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
|
// Copyright 2017 Yahoo Holdings. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
import org.apache.spark.mllib.recommendation.ALS
import org.apache.spark.mllib.recommendation.MatrixFactorizationModel
import org.apache.spark.mllib.recommendation.Rating
import scala.util.parsing.json.JSONObject
// Prepare data
val data_path = "data/original_data/trainPosts.json"
val sqlContext = new org.apache.spark.sql.SQLContext(sc)
val full_dataset = sqlContext.read.json(data_path)
var data = full_dataset.select($"post_id", explode($"likes").as("likes_flat"))
data = data.select($"likes_flat.uid".as("uid"), $"post_id")
data = data.filter("uid is not null and uid != '' and post_id is not null and post_id != ''")
val ratings = data.rdd.map(x => (x(0).toString, x(1).toString) match {
case (user, item) => Rating(user.toInt, item.toInt, 1)
})
// Train the model
val rank = 10
val numIterations = 10
val model = ALS.train(ratings, rank, numIterations, 0.01)
// Convert latent vectors from model to Vespa Tensor model
def writeModelFeaturesAsTensor (modelFeatures:(Int, Array[Double]), id_string:String) = {
val id = modelFeatures._1
val latentVector = modelFeatures._2
var latentVectorMap:Map[String,Double] = Map()
var output:Map[String,Any] = Map()
for ( i <- 0 until latentVector.length ){
latentVectorMap += (("user_item_cf:" + i.toString, latentVector(i)))
}
output += ((id_string, id))
output += (("user_item_cf", scala.util.parsing.json.JSONObject(latentVectorMap)))
JSONObject(output)
}
// Write user and item latent factors to disk
val product_features = model.productFeatures.map(x => writeModelFeaturesAsTensor(x, "post_id"))
product_features.saveAsTextFile("data/user_item_cf/product_features")
val user_features = model.userFeatures.map(x => writeModelFeaturesAsTensor(x, "user_id"))
user_features.saveAsTextFile("data/user_item_cf/user_features")
|
