-
Notifications
You must be signed in to change notification settings - Fork 1
/
RunCalculators.py
208 lines (184 loc) · 11.8 KB
/
RunCalculators.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
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
from CentralityMeasures import CentralityMeasures
from UserFeatures import UserFeatures
from pyspark import SparkContext
from pyspark import SparkConf
from pyspark.sql.session import SparkSession
from pyspark.sql.functions import col, udf, struct
from pyspark.sql.types import DoubleType
from pyspark.ml.feature import MinMaxScaler, StandardScaler, VectorAssembler
from pyspark.ml import Pipeline
from neo4j import GraphDatabase, basic_auth
from pyspark.sql.utils import AnalysisException
class RunCalculators:
def __init__(self, mongo_port, databaseName, tweetCollection, userCollection, edgesCollection, outputFile, outputCollection, categoryColumnName, method_name):
self.mongo_port = mongo_port
self.databaseName = databaseName
self.tweetCollection = tweetCollection
self.userCollection = userCollection
self.edgesCollection = edgesCollection
self.outputFile = outputFile
self.outputCollection = outputCollection
self.categoryColumnName = categoryColumnName
self.method_name = method_name
self.categories = []
self.user_features_topics = []
self.centralities_topics = []
self.centrality_categories = []
self.userFeatures_df = None
self.centralities_df = None
self.merged_df = None
self.normalized_df = None
self.sparkSession = None
self.neo4jDriver = None
self.user_df = None
self.tweet_df = None
self.edges_df = None
self.spammers_df = None
def main(self):
#set Spark Session
self.setSpark()
#connect Neo4j Session
self.setNeo4j()
#connect MongoDB and set DataFrames
self.setDataFrames(self.mongo_port, self.databaseName, self.tweetCollection, self.userCollection, self.edgesCollection, self.categoryColumnName)
#exclude spammer users
print("before spam module count of user_df:{}".format(self.user_df.count()))
self.excludeSpammers(self.user_df, self.tweet_df, self.edges_df)
print("after spam module:{}".format(self.user_df.count()))
#run user feature calculator
myUserFeatures = UserFeatures(self.user_df, self.tweet_df, self.sparkSession, self.method_name)
myUserFeatures.startCalculation()
self.userFeatures_df = myUserFeatures.getResults()
self.categories = myUserFeatures.getCategories()
self.user_features_topics = myUserFeatures.getUserFeaturesTopics()
print("after UserFeatures module count of userFeatures_df:{}".format(self.userFeatures_df.count()))
self.userFeatures_df.printSchema()
print("categories:",self.categories)
#run centralities calculator
myCentralities = CentralityMeasures(self.edges_df, self.userFeatures_df, self.categories, self.sparkSession, self.neo4jDriver, self.method_name)
myCentralities.insertDataToGraph() #RUN THIS METHOD ONLY ONCE FOR EVERY GRAPH
myCentralities.startCalculation()
self.centralities_df = myCentralities.getResults()
self.centrality_categories = myCentralities.getCentralityCategories()
self.centralities_topics = myCentralities.getCentralitiesTopics()
print("centralities result: ", self.centralities_df.count())
self.centralities_df.printSchema()
#merge, normalize and export dataframes
self.mergeDataFrames(self.userFeatures_df, self.centralities_df, self.method_name)
print("after merge: ", self.merged_df.count())
self.merged_df.printSchema()
self.normalizeDataFrame(self.merged_df, self.user_features_topics, self.centralities_topics, self.categories, self.centrality_categories, self.method_name)
print("before exporting normalized_df:", self.normalized_df.count())
self.normalized_df.printSchema()
self.exportDataFrame(self.normalized_df, self.outputFile)
#save results to database
self.saveResultsToDB(self.normalized_df, self.user_df, self.mongo_port, self.databaseName, self.outputCollection)
#stop spark and neo4j
self.neo4jDriver.session().close()
self.sparkSession.stop()
def setSpark(self):
#spark configurations
conf=SparkConf()
conf.setMaster("local[4]") # if you got java memory error, you can reduce running cores
# conf.set("spark.executor.memory", "32g") # this doesn't important when you run on local
conf.set("spark.driver.memory", "20g")
conf.set("spark.dirver.maxResultSize", "6g")
conf.set("spark.memory.offHeap.size","18g") # you can increase this to overcome java memory issues
# conf.set("spark.executor.extraJavaOptions", "-Xmx1024m")
conf.set("spark.executor.extraJavaOptions","-XX:+UseCompressedOops")
# conf.set("spark.cores.max", "2")
# conf.set("spark.driver.extraClassPath",
# driver_home+'/jdbc/postgresql-9.4-1201-jdbc41.jar:'\
# +driver_home+'/jdbc/clickhouse-jdbc-0.1.52.jar:'\
# +driver_home+'/mongo/mongo-spark-connector_2.11-2.2.3.jar:'\
# +driver_home+'/mongo/mongo-java-driver-3.8.0.jar')
sc = SparkContext.getOrCreate(conf)
self.sparkSession = SparkSession(sc)
def setNeo4j(self):
driver = GraphDatabase.driver("bolt://localhost", auth=basic_auth("neo4j", "12345678"))
self.neo4jDriver = driver
def setDataFrames(self, mongo_port, db, tweetColl, userColl, edgesColl, categoryColumnName):
spark = self.sparkSession
temp_u_df = spark.read.format("mongo").option("uri","mongodb://127.0.0.1:{}/{}.{}".format(mongo_port,db, userColl)).option('sampleSize', 40000).load()
#We don't like "id" column :)
if "id" in temp_u_df.columns:
temp_u_df = temp_u_df.drop("id")
self.user_df = temp_u_df.withColumnRenamed("id_str", "id")
temp_t_df = spark.read.format("mongo").option("uri","mongodb://127.0.0.1:{}/{}.{}".format(mongo_port,db, tweetColl)).option('sampleSize', 50000).load()
self.tweet_df = temp_t_df.select(col('id_str').alias('tweetId'),col('user.id_str').alias("userId"),col('user.screen_name').alias('screen_name'),col('created_at'),col('{}'.format(categoryColumnName)).alias('category'))
temp_e_df= spark.read.format("mongo").option("uri","mongodb://127.0.0.1:{}/{}.{}".format(mongo_port,db, edgesColl)).option('sampleSize', 5000).load()
self.edges_df = temp_e_df.dropDuplicates(["Source","Target"])
def excludeSpammers(self, userdf, tweetdf, edgesdf):
try:
self.spammers_df = userdf.filter(userdf['experiment_spam.category_rf']=="1").select(col("id"))
spammers_list = self.spammers_df.rdd.map(lambda r: r[0]).collect()
self.tweetdf = tweetdf.filter(~tweetdf.userId.isin(spammers_list))
self.user_df = userdf.filter(~userdf.id.isin(spammers_list))
self.edges_df = edgesdf.filter(~edgesdf.Source.isin(spammers_list) & ~edgesdf.Target.isin(spammers_list))
except AnalysisException:
pass
def mergeDataFrames(self, userFeatures_df, centralities_df, method_name):
self.merged_df = centralities_df.join(userFeatures_df,centralities_df.id==userFeatures_df.id) \
.select(userFeatures_df["id"], \
"user_features{}".format(("_"+method_name) if len(method_name)>0 else ""), \
"centralities{}".format(("_"+method_name) if len(method_name)>0 else ""))
def exportDataFrame(self, df, outputFile):
df.coalesce(1).write.format('json').save(outputFile)
def saveResultsToDB(self, results_df, user_df, mongo_port, db, outputCollection):
user_info_with_features=user_df.join(results_df,user_df.id == results_df.id,how='left').withColumnRenamed("id", "id_str")
user_info_with_features.write.format("mongo").mode("append").option("uri","mongodb://127.0.0.1:{}/{}.{}".format(mongo_port, db, outputCollection)).save()
def normalizeDataFrame(self, dataframe, features, centralities, categories, centrality_categories, method_name, method="minmax"):
# UDF for converting column type from vector to double type
unlist = udf(lambda x: round(float(list(x)[0]),7), DoubleType())
# scale "user_features"
df=dataframe.select("id")
for clm in features:
sub_df=dataframe.select("id")
for sub_clm in categories:
temp=dataframe.select("id","user_features{}.{}.{}".format(("_"+method_name) if len(method_name)>0 else "",clm,sub_clm)).fillna(0)
# VectorAssembler Transformation - Converting column to vector type
assembler = VectorAssembler(inputCols=[sub_clm],outputCol=sub_clm+"_Vect")
if method=="minmax":
# MinMaxScaler Transformation
scaler = MinMaxScaler(min=0.0,max=1.0,inputCol=sub_clm+"_Vect", outputCol=sub_clm+"_Scaled")
if method=="std":
# StandartScaler Transformation
scaler = StandardScaler(inputCol=sub_clm+"_Vect", outputCol=sub_clm+"_Scaled",withStd=True, withMean=False)
# Pipeline of VectorAssembler and selected scaler
pipeline = Pipeline(stages=[assembler, scaler])
# Fitting pipeline on dataframe
temp = pipeline.fit(temp).transform(temp).withColumn(sub_clm+"_Scaled", unlist(sub_clm+"_Scaled")).drop(sub_clm+"_Vect").drop(sub_clm).withColumnRenamed(sub_clm+"_Scaled",sub_clm)
# Add scaled columns to sub_df
sub_df=sub_df.join(temp,"id")
# Assign categories as sub-columns to feature columns
sub_df=sub_df.select("id",struct(sub_df.columns[1:]).alias(clm))
# Add feature columns to df
df=df.join(sub_df,"id")
# Assign features as sub-columns to "user_features"
features_df=df.select("id",struct(df.columns[1:]).alias("user_features{}".format(("_"+method_name) if len(method_name)>0 else "")))
# scale "centralities"
df=dataframe.select("id")
for clm in centralities:
sub_df=dataframe.select("id")
for sub_clm in centrality_categories:
temp=dataframe.select("id","centralities{}.{}.{}".format(("_"+method_name) if len(method_name)>0 else "",clm,sub_clm)).fillna(0)
# VectorAssembler Transformation - Converting column to vector type
assembler = VectorAssembler(inputCols=[sub_clm],outputCol=sub_clm+"_Vect")
# MinMaxScaler Transformation
scaler = MinMaxScaler(min=0.0,max=1.0,inputCol=sub_clm+"_Vect", outputCol=sub_clm+"_Scaled")
# Pipeline of VectorAssembler and MinMaxScaler
pipeline = Pipeline(stages=[assembler, scaler])
# Fitting pipeline on dataframe
temp = pipeline.fit(temp).transform(temp).withColumn(sub_clm+"_Scaled", unlist(sub_clm+"_Scaled")).drop(sub_clm+"_Vect").drop(sub_clm).withColumnRenamed(sub_clm+"_Scaled",sub_clm)
# Add scaled columns to sub_df
sub_df=sub_df.join(temp,"id")
# Assign categories as sub-columns to centrality columns
sub_df=sub_df.select("id",struct(sub_df.columns[1:]).alias(clm))
# Add centrality columns to df
df=df.join(sub_df,"id")
# Assign centralities as sub-columns to "centralities"
centralities_df=df.select("id",struct(df.columns[1:]).alias("centralities{}".format(("_"+method_name) if len(method_name)>0 else "")))
# join centralities and user features
self.normalized_df=features_df.join(centralities_df,"id")
# def getResultsAsPandas(self):
# return self.normalized_df.toPandas()