create recommender my_recommender = recommender(sql_conn)
import preprocessing
import recommender
## dev
target_users_list, users, metadata, magazine, read, read_each_article = preprocessing.data_preprocessing(
'dev')
recommed = recommender.recommender(target_users_list, './recommend.txt')
## test
#target_users_list, users, metadata, magazine, read, read_each_article = data_preprocessing('test')
#recommed = recommender(target_users_list, './recommend.txt')
def main(movieId):
reco = recommender(movieId=movieId, recommendationsNumber=3)
return reco
movie.append(row[0])
i = 0
for rating in row:
if i > 0:
if rating != '':
users[i].ratings[row[0]] = float(rating)
i=i+1
# convert list of User objects to dictionary of user ratings - {'name' : {'movie' : 0-5, 'movie2' : 0-5 ... 'movie_n' : 0-5}}
userDict = {}
for user in users:
userDict[user.name] = user.ratings
#
r = recommender(userDict)
#print('movie list', movie)
#use movie list to populate recommender.productid2name dictionary - {'film name' : 'film name'}
for film in movie:
r.productid2name[film] = film
#use user User list to populate recommender.username2id and recommender.userid2name = {'name' : 'name'}
for user in users:
r.username2id[user.name] = user.name
r.userid2name[user.name] = user.name
#print(r.productid2name)
#here, for recommender.recommend and recommend.userRatings to function,
# userID and productId lists need to be constructed if they are to utilize the
#!/usr/bin/python3
from flask import Flask, request, jsonify, render_template
from flask_restful import Resource, Api
from json import dumps
from movielens import *
app = Flask(__name__)
api = Api(app)
from recommender import recommender
r = recommender({})
r.loadmovieDB()
user = []
item = []
d = Dataset()
d.load_users("data/u.user", user)
d.load_items("data/u.item", item)
# print(item[0])
class Users(Resource):
def get(self):
return jsonify({'users': [i for i in range(1,1683)]})
class UsersDetail(Resource):
def get(self,user_id):
result = user[int(user_id)-1]
result_dict = {'id':result.id, 'age':result.age, 'sex':result.sex, 'occupation':result.occupation} #avg_r to be added
return jsonify(result_dict)
uri = '4HJ7mSMtHAdU55lLjGE4zW'
# uri = '17kv7O2z3dLHW0LnXaU5Pn'
# uri = '0SuFqlCe5i30Fr75ZlPQVT'
# genre = 'popindie'
genre = input('Enter a genre: ')
num_tracks = 15
### CALL RECOMMENDER
playlist = r.recommender(uri, genre, num_tracks)
# r.create_playlist_file(playlist['track_id'])
print(playlist)
# results = r.pop_track_reccommender(df, track)
# ## SELECT TOP NUMBER
# rec_tracks = r.top_recommended_tracks(results, 15)
# print(r.display_playlist(rec_tracks['track_id']))
# playlist_file = open('playlist.txt', 'r+')
# playlist = playlist_file.read().splitlines()
# ## create playlist
import recommender
rec_unit = recommender.recommender()
for uid in range(100):
print(uid)
res = rec_unit.predict([[uid]], 'feature_name')
print(res)
"Enter movie names and rating (both likes and dislikes) in the side panel to build your profile."
)
st.write("")
st.write("")
st.write("")
st.sidebar.markdown("")
userLikes = []
n = st.sidebar.number_input("How many movies do you want to enter?", value=0)
i = 0
while i < n:
movies = {}
movie = st.sidebar.text_input("Movie Title", key=i)
rating = st.sidebar.slider("Rating", min_value=0.01, max_value=5.01, key=i)
movies["title"] = movie.lower()
movies["rating"] = rating
userLikes.append(movies)
i += 1
st.sidebar.markdown("")
st.sidebar.markdown("")
year = st.sidebar.text_input(
"Do you want movies after a specific year, if yes, mention the year?",
value=1900)
noOfMovies = st.sidebar.number_input("How many movies do you want?", value=10)
if st.sidebar.button("Press for Recommendations", key=1):
recommended = recommender(userLikes, year, noOfMovies)
st.write(recommended)
import recommender
if __name__=='__main__':
dictPath='prediction_dict.pkl'
n=10
rc=recommender.recommender(dictPath,n)
rc.train()
print('train.py complete')
def recommend(): #controller function
df = load_data()
recc = r.recommender()
song = song_in(df)
recs = recc.recommend(data=df, song=song)
printresults(song, recs)
def start_gui(self):
#start mouse and keyboard listeners
self.keyboard_listener.start()
self.mouse_listener.start()
#######this blocks dont use########
#self.keyboard_listener.join()
#self.mouseListener.join()
###################################
in_text=""
out_text=""
extra=""
while True:
event, values = self.window.read()
if event=='in':
print()
if event=='out':
out_text=values['out']
if event=='extra':
extra=values['extra']
if event == "_grab_" and self.grab_area_isSet()==True:
img=ImageGrab.grab(bbox=(int(self.window['mouse_x_start'].get()),int(self.window['mouse_y_start'].get()),int(self.window['mouse_x_end'].get()),int(self.window['mouse_y_end'].get())),include_layered_windows=True,all_screens=True)
#img.show()
img.save(self.img_path)
self.window['_image_'].update(filename= self.img_path)
#print("grab")
img_dealer=img=ImageGrab.grab(bbox=(int(self.window['mouse_x_start'].get()),int(self.window['mouse_y_start'].get()),int(self.window['mouse_x_end'].get()),int(self.window['mouse_y_start'].get())+(int(self.window['mouse_y_end'].get())-int(self.window['mouse_y_start'].get()))/2),include_layered_windows=True,all_screens=True)
img_dealer.save(self.img_dealer_path)
img_player=img=ImageGrab.grab(bbox=(int(self.window['mouse_x_start'].get()),int(self.window['mouse_y_start'].get())+(int(self.window['mouse_y_end'].get())-int(self.window['mouse_y_start'].get()))/2,int(self.window['mouse_x_end'].get()),int(self.window['mouse_y_end'].get())),include_layered_windows=True,all_screens=True)
img_player.save(self.img_player_path)
#img_dealer.show()
#img_player.show()
if event=='_calc_':
#send to model
img=Image.open( self.img_path).convert("RGB")
img_dealer=Image.open( self.img_dealer_path).convert("RGB")
img_player=Image.open(self.img_player_path).convert("RGB")
#img=(read_image('data\grab1.png')).float()
a = transforms.Compose([transforms.ToTensor(),transforms.Normalize(mean=[0.485, 0.456, 0.406],std=[0.229, 0.224, 0.225] )])
img=a(img)
img_dealer=Image.open(self.img_dealer_path).convert("RGB")
img_dealer=a(img_dealer)
img_player=Image.open(self.img_player_path).convert("RGB")
img_player=a(img_player)
#predictions = self.model([img])
predictions_dealer = self.model([img_dealer])
predictions_player = self.model([img_player])
#pred_boxes=predictions[0]['boxes'].detach().numpy()
#pred_labels=predictions[0]['labels'].detach().numpy()
#pred_scores=predictions[0]['scores'].detach().numpy()
half_image=(int(self.window['mouse_y_end'].get())-int(self.window['mouse_y_start'].get()))/2
#print('half image:',half_image)
dealer_pred_boxes=predictions_dealer[0]['boxes'].detach().numpy()
dealer_pred_labels=predictions_dealer[0]['labels'].detach().numpy()
dealer_pred_scores=list(predictions_dealer[0]['scores'].detach().numpy())
player_pred_boxes=predictions_player[0]['boxes'].detach().numpy()
player_pred_labels=predictions_player[0]['labels'].detach().numpy()
player_pred_scores=predictions_player[0]['scores'].detach().numpy()
out_img=Image.open( self.img_path)
draw=ImageDraw.Draw(out_img)
#for i in range(0,len(pred_scores)):
# if pred_boxes[i][1]< half_image and pred_boxes[i][3]<half_image:
# #dealers
# dealer_pred_boxes.append(pred_boxes[i])
# draw.rectangle([(pred_boxes[i][0],pred_boxes[i][1]),(pred_boxes[i][2],pred_boxes[i][3])],outline='red',width=2)
# dealer_pred_labels.append(pred_labels[i])
# dealer_pred_scores.append(pred_scores[i])
# elif pred_boxes[i][1]> half_image and pred_boxes[i][3]>half_image:
# #players
# player_pred_boxes.append(pred_boxes[i])
# draw.rectangle([(pred_boxes[i][0],pred_boxes[i][1]),(pred_boxes[i][2],pred_boxes[i][3])],outline='black',width=2)
# player_pred_labels.append(pred_labels[i])
# player_pred_scores.append(pred_scores[i])
for i in range(0,len(dealer_pred_scores)):
draw.rectangle([(dealer_pred_boxes[i][0],dealer_pred_boxes[i][1]),(dealer_pred_boxes[i][2],dealer_pred_boxes[i][3])],outline='red',width=1)
for i in range(0,len(player_pred_scores)):
draw.rectangle([(player_pred_boxes[i][0],half_image+player_pred_boxes[i][1]),(player_pred_boxes[i][2],half_image+player_pred_boxes[i][3])],outline='black',width=1)
player_pred_labels=list(set(player_pred_labels))
print(player_pred_labels)
out_img.save( self.img_path)
self.window['_image_'].update(filename= self.img_path)
print('dealer:')
print(dealer_pred_boxes)
print(dealer_pred_labels)
print(dealer_pred_scores)
print('player:')
print(player_pred_boxes)
print(player_pred_labels)
print(player_pred_scores)
self.list=[]
self.window['_list_'].update(self.list)
dealer_sum=(dealer_pred_labels[dealer_pred_scores.index(max(dealer_pred_scores))] if len(dealer_pred_scores)>0 else -1) if (dealer_pred_labels[dealer_pred_scores.index(max(dealer_pred_scores))] if len(dealer_pred_scores)>0 else -1) <10 else 10
n=int(self.window['in'].get() if self.window['in'].get() !='' else 2 )
print('nbr of cards:',n)
player_sum=0
player_cards=[]
for i,j in zip(range(0,n), range(0,len(player_pred_labels))):
player_sum=player_sum+(player_pred_labels[i] if player_pred_labels[i] <10 else 10)
player_cards.append(player_pred_labels[i])
print(player_sum)
self.list.append(('dealer sum:', dealer_sum ))
self.list.append(('dealer card:', (dealer_pred_labels[dealer_pred_scores.index(max(dealer_pred_scores))] if len(dealer_pred_scores)>0 else -1)))
self.window['_list_'].update(self.list)
self.list.append(('player sum:',player_sum))
self.list.append(('player cards:',player_cards))
self.window['_list_'].update(self.list)
if player_sum <21 and dealer_sum !=-1 and player_sum:
#send to recommender system
recommendation = recommender.recommender(a=0,p=player_sum,d=dealer_sum)
#make a decision
if recommendation.thorpe()==0:
self.list.append(('recommendation: STAND'))
self.window['_list_'].update(self.list)
#self.click_stand()
elif recommendation.thorpe()==1:
self.list.append(('recommendation: HIT'))
self.window['_list_'].update(self.list)
#self.click_hit()
elif recommendation.thorpe()==2:
self.list.append(('recommendation: DOUBLE'))
self.window['_list_'].update(self.list)
#self.click_double()
else:
self.list.append(('!!! BUST !!!!'))
self.window['_list_'].update(self.list)
if event=='_remove_':
if len(values['_list_'])>0:
self.list.remove(in_text+';'+out_text+';'+extra)
self.window['_list_'].update(self.list)
if event=='_list_':
if len(values['_list_'])>1:
list_item_split=values['_list_'][0].strip().split(';')
self.window['in'].update(list_item_split[0])
in_text=values['in']
self.window['out'].update(list_item_split[1])
out_text=values['out']
self.window['extra'].update(list_item_split[2])
extra=values['extra']
if event == sg.WIN_CLOSED:
break
"Chan": {"Blues Traveler": 5.0, "Broken Bells": 1.0,
"Deadmau5": 1.0, "Norah Jones": 3.0, "Phoenix": 5,
"Slightly Stoopid": 1.0},
"Dan": {"Blues Traveler": 3.0, "Broken Bells": 4.0,
"Deadmau5": 4.5, "Phoenix": 3.0,
"Slightly Stoopid": 4.5, "The Strokes": 4.0,
"Vampire Weekend": 2.0},
"Hailey": {"Broken Bells": 4.0, "Deadmau5": 1.0,
"Norah Jones": 4.0, "The Strokes": 4.0,
"Vampire Weekend": 1.0},
"Jordyn": {"Broken Bells": 4.5, "Deadmau5": 4.0,
"Norah Jones": 5.0, "Phoenix": 5.0,
"Slightly Stoopid": 4.5, "The Strokes": 4.0,
"Vampire Weekend": 4.0},
"Sam": {"Blues Traveler": 5.0, "Broken Bells": 2.0,
"Norah Jones": 3.0, "Phoenix": 5.0,
"Slightly Stoopid": 4.0, "The Strokes": 5.0},
"Veronica": {"Blues Traveler": 3.0, "Norah Jones": 5.0,
"Phoenix": 4.0, "Slightly Stoopid": 2.5,
"The Strokes": 3.0}
}
r = recommender.recommender(users)
#print "hello"
#r.loadBookDB('/home/abhiroop/Downloads/WISHLIST/BX-Dump/')
#print r.recommend('171118')
print r.recommend('Jordyn')
def receive_data(idSession):
if request.method == "POST":
try:
if verify_active_session(idSession) == False:
create_session(idSession)
idUser = request.form["idUser"]
event = request.form["tipo"]
resource = request.form["tag"]
timestamp = request.form["timeStamp"]
x = request.form["x"]
y = request.form["y"]
#idView = request.form["id"]
tela = request.form["tela"]
adaptacao = request.form["adaptacao"]
idView = request.form["classId"]
#print tela
if x == "" and y == "" and resource == "":
#Neste bloco está sendo atualizado a última linha que significa o timestamp e a tela atual passado pelo Player
with open(
"sessions-Logs/" + idSession + "/" + idUser +
"_log.csv", "a+") as f:
lines = f.readlines()
if len(lines) > 0:
lineOld = lines[-1] #pega ultima linha
lineNew = idSession + ";" + idUser + ";" + timestamp + ";" + event + ";" + tela + ";" + idView + ";" + resource + ";" + x + ";" + y + "\n"
f.close()
fileaux = open(
"sessions-Logs/" + idSession + "/" + idUser +
"_log.csv", 'r')
filedata = fileaux.read()
fileaux.close()
newdata = filedata.replace(lineOld, lineNew)
fileaux = open(
"sessions-Logs/" + idSession + "/" + idUser +
"_log.csv", 'w')
fileaux.write(newdata)
fileaux.close()
else:
newdata = idSession + ";" + idUser + ";" + timestamp + ";" + event + ";" + tela + ";" + idView + ";" + resource + ";" + x + ";" + y + "\n"
fileaux = open(
"sessions-Logs/" + idSession + "/" + idUser +
"_log.csv", 'w')
fileaux.write(newdata)
fileaux.close()
else:
with open(
"sessions-Logs/" + idSession + "/" + idUser +
"_log.csv", "a+") as f:
lines = f.readlines()
lineOld = lines[-1] #pega ultima linha
f.close()
#print lines
fileaux = open(
"sessions-Logs/" + idSession + "/" + idUser +
"_log.csv", 'r')
filedata = fileaux.read()
fileaux.close()
lineNew = idSession + ";" + idUser + ";" + timestamp + ";" + event + ";" + tela + ";" + idView + ";" + resource + ";" + x + ";" + y + "\n"
newdata = filedata.replace(lineOld, lineNew)
newdata = newdata + lineOld
#print newdata
fileaux = open(
"sessions-Logs/" + idSession + "/" + idUser +
"_log.csv", 'w')
fileaux.write(newdata)
fileaux.close()
#data_received = idSession+";"+idUser+";"+timestamp+";"+event +";"+idView+";"+resource+";"+x+";"+y+"\n"
#fileBuffer = open("sessions-Logs/"+idSession+"/"+idUser+"_log.csv", "a")
#fileBuffer.write(data_received)
#fileBuffer.close()
print idSession + ";" + idUser + ";" + timestamp + ";" + event + ";" + tela + ";" + idView + ";" + resource + ";" + x + ";" + y
print ""
dt = datetime.now()
strTimeDate = str(dt.day) + "/" + str(dt.month) + "/" + str(
dt.year
) + "-" + str(dt.hour) + ":" + str(dt.minute) + ":" + str(
dt.second
) + "-" + idUser + ";" + timestamp + ";" + event + ";" + tela + ";" + idView + ";" + resource + ";" + x + ";" + y + "\n"
logService = open("SERVICE_log.csv", "a+")
logService.write(strTimeDate)
logService.close()
#print "btn Storage" + btnTroca[0]
#Solicita recomendacao caso o aluno estiver em uma questao
try:
if idView != "" and (
idView == "troca-Q1 btn btn-info btn-lg"
or idView == "troca-Q2 btn btn-info btn-lg"
or idView == "troca-Q3 btn btn-info btn-lg"
or idView == "troca-Q4 btn btn-info btn-lg"
or idView == "troca-Q5 btn btn-info btn-lg"
or idView == "troca-Q6 btn btn-info btn-lg"
or idView == "troca-Q7 btn btn-info btn-lg"
or idView == "troca-Q8 btn btn-info btn-lg"
or idView == "troca-Q9 btn btn-info btn-lg"
or idView == "troca-Q10 btn btn-info btn-lg"):
#btnTroca[0] = "1"
print "IDVIEW: " + idView
auxIdView = idView.split(" ")
aux = auxIdView[0]
print "aux" + aux
i = aux.find("Q")
print "AUX: " + aux[i + 1:]
changeBtnQuestionEasy(idUser, recomendation,
int(aux[i + 1:]), int(timestamp))
#print "AQUI BTN"
#print "aquiii"
#print sumario
if idView != "" and adaptacao == "1" and (
idView == "R:1:E:1-1 btn btn-info btn-lg"
or idView == "R:2:E:1-1 btn btn-info btn-lg"
or idView == "R:3:E:1-1 btn btn-info btn-lg"
or idView == "R:4:E:1-1 btn btn-info btn-lg"
or idView == "R:5:E:1-1 btn btn-info btn-lg"
or idView == "R:6:E:1-1 btn btn-info btn-lg"
or idView == "R:7:E:1-1 btn btn-info btn-lg"
or idView == "R:8:E:1-2 btn btn-info btn-lg"
or idView == "R:9:E:1-2 btn btn-info btn-lg"
or idView == "R:10:E:1-2 btn btn-info btn-lg"):
#btnTroca[0] = "1"
print "IDVIEW: " + idView
auxIdView = idView.split(":")
aux = auxIdView[1]
print "aux: " + aux
answerBtnQuestionEasy(idUser, recomendation, int(aux),
int(timestamp))
#idViewSplit = idView.split(":")
#print idViewSplit
#updateQuestionsTime(idUser, timestamp, idView)
#print "aqui1"
#print "idview: " + idView
idQuestion = idView.split(":")
idQuestion[0] = "Q"
#print "aqui2"
feedback = []
feedback = recommender(idUser, recomendation,
int(idQuestion[1]), int(timestamp))
print "Recomendação(SENT): " + feedback[1]
print "Questions (SENT): " + feedback[2]
dt = datetime.now()
strTimeDate = str(dt.day) + "/" + str(dt.month) + "/" + str(
dt.year
) + "-" + str(dt.hour) + ":" + str(dt.minute) + ":" + str(
dt.second
) + "-" + idUser + ";" + "Recomendacao(SENT): " + feedback[
1] + ";" + "Questions (SENT): " + feedback[2] + "\n"
#strTimeDate = "ok\n"
logService = open("SERVICE_log.csv", "a+")
logService.write(strTimeDate)
logService.close()
#Zerando sumario
sumarioFile = open("sumario.csv", "r")
dataSumario = sumarioFile.readlines()
sumarioAux = list()
for i in dataSumario:
sumarioAux.append(i[0:-1])
sumarioFile.close()
print "\nSUMARIO ANTES DE ZERAR CONTADOR"
print sumarioAux
sumario = ClearSummarizerByUser(idUser, event, timestamp,
idView, sumarioAux)
print "\nSUMARIO ZERANDO CONTADOR"
print sumario
sumarioAux = open("sumario.csv", "w")
sumarioAux.close
sumarioAux = open("sumario.csv", "a+")
for i in sumario:
sumarioAux.write(i + "\n")
sumarioAux.close
recommendation = [{
"recommendation": feedback[1]
}, {
"questions": feedback[2]
}]
#print "----Teste Recommendation OOOOOI------", recommendation
return jsonify({'recommendation': recommendation})
if idView != "" and idView[0] == "Q" and adaptacao == "1":
idViewSplit = idView.split(":")
#print idViewSplit
#updateQuestionsTime(idUser, timestamp, idView)
feedback = searchAdaptation(idUser, timestamp, event,
idView)
#print feedback
if len(feedback) > 0:
print "Recomendação(SENT): " + feedback[1]
print "Questions (SENT): " + feedback[2]
dt = datetime.now()
strTimeDate = str(dt.day) + "/" + str(
dt.month
) + "/" + str(dt.year) + "-" + str(
dt.hour
) + ":" + str(dt.minute) + ":" + str(
dt.second
) + "-" + idUser + ";" + "Recomendacao(SENT): " + feedback[
1] + ";" + "Questions (SENT): " + feedback[2] + "\n"
#strTimeDate = "ok\n"
logService = open("SERVICE_log.csv", "a+")
logService.write(strTimeDate)
logService.close()
recommendation = [{
"recommendation": feedback[1]
}, {
"questions": feedback[2]
}]
#print "----Teste Recommendation OOOOOI------", recommendation
return jsonify({'recommendation': recommendation})
else:
#print "AQUI ELSE1"
recommendation = [{
"recommendation": "ok"
}, {
"questions": "nada"
}]
return jsonify({'recommendation': recommendation})
else:
#idUser, event, timestamp, idView, sumarioL
#print "aqui2"
#global sumario
#print "aqui"
print "" + idUser + "/" + timestamp + "/" + event + "/" + idView + "/"
sumarioFile = open("sumario.csv", "r")
dataSumario = sumarioFile.readlines()
sumarioAux = list()
for i in dataSumario:
sumarioAux.append(i[0:-1])
sumarioFile.close()
print "\nSUMARIO ANTES DE ZERAR CONTADOR"
print sumarioAux
sumario = ClearSummarizerByUser(idUser, event, timestamp,
idView, sumarioAux)
print "\nSUMARIO ZERANDO CONTADOR"
print sumario
sumarioAux = open("sumario.csv", "w")
sumarioAux.close
sumarioAux = open("sumario.csv", "a+")
for i in sumario:
sumarioAux.write(i + "\n")
sumarioAux.close
recommendation = [{
"recommendation": "ok"
}, {
"questions": "nada"
}]
return jsonify({'recommendation': recommendation})
except:
print "Broken1"
dt = datetime.now()
strTimeDate = str(dt.day) + "/" + str(dt.month) + "/" + str(
dt.year) + "-" + str(dt.hour) + ":" + str(
dt.minute) + ":" + str(
dt.second) + "-" + idUser + ";" + "Broken1" + "\n"
logService = open("SERVICE_log.csv", "a+")
logService.write(strTimeDate)
logService.close()
pass
except:
print "Broken2"
dt = datetime.now()
strTimeDate = str(dt.day) + "/" + str(dt.month) + "/" + str(
dt.year) + "-" + str(dt.hour) + ":" + str(
dt.minute) + ":" + str(
dt.second) + "-" + idUser + ";" + "Broken2" + "\n"
logService = open("SERVICE_log.csv", "a+")
logService.write(strTimeDate)
logService.close()
pass
def app():
##### Set up data for modelling
# import data to model CV against
workpac_data = pd.read_csv('job_descriptions.csv')
# remove id column as we can use index
workpac_data = workpac_data[['job_link', 'job_description', 'job_title']]
# clean Na's so functions can run
workpac_data = workpac_data.replace(np.nan, ' ', regex=True)
# clean job description
workpac_data['job_description_cleaned'] = workpac_data[
'job_description'].apply(_removeNonAscii)
workpac_data[
'job_description_cleaned'] = workpac_data.job_description_cleaned.apply(
func=make_lower_case)
workpac_data[
'job_description_cleaned'] = workpac_data.job_description_cleaned.apply(
func=remove_stop_words)
workpac_data[
'job_description_cleaned'] = workpac_data.job_description_cleaned.apply(
func=remove_punctuation)
workpac_data[
'job_description_cleaned'] = workpac_data.job_description_cleaned.apply(
func=remove_html)
workpac_data['job_title_cleaned'] = workpac_data['job_title'].apply(
_removeNonAscii)
workpac_data['job_title_cleaned'] = workpac_data.job_title_cleaned.apply(
func=make_lower_case)
workpac_data['job_title_cleaned'] = workpac_data.job_title_cleaned.apply(
func=remove_stop_words)
workpac_data['job_title_cleaned'] = workpac_data.job_title_cleaned.apply(
func=remove_punctuation)
workpac_data['job_title_cleaned'] = workpac_data.job_title_cleaned.apply(
func=remove_html)
# stem cleaned data
workpac_data['job_description_cleaned'] = workpac_data[
'job_description_cleaned'].apply(stem_sentences)
workpac_data['job_title_cleaned'] = workpac_data[
'job_title_cleaned'].apply(stem_sentences)
workpac_data['job_text_cleaned'] = workpac_data[
'job_title_cleaned'] + ' ' + workpac_data['job_description_cleaned']
model_df = workpac_data[['job_link', 'job_text_cleaned']]
# title of streamlit app
st.title(
'Candidate Job Ranking System Using Natural Language Processing (NLP)')
# sidebar
st.subheader('Files')
# file uploader
uploaded_file = st.file_uploader(label='Please upload your CV',
type=['docx'])
# create message variable
message = ''
doc = ''
if uploaded_file is not None:
try:
doc = docx2python(uploaded_file).text
message = 'File upload successful'
except Exception as e:
message = 'There was an error uploading your file'
# display success/failure message
st.text(message)
# display uploaded file (text only as the method used in docx2python is .text)
try:
st.text(doc)
except Exception as e:
print(e)
st.write('Please upload CV')
if doc != '':
# call recommender function
recommended_jobs = recommender(model_df, doc)
# create output dataframe
recommended_jobs = pd.merge(left=recommended_jobs,
right=workpac_data,
left_on='job_link',
right_on='job_link')
# summarise output
recommender_summary = recommended_jobs[['job_title', 'job_link']]
# display output
fig = go.Figure(data=[
go.Table(header=dict(values=list(recommender_summary.columns),
fill_color='paleturquoise',
align='left'),
cells=dict(values=[
recommender_summary.job_title,
recommender_summary.job_link
],
fill_color='lavender',
align='left'))
])
st.write(fig)
"Vampire Weekend": 1.0
},
"Jordyn": {
"Broken Bells": 4.5,
"Deadmau5": 4.0,
"Norah Jones": 5.0,
"Phoenix": 5.0,
"Slightly Stoopid": 4.5,
"The Strokes": 4.0,
"Vampire Weekend": 4.0
},
"Sam": {
"Blues Traveler": 5.0,
"Broken Bells": 2.0,
"Norah Jones": 3.0,
"Phoenix": 5.0,
"Slightly Stoopid": 4.0,
"The Strokes": 5.0
},
"Veronica": {
"Blues Traveler": 3.0,
"Norah Jones": 5.0,
"Phoenix": 4.0,
"Slightly Stoopid": 2.5,
"The Strokes": 3.0
}
}
r = recommender.recommender(users)
print(r.recommend("Hailey"))
def generateHot(self):
after = self.request.args.get('after')
if after is None:
print 'no after'
msg = None
if g.user:
print 'we have a user'
msg = Database.query(
'''select image from Uploads ul
where (select score from imagescores where image=ul.image) > ?
and ul.image not in (select image from likes where userid=?)
order by (select score from imagescores where image=ul.image) desc
limit ?''', [
config.get('hotFloor'), session['userid'],
config.get('perPage')
])
print[f['image'] for f in msg], 'IMAGES'
else:
print 'no user'
msg = Database.query(
'''select image from Uploads ul
where (select score from imagescores where image=ul.image) > ?
order by (select score from imagescores where image=ul.image) desc
limit ?''',
[config.get('hotFloor'),
config.get('perPage')])
hot = [image(f['image']) for f in msg]
print 'hot: ', hot
recs = []
if g.user:
rc = recommender(session['userid'])
recs = rc.getRecommendations()
print 'recs ', recs
ret = self.__createFeed(hot, recs)
print 'final: ', ret
return ret
print "we found an after parameter, and it is", after
score = Database.query('select score from imagescores where image=?',
[after],
one=True)['score']
msg = None
if g.user:
msg = Database.query(
'''select image from uploads ul
where (select score from imagescores where image=ul.image) < ?
and (select score from imagescores where image=ul.image) > ? and
ul.image not in (select image from likes where userid=?)
order by (select score from imagescores where image=ul.image) desc
limit ?''', [
score,
config.get('hotFloor'), session['userid'],
config.get('perPage')
])
else:
msg = Database.query(
'''select image from uploads ul
where (select score from imagescores where image=ul.image) < ?
and (select score from imagescores where image=ul.image) > ?
order by (select score from imagescores where image=ul.image) desc
limit ?''',
[score, config.get('hotFloor'),
config.get('perPage')])
hot = [image(f['image']) for f in msg]
print 'hot: ', hot
recs = []
if g.user:
rc = recommender(session['userid'])
recs = rc.getRecommendations()
print 'recs ', recs
ret = self.__createFeed(hot, recs)
print 'final: ', ret
return ret
import random
import os
import recommender as rc
users=2100
rcTemp=rc.recommender()
testCase=1
for i in range(testCase):
uid_random=random.randint(1,users)
X=[[uid_random]]
results=rcTemp.predict(X,'featureName')
print('USER ID [%d]' %uid_random)
print(results)
