Build error if the number of the raw features more than 23

[xiaoerhei]

Hi developer,

Thanks for the pio template.

I am facing an build error if I tried more input features like 23 raw input features. the error log shows as below: could you help me out? thanks

[INFO] [Engine$] Using command '/home/zhangyongxing/apache-predictionio-0.12.0-incubating/PredictionIO-0.12.0-incubating/sbt/sbt' at /home/zhangyongxing/apache-predictionio-0.12.0-incubating/templates/predictionio-template-boston-house-prices to build.
[INFO] [Engine$] If the path above is incorrect, this process will fail.
[INFO] [Engine$] Uber JAR disabled. Making sure lib/pio-assembly-0.12.0-incubating.jar is absent.
[INFO] [Engine$] Going to run: /home/zhangyongxing/apache-predictionio-0.12.0-incubating/PredictionIO-0.12.0-incubating/sbt/sbt package assemblyPackageDependency in /home/zhangyongxing/apache-predictionio-0.12.0-incubating/templates/predictionio-template-boston-house-prices
[ERROR] [Engine$] [error] /home/zhangyongxing/apache-predictionio-0.12.0-incubating/templates/predictionio-template-boston-house-prices/src/main/scala/org/jpioug/template/bhp/Algorithm.scala:49: object is not a member of package scala
[ERROR] [Engine$] [error] val data = Seq((
[ERROR] [Engine$] [error] ^
[ERROR] [Engine$] [error] one error found
[ERROR] [Engine$] [error] (compile:compileIncremental) Compilation failed
[ERROR] [Engine$] [error] Total time: 4 s, completed Nov 29, 2017 10:04:10 PM
[ERROR] [Engine$] Return code of build command: /home/zhangyongxing/apache-predictionio-0.12.0-incubating/PredictionIO-0.12.0-incubating/sbt/sbt package assemblyPackageDependency is 1. Aborting.

And my code as below,just modify the src/main/scala/org/jpioug/template/bhp/Algorithm.scala:

package org.jpioug.template.bhp

import org.apache.predictionio.controller.{P2LAlgorithm, Params}
import org.apache.spark.SparkContext
import org.apache.spark.ml.PipelineModel
import org.apache.spark.sql.SparkSession
import org.jpioug.template.python.{Engine, PreparedData}

case class AlgorithmParams(name: String) extends Params

case class Query(AGE: Double,
B: Double,
CHAS: Double,
CRIM: Double,
DIS: Double,
INDUS: Double,
LSTAT: Double,
NOX: Double,
PTRATIO: Double,
RAD: Double,
RM: Double,
TAX: Double,
ZN: Double,
index_14: Double,
index_15: Double,
index_16: Double,
index_17: Double,
index_18: Double,
index_19: Double,
index_20: Double,
index_21: Double,
index_22: Double,
index_23: Double )

case class PredictedResult(label: Double) extends Serializable

class Algorithm(val ap: AlgorithmParams)
extends P2LAlgorithm[PreparedData, PipelineModel, Query, PredictedResult] {

def train(sc: SparkContext, data: PreparedData): PipelineModel = {
Engine.modelRef.get()
}

def predict(model: PipelineModel, query: Query): PredictedResult = {
val spark = SparkSession
.builder
.getOrCreate()
import spark.implicits._
val data = Seq((
query.AGE,
query.B,
query.CHAS,
query.CRIM,
query.DIS,
query.INDUS,
query.LSTAT,
query.NOX,
query.PTRATIO,
query.RAD,
query.RM,
query.TAX,
query.ZN,
query.index_14,
query.index_15,
query.index_16,
query.index_17,
query.index_18,
query.index_19,
query.index_20,
query.index_21,
query.index_22,
query.index_23
))
val df = spark.createDataset(data).toDF("AGE",
"B",
"CHAS",
"CRIM",
"DIS",
"INDUS",
"LSTAT",
"NOX",
"PTRATIO",
"RAD",
"RM",
"TAX",
"ZN",
"index_14",
"index_15",
"index_16",
"index_17",
"index_18",
"index_19",
"index_20",
"index_21",
"index_22",
"index_23"
)
val labelDf = model.transform(df)
PredictedResult(labelDf.select("prediction").first().getAs(0))
}
}

Thanks

[marevol]

It looks like Tuple 22 problem.

[xiaoerhei]

Thanks for the reply. Yes, indeed. It is the tuple 22 problem. If needed, you can close this issue, I am working to get around this 22 limit problem. Thanks

Fixing this issue: modify the "data" related code as below:

val data = mutable.MutableList [Query] ()
data += query
val df = data.toDF()