Merge pull request #16 from lijunweiyhn/master

DiffNet++ pull request

Author: PeiJieSun

.DS_Store

@@ -0,0 +1 @@
+python: 2.7

Diffnet++/.DS_Store

@@ -0,0 +1,33 @@
+import os
+from time import time
+from DataModule import DataModule
+
+class DataUtil():
+ def __init__(self, conf):
+ self.conf = conf
+ #print('DataUtil, Line12, test- conf data_dir:%s' % self.conf.data_dir)
+
+ def initializeRankingHandle(self):
+ #t0 = time()
+ self.createTrainHandle()
+ self.createEvaluateHandle()
+ #t1 = time()
+ #print('Prepare data cost:%.4fs' % (t1 - t0))
+ 
+ def createTrainHandle(self):
+ data_dir = self.conf.data_dir
+ train_filename = "%s/%s.train.rating" % (data_dir, self.conf.data_name)
+ val_filename = "%s/%s.val.rating" % (data_dir, self.conf.data_name)
+ test_filename = "%s/%s.test.rating" % (data_dir, self.conf.data_name)
+
+ self.train = DataModule(self.conf, train_filename)
+ self.val = DataModule(self.conf, val_filename)
+ self.test = DataModule(self.conf, test_filename)
+
+ def createEvaluateHandle(self):
+ data_dir = self.conf.data_dir
+ val_filename = "%s/%s.val.rating" % (data_dir, self.conf.data_name)
+ test_filename = "%s/%s.test.rating" % (data_dir, self.conf.data_name)
+
+ self.val_eva = DataModule(self.conf, val_filename)
+ self.test_eva = DataModule(self.conf, test_filename)

Diffnet++/.travis.yml

@@ -0,0 +1,323 @@
+import math
+import numpy as np
+
+
+class Evaluate():
+ def __init__(self, conf):
+ self.conf = conf
+
+ def getIdcg(self, length):
+ idcg = 0.0
+ for i in range(length):
+ idcg = idcg + math.log(2) / math.log(i + 2)
+ return idcg
+
+ def getDcg(self, value):
+ dcg = math.log(2) / math.log(value + 2)
+ return dcg
+
+ def getHr(self, value):
+ hit = 1.0
+ return hit
+
+ def evaluateRankingPerformance(self, evaluate_index_dict, evaluate_real_rating_matrix, \
+ evaluate_predict_rating_matrix, topK, num_procs, exp_flag=0, sp_name=None, result_file=None):
+ user_list = list(evaluate_index_dict.keys())
+ batch_size = len(user_list) / num_procs
+
+ hr_list, ndcg_list = [], []
+ index = 0
+ for _ in range(num_procs):
+ if index + batch_size < len(user_list):
+ batch_user_list = user_list[index:index+batch_size]
+ index = index + batch_size
+ else:
+ batch_user_list = user_list[index:len(user_list)]
+ tmp_hr_list, tmp_ndcg_list = self.getHrNdcgProc(evaluate_index_dict, evaluate_real_rating_matrix, \
+ evaluate_predict_rating_matrix, topK, batch_user_list)
+ hr_list.extend(tmp_hr_list)
+ ndcg_list.extend(tmp_ndcg_list)
+ return np.mean(hr_list), np.mean(ndcg_list)
+
+ def getHrNdcgProc(self, 
+ evaluate_index_dict, 
+ evaluate_real_rating_matrix,
+ evaluate_predict_rating_matrix, 
+ topK, 
+ user_list):
+
+ tmp_hr_list, tmp_ndcg_list = [], []
+
+ for u in user_list:
+ real_item_index_list = evaluate_index_dict[u]
+ real_item_rating_list = list(np.concatenate(evaluate_real_rating_matrix[real_item_index_list]))
+ positive_length = len(real_item_rating_list)
+ target_length = min(positive_length, topK)
+ 
+ predict_rating_list = evaluate_predict_rating_matrix[u]
+ real_item_rating_list.extend(predict_rating_list)
+ sort_index = np.argsort(real_item_rating_list)
+ sort_index = sort_index[::-1]
+
+ user_hr_list = []
+ user_ndcg_list = []
+ hits_num = 0
+ for idx in range(topK):
+ ranking = sort_index[idx]
+ if ranking < positive_length:
+ hits_num += 1
+ user_hr_list.append(self.getHr(idx))
+ user_ndcg_list.append(self.getDcg(idx))
+
+ idcg = self.getIdcg(target_length)
+
+ tmp_hr = np.sum(user_hr_list) / target_length
+ tmp_ndcg = np.sum(user_ndcg_list) / idcg
+ tmp_hr_list.append(tmp_hr)
+ tmp_ndcg_list.append(tmp_ndcg)
+
+ return tmp_hr_list, tmp_ndcg_list
+
+ def evaluateRankingPerformance_sparsity(self, evaluate_index_dict, social_sparsity_dict, interest_sparsity_dict, evaluate_real_rating_matrix, \
+ evaluate_predict_rating_matrix, topK, num_procs, exp_flag=0, sp_name=None, result_file=None):
+ user_list = list(evaluate_index_dict.keys())
+ batch_size = len(user_list) / num_procs
+
+ social_hr_list_0_4, social_ndcg_list_0_4 = [], []
+ social_hr_list_4_8, social_ndcg_list_4_8 = [], []
+ social_hr_list_8_16, social_ndcg_list_8_16 = [], []
+ social_hr_list_16_32, social_ndcg_list_16_32 = [], []
+ social_hr_list_32_64, social_ndcg_list_32_64 = [], []
+ social_hr_list_64, social_ndcg_list_64 = [], []
+ 
+
+
+ interest_hr_list_0_4, interest_ndcg_list_0_4 = [], []
+ interest_hr_list_4_8, interest_ndcg_list_4_8 = [], []
+ interest_hr_list_8_16, interest_ndcg_list_8_16 = [], []
+ interest_hr_list_16_32, interest_ndcg_list_16_32 = [], []
+ interest_hr_list_32_64, interest_ndcg_list_32_64 = [], []
+ interest_hr_list_64, interest_ndcg_list_64 = [], []
+
+
+
+ index = 0
+ for _ in range(num_procs):
+ if index + batch_size < len(user_list):
+ batch_user_list = user_list[index:index+batch_size]
+ index = index + batch_size
+ else:
+ batch_user_list = user_list[index:len(user_list)]
+ #set_trace()
+ #social
+ social_tmp_hr_list_0_4, social_tmp_ndcg_list_0_4,\
+ social_tmp_hr_list_4_8, social_tmp_ndcg_list_4_8,\
+ social_tmp_hr_list_8_16, social_tmp_ndcg_list_8_16,\
+ social_tmp_hr_list_16_32, social_tmp_ndcg_list_16_32,\
+ social_tmp_hr_list_32_64, social_tmp_ndcg_list_32_64, \
+ social_tmp_hr_list_64, social_tmp_ndcg_list_64 = self.getHrNdcgProc_social_sparsity(evaluate_index_dict, social_sparsity_dict, interest_sparsity_dict, evaluate_real_rating_matrix, \
+ evaluate_predict_rating_matrix, topK, batch_user_list)
+
+ social_hr_list_0_4.extend(social_tmp_hr_list_0_4)
+ social_ndcg_list_0_4.extend(social_tmp_ndcg_list_0_4)
+ social_hr_list_4_8.extend(social_tmp_hr_list_4_8)
+ social_ndcg_list_4_8.extend(social_tmp_ndcg_list_4_8)
+ social_hr_list_8_16.extend(social_tmp_hr_list_8_16)
+ social_ndcg_list_8_16.extend(social_tmp_ndcg_list_8_16)
+ social_hr_list_16_32.extend(social_tmp_hr_list_16_32)
+ social_ndcg_list_16_32.extend(social_tmp_ndcg_list_16_32)
+ social_hr_list_32_64.extend(social_tmp_hr_list_32_64)
+ social_ndcg_list_32_64.extend(social_tmp_ndcg_list_32_64)
+ social_hr_list_64.extend(social_tmp_hr_list_64)
+ social_ndcg_list_64.extend(social_tmp_ndcg_list_64)
+
+
+ #interest
+ interest_tmp_hr_list_0_4, interest_tmp_ndcg_list_0_4,\
+ interest_tmp_hr_list_4_8, interest_tmp_ndcg_list_4_8,\
+ interest_tmp_hr_list_8_16, interest_tmp_ndcg_list_8_16,\
+ interest_tmp_hr_list_16_32, interest_tmp_ndcg_list_16_32,\
+ interest_tmp_hr_list_32_64, interest_tmp_ndcg_list_32_64, \
+ interest_tmp_hr_list_64, interest_tmp_ndcg_list_64 = self.getHrNdcgProc_interest_sparsity(evaluate_index_dict, social_sparsity_dict, interest_sparsity_dict, evaluate_real_rating_matrix, \
+ evaluate_predict_rating_matrix, topK, batch_user_list)
+
+ interest_hr_list_0_4.extend(interest_tmp_hr_list_0_4)
+ interest_ndcg_list_0_4.extend(interest_tmp_ndcg_list_0_4)
+ interest_hr_list_4_8.extend(interest_tmp_hr_list_4_8)
+ interest_ndcg_list_4_8.extend(interest_tmp_ndcg_list_4_8)
+ interest_hr_list_8_16.extend(interest_tmp_hr_list_8_16)
+ interest_ndcg_list_8_16.extend(interest_tmp_ndcg_list_8_16)
+ interest_hr_list_16_32.extend(interest_tmp_hr_list_16_32)
+ interest_ndcg_list_16_32.extend(interest_tmp_ndcg_list_16_32)
+ interest_hr_list_32_64.extend(interest_tmp_hr_list_32_64)
+ interest_ndcg_list_32_64.extend(interest_tmp_ndcg_list_32_64)
+ interest_hr_list_64.extend(interest_tmp_hr_list_64)
+ interest_ndcg_list_64.extend(interest_tmp_ndcg_list_64)
+
+
+ #set_trace()
+ return np.sum(social_hr_list_0_4)/len(social_sparsity_dict['0-4']), np.sum(social_ndcg_list_0_4)/len(social_sparsity_dict['0-4']),\
+ np.sum(social_hr_list_4_8)/len(social_sparsity_dict['4-8']), np.sum(social_ndcg_list_4_8)/len(social_sparsity_dict['4-8']),\
+ np.sum(social_hr_list_8_16)/len(social_sparsity_dict['8-16']), np.sum(social_ndcg_list_8_16)/len(social_sparsity_dict['8-16']),\
+ np.sum(social_hr_list_16_32)/len(social_sparsity_dict['16-32']), np.sum(social_ndcg_list_16_32)/len(social_sparsity_dict['16-32']),\
+ np.sum(social_hr_list_32_64)/len(social_sparsity_dict['32-64']), np.sum(social_ndcg_list_32_64)/len(social_sparsity_dict['32-64']), \
+ np.sum(social_hr_list_64)/len(social_sparsity_dict['64-']), np.sum(social_ndcg_list_64)/len(social_sparsity_dict['64-']), \
+ np.sum(interest_hr_list_0_4)/len(interest_sparsity_dict['0-4']), np.sum(interest_ndcg_list_0_4)/len(interest_sparsity_dict['0-4']),\
+ np.sum(interest_hr_list_4_8)/len(interest_sparsity_dict['4-8']), np.sum(interest_ndcg_list_4_8)/len(interest_sparsity_dict['4-8']),\
+ np.sum(interest_hr_list_8_16)/len(interest_sparsity_dict['8-16']), np.sum(interest_ndcg_list_8_16)/len(interest_sparsity_dict['8-16']),\
+ np.sum(interest_hr_list_16_32)/len(interest_sparsity_dict['16-32']), np.sum(interest_ndcg_list_16_32)/len(interest_sparsity_dict['16-32']),\
+ np.sum(interest_hr_list_32_64)/len(interest_sparsity_dict['32-64']), np.sum(interest_ndcg_list_32_64)/len(interest_sparsity_dict['32-64']), \
+ np.sum(interest_hr_list_64)/len(interest_sparsity_dict['64-']), np.sum(interest_ndcg_list_64)/len(interest_sparsity_dict['64-'])
+
+
+ def getHrNdcgProc_social_sparsity(self, 
+ evaluate_index_dict, 
+ social_sparsity_dict,
+ interest_sparsity_dict,
+ evaluate_real_rating_matrix,
+ evaluate_predict_rating_matrix, 
+ topK, 
+ user_list):
+
+ social_tmp_hr_list_0_4, social_tmp_ndcg_list_0_4= [], []
+ social_tmp_hr_list_4_8, social_tmp_ndcg_list_4_8= [], []
+ social_tmp_hr_list_8_16, social_tmp_ndcg_list_8_16= [], []
+ social_tmp_hr_list_16_32, social_tmp_ndcg_list_16_32= [], []
+ social_tmp_hr_list_32_64, social_tmp_ndcg_list_32_64= [], []
+ social_tmp_hr_list_64, social_tmp_ndcg_list_64= [], []
+ 
+
+ for u in user_list:
+ real_item_index_list = evaluate_index_dict[u]
+ real_item_rating_list = list(np.concatenate(evaluate_real_rating_matrix[real_item_index_list]))
+ positive_length = len(real_item_rating_list)
+ target_length = min(positive_length, topK)
+ 
+ predict_rating_list = evaluate_predict_rating_matrix[u]
+ real_item_rating_list.extend(predict_rating_list)
+ sort_index = np.argsort(real_item_rating_list)
+ sort_index = sort_index[::-1]
+
+ user_hr_list = []
+ user_ndcg_list = []
+ hits_num = 0
+ for idx in range(topK):
+ ranking = sort_index[idx]
+ if ranking < positive_length:
+ hits_num += 1
+ user_hr_list.append(self.getHr(idx))
+ user_ndcg_list.append(self.getDcg(idx))
+
+ idcg = self.getIdcg(target_length)
+
+ tmp_hr = np.sum(user_hr_list) / target_length
+ tmp_ndcg = np.sum(user_ndcg_list) / idcg
+ #set_trace()
+ if( u in social_sparsity_dict['64-'] ):
+ social_tmp_hr_list_64.append(tmp_hr)
+ social_tmp_ndcg_list_64.append(tmp_ndcg) 
+ elif( u in social_sparsity_dict['32-64'] ):
+ social_tmp_hr_list_32_64.append(tmp_hr)
+ social_tmp_ndcg_list_32_64.append(tmp_ndcg)
+ elif( u in social_sparsity_dict['16-32'] ):
+ social_tmp_hr_list_16_32.append(tmp_hr)
+ social_tmp_ndcg_list_16_32.append(tmp_ndcg) 
+ elif( u in social_sparsity_dict['8-16'] ):
+ social_tmp_hr_list_8_16.append(tmp_hr)
+ social_tmp_ndcg_list_8_16.append(tmp_ndcg) 
+ elif( u in social_sparsity_dict['4-8'] ):
+ social_tmp_hr_list_4_8.append(tmp_hr)
+ social_tmp_ndcg_list_4_8.append(tmp_ndcg) 
+ elif( u in social_sparsity_dict['0-4'] ):
+ social_tmp_hr_list_0_4.append(tmp_hr)
+ social_tmp_ndcg_list_0_4.append(tmp_ndcg) 
+
+
+ return social_tmp_hr_list_0_4, social_tmp_ndcg_list_0_4, \
+ social_tmp_hr_list_4_8, social_tmp_ndcg_list_4_8, \
+ social_tmp_hr_list_8_16, social_tmp_ndcg_list_8_16, \
+ social_tmp_hr_list_16_32, social_tmp_ndcg_list_16_32, \
+ social_tmp_hr_list_32_64, social_tmp_ndcg_list_32_64, \
+ social_tmp_hr_list_64, social_tmp_ndcg_list_64
+ 
+
+
+
+ def getHrNdcgProc_interest_sparsity(self, 
+ evaluate_index_dict, 
+ social_sparsity_dict,
+ interest_sparsity_dict,
+ evaluate_real_rating_matrix,
+ evaluate_predict_rating_matrix, 
+ topK, 
+ user_list):
+
+ interest_tmp_hr_list_0_4, interest_tmp_ndcg_list_0_4= [], []
+ interest_tmp_hr_list_4_8, interest_tmp_ndcg_list_4_8= [], []
+ interest_tmp_hr_list_8_16, interest_tmp_ndcg_list_8_16= [], []
+ interest_tmp_hr_list_16_32, interest_tmp_ndcg_list_16_32= [], []
+ interest_tmp_hr_list_32_64, interest_tmp_ndcg_list_32_64= [], []
+ interest_tmp_hr_list_64, interest_tmp_ndcg_list_64= [], []
+
+
+ for u in user_list:
+ real_item_index_list = evaluate_index_dict[u]
+ real_item_rating_list = list(np.concatenate(evaluate_real_rating_matrix[real_item_index_list]))
+ positive_length = len(real_item_rating_list)
+ target_length = min(positive_length, topK)
+ 
+ predict_rating_list = evaluate_predict_rating_matrix[u]
+ real_item_rating_list.extend(predict_rating_list)
+ sort_index = np.argsort(real_item_rating_list)
+ sort_index = sort_index[::-1]
+
+ user_hr_list = []
+ user_ndcg_list = []
+ hits_num = 0
+ for idx in range(topK):
+ ranking = sort_index[idx]
+ if ranking < positive_length:
+ hits_num += 1
+ user_hr_list.append(self.getHr(idx))
+ user_ndcg_list.append(self.getDcg(idx))
+
+ idcg = self.getIdcg(target_length)
+
+ tmp_hr = np.sum(user_hr_list) / target_length
+ tmp_ndcg = np.sum(user_ndcg_list) / idcg
+ #set_trace()
+ if( u in interest_sparsity_dict['64-'] ):
+ interest_tmp_hr_list_64.append(tmp_hr)
+ interest_tmp_ndcg_list_64.append(tmp_ndcg) 
+ elif( u in interest_sparsity_dict['32-64']):
+ interest_tmp_hr_list_32_64.append(tmp_hr)
+ interest_tmp_ndcg_list_32_64.append(tmp_ndcg) 
+ elif( u in interest_sparsity_dict['16-32']):
+ interest_tmp_hr_list_16_32.append(tmp_hr)
+ interest_tmp_ndcg_list_16_32.append(tmp_ndcg) 
+ elif( u in interest_sparsity_dict['8-16']):
+ interest_tmp_hr_list_8_16.append(tmp_hr)
+ interest_tmp_ndcg_list_8_16.append(tmp_ndcg) 
+ elif( u in interest_sparsity_dict['4-8'] ):
+ interest_tmp_hr_list_4_8.append(tmp_hr)
+ interest_tmp_ndcg_list_4_8.append(tmp_ndcg) 
+ elif( u in interest_sparsity_dict['0-4']):
+ interest_tmp_hr_list_0_4.append(tmp_hr)
+ interest_tmp_ndcg_list_0_4.append(tmp_ndcg)
+
+ return interest_tmp_hr_list_0_4, interest_tmp_ndcg_list_0_4, \
+ interest_tmp_hr_list_4_8, interest_tmp_ndcg_list_4_8, \
+ interest_tmp_hr_list_8_16, interest_tmp_ndcg_list_8_16, \
+ interest_tmp_hr_list_16_32, interest_tmp_ndcg_list_16_32, \
+ interest_tmp_hr_list_32_64, interest_tmp_ndcg_list_32_64, \
+ interest_tmp_hr_list_64, interest_tmp_ndcg_list_64
+ 
+
+
+
+
+
+
+
+
+

Diffnet++/class/DataModule.py

@@ -0,0 +1,13 @@
+import os, shutil
+import ConfigParser as cp
+
+class Logging():
+ def __init__(self, filename):
+ self.filename = filename
+ 
+ def record(self, str_log):
+ filename = self.filename
+ print(str_log)
+ with open(filename, 'a') as f:
+ f.write("%s\r\n" % str_log)
+ f.flush()