feat:ros二分类任务分桶脚本

recommend-model-produce/src/main/scala/com/tzld/piaoquan/recommend/model/recsys_01_ros_binary_weight_xgb_train.scala

@@ -0,0 +1,217 @@
+package com.tzld.piaoquan.recommend.model
+
+import com.alibaba.fastjson.JSON
+import com.tzld.piaoquan.recommend.utils.{MyHdfsUtils, ParamUtils}
+import ml.dmlc.xgboost4j.scala.spark.XGBoostClassifier
+import org.apache.commons.lang.math.NumberUtils
+import org.apache.commons.lang3.StringUtils
+import org.apache.hadoop.io.compress.GzipCodec
+import org.apache.spark.ml.evaluation.BinaryClassificationEvaluator
+import org.apache.spark.ml.feature.VectorAssembler
+import org.apache.spark.rdd.RDD
+import org.apache.spark.sql.types.DataTypes
+import org.apache.spark.sql.{Dataset, Row, SparkSession}
+
+import java.time.LocalDateTime
+import java.time.format.DateTimeFormatter
+import java.util
+import scala.io.Source
+
+object recsys_01_ros_binary_weight_xgb_train {
+  def main(args: Array[String]): Unit = {
+
+    val dt = DateTimeFormatter.ofPattern("yyyyMMddHHmm").format(LocalDateTime.now())
+
+    val spark = SparkSession
+      .builder()
+      .appName(this.getClass.getName + " : " + dt)
+      .getOrCreate()
+    val sc = spark.sparkContext
+
+    val param = ParamUtils.parseArgs(args)
+    val featureFile = param.getOrElse("featureFile", "20240703_ad_feature_name.txt")
+    val trainPath = param.getOrElse("trainPath", "/dw/recommend/model/43_recsys_ros_data_bucket/20250301")
+    val testPath = param.getOrElse("testPath", "/dw/recommend/model/43_recsys_ros_data_bucket/20250302")
+    val savePath = param.getOrElse("savePath", "/dw/recommend/model/44_recsys_ros_predict/")
+    val featureFilter = param.getOrElse("featureFilter", "XXXXXX").split(",")
+    val eta = param.getOrElse("eta", "0.01").toDouble
+    val gamma = param.getOrElse("gamma", "0.0").toDouble
+    val max_depth = param.getOrElse("max_depth", "5").toInt
+    val num_round = param.getOrElse("num_round", "100").toInt
+    val num_worker = param.getOrElse("num_worker", "20").toInt
+    val func_object = param.getOrElse("func_object", "binary:logistic")
+    val func_metric = param.getOrElse("func_metric", "auc")
+    val repartition = param.getOrElse("repartition", "20").toInt
+    val subsample = param.getOrElse("subsample", "0.8").toDouble
+    val modelPath = param.getOrElse("modelPath", "/dw/recommend/model/45_recommend_model/20250310_ros_binary_weight_1000")
+    val modelFile = param.getOrElse("modelFile", "model.tar.gz")
+    val weightField = param.getOrElse("weight_field", "return_n_uv");
+
+    val loader = getClass.getClassLoader
+    val resourceUrl = loader.getResource(featureFile)
+    val content =
+      if (resourceUrl != null) {
+        val content = Source.fromURL(resourceUrl).getLines().mkString("\n")
+        Source.fromURL(resourceUrl).close()
+        content
+      } else {
+        ""
+      }
+    println(content)
+
+    val features = content.split("\n")
+      .map(r => r.replace(" ", "").replaceAll("\n", ""))
+      .filter(r => r.nonEmpty || !featureFilter.contains(r))
+    println("features.size=" + features.length)
+
+    val trainData = createData4Ad(
+      sc.textFile(trainPath),
+      features,
+      weightField
+    )
+    println("zhangbo:train data size：" + trainData.count())
+
+    var fields = Array(
+      DataTypes.createStructField("label", DataTypes.IntegerType, true)
+    ) ++ features.map(f => DataTypes.createStructField(f, DataTypes.DoubleType, true))
+
+    fields = fields ++ Array(
+      DataTypes.createStructField("logKey", DataTypes.StringType, true),
+      DataTypes.createStructField("weight", DataTypes.DoubleType, true)
+    )
+    val schema = DataTypes.createStructType(fields)
+    val trainDataSet: Dataset[Row] = spark.createDataFrame(trainData, schema)
+    val vectorAssembler = new VectorAssembler().setInputCols(features).setOutputCol("features")
+    val xgbInput = vectorAssembler.transform(trainDataSet).select("features", "label", "weight")
+    //    val xgbParam = Map("eta" -> 0.01f,
+    //      "max_depth" -> 5,
+    //      "objective" -> "binary:logistic",
+    //      "num_class" -> 3)
+    val xgbClassifier = new XGBoostClassifier()
+      .setEta(eta)
+      .setGamma(gamma)
+      .setMissing(0.0f)
+      .setMaxDepth(max_depth)
+      .setNumRound(num_round)
+      .setSubsample(subsample)
+      .setColsampleBytree(0.8)
+      .setScalePosWeight(1)
+      .setObjective(func_object)
+      .setEvalMetric(func_metric)
+      .setFeaturesCol("features")
+      .setLabelCol("label")
+      .setWeightCol("weight")
+      .setNthread(1)
+      .setNumWorkers(num_worker)
+      .setSeed(2024)
+      .setMinChildWeight(1)
+    val model = xgbClassifier.fit(xgbInput)
+
+    if (modelPath.nonEmpty && modelFile.nonEmpty) {
+      model.write.overwrite().save(modelPath)
+    }
+
+    val testData = createData4Ad(
+      sc.textFile(testPath),
+      features,
+      weightField
+    )
+    val testDataSet = spark.createDataFrame(testData, schema)
+    val testDataSetTrans = vectorAssembler.transform(testDataSet).select("features", "label", "logKey")
+    val predictions = model.transform(testDataSetTrans)
+    //     [label, features, probability, prediction, rawPrediction]
+    println("zhangbo:columns:" + predictions.columns.mkString(","))
+    val saveData = predictions.select("label", "rawPrediction", "probability", "logKey").rdd
+      .map(r => {
+        (r.get(0), r.get(1), r.get(2), r.get(3)).productIterator.mkString("\t")
+      })
+    val hdfsPath = savePath
+    if (hdfsPath.nonEmpty && hdfsPath.startsWith("/dw/recommend/model/")) {
+      println("删除路径并开始数据写入:" + hdfsPath)
+      MyHdfsUtils.delete_hdfs_path(hdfsPath)
+      saveData.repartition(repartition).saveAsTextFile(hdfsPath, classOf[GzipCodec])
+    } else {
+      println("路径不合法，无法写入:" + hdfsPath)
+    }
+
+
+    val evaluator = new BinaryClassificationEvaluator()
+      .setLabelCol("label")
+      .setRawPredictionCol("probability")
+      .setMetricName("areaUnderROC")
+    val auc = evaluator.evaluate(predictions.select("label", "probability"))
+    println("zhangbo:auc:" + auc)
+
+    // 统计分cid的分数
+    sc.textFile(hdfsPath).map(r => {
+      val rList = r.split("\t")
+      val vid = JSON.parseObject(rList(3)).getString("vid")
+      val score = rList(2).replace("[", "").replace("]", "")
+        .split(",")(1).toDouble
+      val label = rList(0).toDouble
+      (vid, (1, label, score))
+    }).reduceByKey {
+      case (a, b) => (a._1 + b._1, a._2 + b._2, a._3 + b._3)
+    }.map {
+      case (vid, (all, zheng, scores)) =>
+        (vid, all, zheng, scores, zheng / all, scores / all)
+    }.collect().sortBy(_._1).map(_.productIterator.mkString("\t")).foreach(println)
+
+  }
+
+
+  def createData4Ad(data: RDD[String], features: Array[String], weightField: String): RDD[Row] = {
+    data.map(r => {
+      val line: Array[String] = StringUtils.split(r, '\t')
+      val logKey = line(0)
+
+      val logJson = JSON.parseObject(logKey)
+      val weight = logJson.getDouble(weightField);
+
+      val label: Int = NumberUtils.toInt(line(1))
+      val map: util.Map[String, Double] = new util.HashMap[String, Double]
+      for (i <- 2 until line.length) {
+        val fv: Array[String] = StringUtils.split(line(i), ':')
+        map.put(fv(0), NumberUtils.toDouble(fv(1), 0.0))
+      }
+
+      val v: Array[Any] = new Array[Any](features.length + 3)
+      v(0) = label
+      for (i <- 0 until features.length) {
+        v(i + 1) = map.getOrDefault(features(i), 0.0d)
+      }
+      v(features.length + 1) = logKey
+      v(features.length + 2) = weight + 1
+      Row(v: _*)
+    })
+  }
+}
+
+
+
+//rabit_timeout -> -1
+//scale_pos_weight -> 1.0
+//seed -> 0
+//handle_invalid -> error
+//features_col -> features
+//label_col -> label
+//num_workers -> 1
+//subsample -> 0.8
+//max_depth -> 5
+//probability_col -> probability
+//raw_prediction_col -> rawPrediction
+//tree_limit -> 0
+//dmlc_worker_connect_retry -> 5
+//train_test_ratio -> 1.0
+//use_external_memory -> false
+//objective -> binary:logistic
+//eval_metric -> auc
+//num_round -> 1000
+//missing -> 0.0
+//rabit_ring_reduce_threshold -> 32768
+//tracker_conf -> TrackerConf(0,python,,)
+//eta -> 0.009999999776482582
+//colsample_bytree -> 0.8
+//allow_non_zero_for_missing -> false
+//nthread -> 8
+//prediction_col -> prediction