def run_assignment_2_4():
print "-"*50, "\n", "assignment 2.4: Item based collaborative filtering (ICF)\n", "-"*50
transCritics = reco.transformCritics(critics)
person = 'Toby'
recommendations_euclid = reco.getRecommendedItems(transCritics, person, reco.sim_euclid)
print "Recommendations for %s (calculated with Euclidean Distance):\n " % (person), recommendations_euclid
recommendations_pearson = reco.getRecommendedItems(transCritics, person, reco.sim_pearson)
print "Recommendations for %s (calculated with Pearson Distance):\n " % (person), recommendations_pearson
print "-"*50
def get(self):
user = self.user
user_id = user.get_id()
results = UserRating.query(UserRating.userid == str(user_id))
user_ratings = dict()
for ur in results:
user_ratings[ur.userid] = dict(zip(ur.movies, ur.ratings))
sim = item.itemMatch
recs = recommendations.getRecommendedItems(user_ratings, sim, str(user_id))
dictLs = []
keys = ['index', 'rating', 'movie']
for rec in recs:
ls = []
(rating, movie) = rec
ls.append(len(dictLs)+1)
ls.append(round(rating, 1))
ls.append(movie)
dictLs.append(dict(zip(keys, ls)))
logging.info(dictLs[0])
params = {
'recs': dictLs[:10],
'title': 'recommend',
'icon': 'icon-thumbs-up'
}
self.render_template('recommend.html', params)
def recommendByItems(self, event):
if not self.isValid(): # 保证输入框存在有效值
return
itemsim = recommendations.calculateSimilarItems(self.prefs, n = 50)
self.resultList = recommendations.getRecommendedItems(self.prefs, itemsim,
self.userId)[0:self.maxNum]
self.showResult()
def test_getRecommendedItems(self):
self.assertAlmostEqual(
recommendations.getRecommendedItems(critics, self.item_sim,
'Toby'),
[(3.182634730538922, 'The Night Listener'),
(2.5983318700614575, 'Just My Luck'),
(2.4730878186968837, 'Lady in the Water')], 5)
def recommendByItems(self, event):
if not self.isValid(): # 保证输入框存在有效值
return
itemsim = recommendations.calculateSimilarItems(self.prefs, n=50)
self.resultList = recommendations.getRecommendedItems(
self.prefs, itemsim, self.userId)[0:self.maxNum]
self.showResult()
def test_getRecommendedItems(self):
self.assertEqual(
recommendations.getRecommendedItems(
recommendations.critics,
recommendations.calculateSimilarItems(recommendations.critics),
'Toby'), [(3.182634730538922, 'The Night Listener'),
(2.5983318700614575, 'Just My Luck'),
(2.4730878186968837, 'Lady in the Water')])
def testMovieLens():
import recommendations
prefs = loadMovieLens()
userId = '87'
print prefs[userId]
start = time.clock()
print recommendations.getRecommendations(prefs, userId)[0:30]
print "Total Time:" + str(time.clock() - start)
print
print "Preprocessing..."
itemMatches = recommendations.calculateSimilarItems(prefs, n=50)
print
start = time.clock()
print recommendations.getRecommendedItems(prefs, itemMatches, userId)[0:30]
print "Total Time:" + str(time.clock() - start)
def getRecommendations(ratings, number):
# logging.info(ratings)
user_rating = dict()
for key in ratings:
user_rating[int(key)] = ratings[key]
user_ratings = {'current':user_rating}
rec_list = recommendations.getRecommendedItems(user_ratings, jokesim.sim_critics, 'current')
if len(rec_list) > number:
rec_list = rec_list[0:number]
# logging.info(rec_list)
rec_jokes = list()
for item in rec_list:
# rec_jokes[item[1]] = jokes.jokes[item[1]]
rec_jokes.append((item[1], jokes.jokes[item[1]], item[0]))
return rec_jokes
def get_recommendations(self, type='user_similarity', similarity='pearson'):
from recommendations import getRecommendations, calculateSimilarItems, get_similarity, getRecommendedItems
if type == 'user_similarity':
users_prefs = User.get_users_ratings_dict(lazy_evaluation=False)
sim = get_similarity(similarity)
return [
(star, Product.get_by_id(productid))
for star, productid in getRecommendations(users_prefs, self.id, similarity=sim)
]
elif type == 'item_similarity':
users_prefs = User.get_users_ratings_dict(lazy_evaluation=True)
similar_items = ItemSimilarity.get_similarity_dict()
return [
(star, Product.get_by_id(productid))
for star, productid in getRecommendedItems(users_prefs, similar_items, self.id)
]
else:
raise ValueError('Unknown recommendation type %s' % type)
# recommendationTest.py
import recommendations
itemsim = recommendations.calculateSimilarItems(recommendations.critics)
'''
recommendedList = recommendations.getRecommendedItems(recommendations.critics, itemsim, 'Toby')
print(recommendedList)
'''
prefs = recommendations.loadMovieLens()
# print prefs['87']
print 'user-based recommendations : '
recommendeduser = recommendations.getRecommendations(prefs, '87')[0:30]
print recommendeduser
print 'item-based recommendations : '
itemsim = recommendations.calculateSimilarItems(prefs, n=50)
recommendeditems = recommendations.getRecommendedItems(prefs, itemsim,
'87')[0:30]
print recommendeditems
import recommendations, jokesim, jokes
# The key of new user data must be in the following list, which repsents all the keys in the sim_critics
# [5, 7, 8, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150]
newusers = {
"user1":{
5: 4.5,
8: 7.0,
## add more testing ratings here
}
}
print len(recommendations.getRecommendedItems(newusers, jokesim.sim_critics, 'user1'))
print recommendations.getRecommendedItems(newusers, jokesim.sim_critics, 'user1')
print jokes.jokes[5]
import recommendations
import time;
def loadMovieLens(path='/Users/yupeng/Downloads/ml-100K'):
movies = {}
for line in open(path + '/u.item'):
(id, title) = line.split('|')[0:2]
movies[id] = title
prefs = {}
for line in open(path + '/u.data'):
(user, movieid, rating, ts) = line.split("\t")
prefs.setdefault(user, {})
prefs[user][movies[movieid]] = float(rating)
return prefs
prefs = loadMovieLens()
start = time.time()
result = recommendations.getRecommendations(prefs, '87')[0:30]
print result
print time.time() - start
start = time.time()
itemsim = recommendations.calculateSimilarItems(prefs, n=50)
print itemsim
print recommendations.getRecommendedItems(prefs, itemsim, '87')[0:30]
print time.time() - start
print sim_distance(critics, 'Lisa Rose', 'Gene Seymour') print 'Pearson Correlation Score of Lisa Rose and Gene Seymour is ' print sim_pearson(critics, 'Lisa Rose', 'Gene Seymour') print 'TopMatches 3 for Toby is ' print topMatches(critics, 'Toby', n=3) # User-Based CF print '通过按人群与 Toby 相似度，加权重新评分，为影片排名获得推荐: ' print getRecommendations(critics, 'Toby') print '通过查看哪些人喜欢 Superman Returns，以及这些人喜欢哪些其他物品来确定相似度：' movies = transformPrefs(critics) print topMatches(movies, 'Superman Returns') print '可能最喜欢 Just My Luck 的人群列表（对调人和物不一定能获得有用信息）：' print getRecommendations(movies, 'Just My Luck') # Item-Based CF print '构造物品比较数据集：' itemsim = calculateSimilarItems(critics) print itemsim print '基于物品的推荐为 Toby 提供推荐列表：' print getRecommendedItems(critics, itemsim, 'Toby')
def testBasics(self):
d = {'N': {'p': 1.0, 'j': 0.3}, 'Y':{'p': 0.8, 'j': 0.2, 'r':1.0} }
itemsim = recommendations.calculateSimilarItems(d)
r = recommendations.getRecommendedItems(d, itemsim, 'N')
self.assertEquals(1, len(r))
self.assertEquals('r', r[0][1])
import recommendations as rec
# 2.4.1.1
# transformation of the critics matrix to {"movie":{"Person1":"similarity", ...}}
transCritics = rec.transposeMatrix(rec.critics)
# 2.4.1.2
#
similar_items_euclidean = rec.calculateSimilarItems(transCritics, rec.sim_euclid_normed)
similar_items_pearson = rec.calculateSimilarItems(transCritics, rec.sim_pearson)
print rec.topMatches(transCritics, 'Lady in the Water', rec.sim_euclid)
# 2.4.1.3
print "Recommended Movies (euclidean): " + str(rec.getRecommendedItems(rec.critics, 'Toby Segaran', similar_items_euclidean))
print "Recommended Movies (pearson): " + str(rec.getRecommendedItems(rec.critics, 'Toby Segaran', similar_items_pearson))
from recommendations import critics import recommendations print critics['Lisa Rose']['Lady in the Water'] critics['Toby']['Snakes on a Plane'] = 4.5 print critics['Toby'] print critics print "-----------" print recommendations.getRecommendations(recommendations.critics, 'Toby') print "-----------" movies = recommendations.transformPrefs(recommendations.critics) print recommendations.topMatches(movies, 'Superman Returns') print "-----------" itemsim = recommendations.calculateSimilarItems(recommendations.critics) print itemsim print "-----------" print recommendations.getRecommendedItems(recommendations.critics, itemsim, 'Toby') print "-----------" prefs = recommendations.loadMovieLens() print prefs['1'] print "-----------" print recommendations.getRecommendations(prefs,'42')[31:60] print "-----------" ##itemsim = recommendations.calculateSimilarItems(prefs, n =50) ##print itemsim print "-----------" ##print recommendations.getRecommendedItems(prefs,itemsim, '87')[0:30] print "-----------" print recommendations.getKNNRecommendations(prefs,'42',30)[31:60]
import recommendations import datetime prefers = recommendations.loadMovieLens() startTime = datetime.datetime.now() itemMatch = recommendations.calculateSimilarItems(prefers, n = 50) print(recommendations.getRecommendedItems(prefers, itemMatch, "87")[0:30]) print((datetime.datetime.now() - startTime).total_seconds()) startTime = datetime.datetime.now() print(recommendations.getRecommendedItems(prefers, itemMatch, "88")[0:30]) print((datetime.datetime.now() - startTime).total_seconds())
created by gjwei on 2017/10/11
"""
import pydelicious
from deliciouserc import initializeUserDict
from recommendations import get_recommmendations, getRecommendedItems, calculate_similarity_items
def load_movielens(path='./data/movielens'):
movies = {}
for line in open(path + '/u.item'):
(id, title) = line.split('|')[0:2]
movies[id] = title
prefs = {}
for line in open(path + '/u.data'):
(user, movieid, rating, ts) = line.split('\t')
prefs.setdefault(user, {})
prefs[user][movies[movieid]] = float(rating)
return prefs
if __name__ == '__main__':
prefs = load_movielens()
print(prefs['87'])
print(get_recommmendations(prefs, '87'))[:3]
print(getRecommendedItems(prefs, calculate_similarity_items(prefs),
'87'))[:5]
movies = recommendations.transformPrefs(recommendations.critics)
print '调换人与物品：'
print movies
print '打印相似类型的电影：'
print recommendations.topMatches(movies, 'Superman Returns')
print recommendations.getRecommendations(movies, 'Just My Luck')
# import pydelicious
# print pydelicious.get_popular(tag='programming')
# print '开始填充数据：'
# from deliciousrec import *
# delusers = initializeUserDict('programming')
# delusers['tsegaran'] ={}
# fillItems(delusers)
# print delusers
itemSim = recommendations.calculateSimilarItems(recommendations.critics)
print itemSim
print '基于物品推荐：'
print recommendations.getRecommendedItems(recommendations.critics, itemSim,
'Toby')
print '电影数据：'
prefs = recommendations.loadMovieLens()
print prefs['87']
print recommendations.getRecommendations(prefs, '87')[0:30]
itemSim2 = recommendations.calculateSimilarItems(prefs, n=50)
print recommendations.getRecommendedItems(prefs, itemSim2, '87')[0:30]
#Get recommendations of movies using Euclidian
print(
recommendations.getRecommendations(
recommendations.critics,
'Toby',
similarity=recommendations.sim_distance))
# Transpose the matrix to movies vs user
movies = recommendations.transformPrefs(recommendations.critics)
#Print similar movies
print(recommendations.topMatches(movies, 'Superman Returns'))
#Similarity matirx of items
itemsim = recommendations.calculateSimilarItems(recommendations.critics)
print(itemsim)
#Get recommendations based on similarity matrix
print(
recommendations.getRecommendedItems(recommendations.critics, itemsim,
'Toby'))
# print(recommendations.sim_pearson(critics,'Lisa Rose','Gene Seymour'))
# print(recommendations.getRecommendations(recommendations.critics,'Toby'))
# delusers=deliciousrec.initializeUserDict('programming')
# itemsim=recommendations.calculateSimilarItems(recommendations.critics)
#
# print(itemsim)
def test_getRecommendedItems(self):
self.assertEqual(recommendations.getRecommendedItems(recommendations.critics, recommendations.calculateSimilarItems(recommendations.critics),'Toby'),
[(3.182634730538922, 'The Night Listener'),(2.5983318700614575, 'Just My Luck'),(2.4730878186968837, 'Lady in the Water')])
print recommendations.topMatches(recommendations.critics,'Toby',n=3,similarity=recommendations.sim_distance) recommendations.getRecommendations(recommendations.critics,'Toby') recommendations.getRecommendations(recommendations.critics,'Toby',similarity=recommendations.sim_distance) movies=recommendations.transformPrefs(recommendations.critics) #print movies recommendations.topMatches(movies,'Superman Returns') recommendations.getRecommendations(movies,'Just My Luck') recommendations.getRecommendations(movies,'Lady in the Water') itemsim=recommendations.calculateSimilarItems(recommendations.critics,n=8) itemsim recommendations.getRecommendedItems(recommendations.critics,itemsim,'Toby') prefs=recommendations.loadMovieLens() prefs['87'] recommendations.getRecommendations(prefs,'87')[0:30] itemsim1=recommendations.calculateSimilarItems(prefs,n=50) recommendations.getRecommendedItems(prefs,itemsim1,'87')[0:30]
def test_getRecommendedItems(self):
self.assertAlmostEqual(recommendations.getRecommendedItems(critics, self.item_sim, 'Toby'),
[(3.182634730538922, 'The Night Listener'),
(2.5983318700614575, 'Just My Luck'),
(2.4730878186968837, 'Lady in the Water')], 5)
