-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathCentralityMeasures.py
183 lines (152 loc) · 8.69 KB
/
CentralityMeasures.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
from pyspark.sql.functions import col, when, explode, struct, array, lit, array_remove,array_intersect
class CentralityMeasures:
def __init__(self, edges_df, user_features_df, categories, sparkSession, neo4jDrvier, method_name):
self.edges_df = edges_df
self.user_features_df = user_features_df
self.list_categories = categories
self.sparkSession = sparkSession
self.neo4jDriver = neo4jDrvier
self.method_name = method_name
self.nodes_df = None
self.results_df = None
self.centrality_categories = []
###(RUN THIS METHOD ONCE)###
def insertDataToGraph(self):
spark = self.sparkSession
neo4j = self.neo4jDriver.session()
sc = spark.sparkContext
feats = self.user_features_df
list_cat=self.list_categories
cat_count=len(list_cat)
#import edges
e = self.edges_df
self.nodes_df = e.select("Source").union(e.select("Target")).distinct().withColumnRenamed('Source', 'id')
n = self.nodes_df
print(feats.count(), list_cat, e.count(), n.count())
feats.printSchema()
#cache dataframes
feats.cache()
e.cache()
n.cache()
#add category property to u
u_focus_rate=feats.select(col('id'),col("user_features{}.dict_focus_rate".format(("_"+self.method_name) if len(self.method_name)>0 else "")).alias("dict_focus_rate"))
u_with_category=u_focus_rate.withColumn("userCategory",array([lit(c) for c in list_cat]))
for cat in list_cat:
u_with_category=u_with_category.withColumn("temp", when(col("dict_focus_rate.{}".format(cat))<1/cat_count, array_remove(u_with_category["userCategory"], cat)).otherwise(u_with_category["userCategory"])).drop("userCategory").withColumnRenamed("temp","userCategory")
u_with_category=u_with_category.select("id","userCategory")
#join n and u_with_category
n_with_category=n.join(u_with_category,"id",how="left")
#add category columns to e
e_with_category=e.join(n_with_category,e.Source==n_with_category.id,how="left").withColumnRenamed("userCategory","sourceCategory").select("Source","Target","sourceCategory")
e_with_category=e_with_category.join(n_with_category,e_with_category.Target==n_with_category.id,how="left").withColumnRenamed("userCategory","targetCategory").select("Source","Target","sourceCategory","targetCategory")
#determine intersection between sourceCategory and targetCategory
e_with_category=e_with_category.withColumn("Categories", array_intersect(e_with_category["sourceCategory"],e_with_category["targetCategory"]))
#flatten out categories of edges
e_with_category=e_with_category.withColumn("Category",explode(col("Categories"))).select("Source","Target","Category")
print("e_with_category", e_with_category.count())
e_with_category.printSchema()
## Insert data
insert_query = '''
UNWIND {triples} as triple
MERGE (p1:User {id:triple[0]})
MERGE (p2:User {id:triple[1]}) WITH p1,p2,triple
CALL apoc.create.relationship(p1, triple[2], {}, p2) YIELD rel
RETURN *
'''
e_listoftriples=e_with_category.toPandas()[['Source','Target','Category']].values.tolist()
print("e_listoftriples:", len(e_listoftriples))
batches = list(self.generate_batches(e_listoftriples, 7000))
for batch in batches:
neo4j.run(insert_query, parameters={"triples": batch})
e_with_category.show()
print("batches size:", len(batches), " last batch:", len(batches[-1]))
#clear cache
# spark.catalog.clearCache()
def startCalculation(self):
spark = self.sparkSession
neo4j = self.neo4jDriver.session()
sc = spark.sparkContext
n = self.nodes_df
#cache dataframes
n.cache()
#get relationship list (some categories would be empty)
rel_query = '''
MATCH ()-[relationship]->()
RETURN DISTINCT TYPE(relationship) AS type
ORDER BY type ASC;
'''
results = neo4j.run(rel_query)
temp_df=spark.createDataFrame(results)
list_cat_updated = temp_df.select("_1").rdd.flatMap(lambda x: x).collect()
self.centrality_categories = list_cat_updated
#Degree Centrality
for cat in list_cat_updated:
degreeCentrality_query = '''
CALL algo.degree.stream("User", "%s", {write: true, direction: "both"})
YIELD nodeId, score as degreeCentrality
RETURN algo.getNodeById(nodeId).id as userId, degreeCentrality
ORDER BY userId DESC
'''%(cat)
results = neo4j.run(degreeCentrality_query)
temp_df=spark.createDataFrame(results).withColumnRenamed("_1", "userId").withColumnRenamed("_2", "degreeCentrality")
n=n.join(temp_df, n.id==temp_df.userId, how="left").withColumnRenamed("degreeCentrality",cat)
n=n.select("id", struct(list_cat_updated).alias("degreeCentrality"))
#Closeness Centrality
for cat in list_cat_updated:
closenessCentrality_query = '''
CALL algo.closeness.stream("User", "%s",{write: true})
YIELD nodeId, centrality as closenessCentrality
RETURN algo.getNodeById(nodeId).id as userId, closenessCentrality
ORDER BY userId DESC
'''%(cat)
results = neo4j.run(closenessCentrality_query)
temp_df=spark.createDataFrame(results).withColumnRenamed("_1", "userId").withColumnRenamed("_2", "closenessCentrality")
n=n.join(temp_df, n.id==temp_df.userId, how="left").withColumnRenamed("closenessCentrality",cat)
n=n.select("id", "degreeCentrality",struct(list_cat_updated).alias("closenessCentrality"))
#Betweenness Centrality
for cat in list_cat_updated:
betweennessCentrality_query = '''
CALL algo.betweenness.stream("User", "%s",{write: true})
YIELD nodeId, centrality as betweennessCentrality
RETURN algo.getNodeById(nodeId).id as userId, betweennessCentrality
ORDER BY userId DESC
'''%(cat)
#(USE THIS FOR LARGER GRAPHS)
#Approximation of Betweenness Centrality(RA-Brandes algorithm)
#aproxBetweennessCentrality_query = '''
#CALL algo.betweenness.sampled.stream("User", "%s", {strategy:"degree"})
#YIELD nodeId, centrality as aproxBetweennessCentrality
#RETURN algo.getNodeById(nodeId).id AS userId, aproxBetweennessCentrality
#ORDER BY userId DESC
#'''%(cat)
results = neo4j.run(betweennessCentrality_query)
temp_df=spark.createDataFrame(results).withColumnRenamed("_1", "userId").withColumnRenamed("_2", "betweennessCentrality")
n=n.join(temp_df, n.id==temp_df.userId, how="left").withColumnRenamed("betweennessCentrality",cat)
n=n.select("id", "degreeCentrality","closenessCentrality",struct(list_cat_updated).alias("betweennessCentrality"))
#PageRank
for cat in list_cat_updated:
pageRank_query = '''
CALL algo.pageRank.stream('User', '%s', {iterations:100, dampingFactor:0.85, write: true})
YIELD nodeId, score as pageRank
RETURN algo.getNodeById(nodeId).id AS userId, pageRank
ORDER BY userId DESC
'''%(cat)
results = neo4j.run(pageRank_query)
temp_df=spark.createDataFrame(results).withColumnRenamed("_1", "userId").withColumnRenamed("_2", "pageRank")
n=n.join(temp_df, n.id==temp_df.userId, how="left").withColumnRenamed("pageRank",cat)
n=n.select("id", "degreeCentrality","closenessCentrality","betweennessCentrality", struct(list_cat_updated).alias("pageRank"))
#set results
self.results_df = n.select("id",struct(n.columns[1:]).alias("centralities{}".format(("_"+self.method_name) if len(self.method_name)>0 else "")))
#clear cache
# spark.catalog.clearCache()
def generate_batches(self, l, n):
# For item i in a range that is a length of l,
for i in range(0, len(l), n):
# Create an index range for l of n items:
yield l[i:i+n]
def getResults(self):
return self.results_df
def getCentralitiesTopics(self):
return ['betweennessCentrality','closenessCentrality','degreeCentrality','pageRank']
def getCentralityCategories(self):
return self.centrality_categories