Constant.scala 14.2 KB
Newer Older
wang-jinfeng committed
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
package mobvista.dmp.datasource.fmp

import java.text.SimpleDateFormat

import com.google.gson.{JsonArray, JsonObject}
import mobvista.dmp.datasource.dm.Constant.{allZero, didPtn}
import mobvista.dmp.util.DateUtil
import mobvista.prd.datasource.util.GsonUtil
import org.apache.commons.lang3.StringUtils
import org.apache.spark.broadcast.Broadcast
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.catalyst.expressions.GenericRowWithSchema
import org.apache.spark.sql.{DataFrame, SparkSession}
import org.apache.spark.storage.StorageLevel

import scala.collection.JavaConversions._

/**
  * @package: mobvista.dmp.datasource.fmp
  * @author: wangjf
  * @date: 2019-06-10
  * @time: 16:57
  * @emial: jinfeng.wang@mobvista.com
  * @phone: 152-1062-7698
  */
object Constant {

  val sdf1 = new SimpleDateFormat("yyyy-MM-dd")
  val sdf2 = new SimpleDateFormat("yyyyMMdd")

  def parseJsonStringJob(jsonString: String): JobProfileEntity = {
    GsonUtil.fromJson(GsonUtil.String2JsonObject(jsonString), classOf[JobProfileEntity])
  }

  def parseDimension(json: JsonObject): DimensionEntity = {
    GsonUtil.fromJson(json, classOf[DimensionEntity])
  }

  def getDF(spark: SparkSession, jobProfileEntity: JobProfileEntity): (DataFrame, Int) = {

    import spark.implicits._
    var user_sql =
      """
        |SHOW PARTITIONS dwh.dm_user_info
      """.stripMargin
    var partDF = spark.sql(user_sql)
    var date = partDF.orderBy(partDF("partition").desc).first.getString(0).split("=")(1)

    user_sql =
      s"""
         |SELECT UPPER(device_id) device_id, frequency FROM dwh.dm_user_info WHERE dt = '${date}'
        """.stripMargin

    if (jobProfileEntity.last_req_day != null) {
      val updateDate = DateUtil.getDayByString(date, "yyyyMMdd", -jobProfileEntity.last_req_day)
      val update_date = sdf1.format(sdf2.parse(updateDate))
      user_sql = user_sql + s" AND update_date >= '${update_date}'"
    }

    if (StringUtils.isNotBlank(jobProfileEntity.behavior) && GsonUtil.String2JsonArray(jobProfileEntity.behavior).size() > 0) {
      var behavior = ""
      GsonUtil.String2JsonArray(jobProfileEntity.behavior).foreach(j => {
        behavior += "'" + j.getAsString.toLowerCase + "',"
      })
      behavior = "(" + behavior.substring(0, behavior.length - 1) + ")"
      user_sql = user_sql + s" AND LOWER(behavior) IN ${behavior}"
    }

    var active_sql =
      """
        |SHOW PARTITIONS dwh.dm_active_tag
      """.stripMargin
    partDF = spark.sql(active_sql)
    date = partDF.orderBy(partDF("partition").desc).take(2)(1).getString(0).split("/")(0).split("=")(1)

    val part = partDF.orderBy(partDF("partition").desc).take(2)(1).getString(0).split("/")(1).split("=")(1)
    /*
    if (!part.equals("month")) {
      date = DateUtil.getDayByString(date, "yyyyMMdd", -1)
    }
    */

    active_sql =
      s"""
         |SELECT UPPER(device_id) device_id, tags, part FROM dwh.dm_active_tag WHERE dt = '${date}'
      """.stripMargin

    if (StringUtils.isNotBlank(jobProfileEntity.country) && GsonUtil.String2JsonArray(jobProfileEntity.country).size() > 0) {
      var country = ""
      GsonUtil.String2JsonArray(jobProfileEntity.country).foreach(j => {
        country += "'" + j.getAsString.toUpperCase + "',"
      })
      country = "(" + country.substring(0, country.length - 1) + ")"
      user_sql = user_sql + s" AND UPPER(country) IN $country"

      active_sql = active_sql + s" AND UPPER(country_code) IN $country"
    }
    if (StringUtils.isNotBlank(jobProfileEntity.platform) && GsonUtil.String2JsonArray(jobProfileEntity.platform).size() > 0) {
      var platform = ""
      //  var platform_ids = ""
      GsonUtil.String2JsonArray(jobProfileEntity.platform).foreach(j => {
        platform += "'" + j.getAsString.toUpperCase + "',"
        //  platform_ids += "'" + (if (j.getAsString.toUpperCase.equals("IOS")) "0" else "1") + "',"
      })
      platform = "(" + platform.substring(0, platform.length - 1) + ")"
      //  platform_ids = "(" + platform_ids.substring(0, platform_ids.length - 1) + ")"
      user_sql = user_sql + s" AND UPPER(platform) IN $platform"

      //  active_sql = active_sql + s" AND UPPER(platform) IN $platform_ids"
    }

    if (StringUtils.isNotBlank(jobProfileEntity.package_name) && GsonUtil.String2JsonArray(jobProfileEntity.package_name).size() > 0) {
      var package_sql =
        """
          |SHOW PARTITIONS dwh.package_mapping
        """.stripMargin
      partDF = spark.sql(package_sql)
      date = partDF.orderBy(partDF("partition").desc).first.getString(0).split("=")(1)
      package_sql =
        s"""
           |SELECT id, package_name FROM dwh.package_mapping WHERE dt = '${date}'
        """.stripMargin
      val bMap = spark.sparkContext.broadcast(spark.sql(package_sql).rdd.map(r => {
        (r.getAs("package_name").toString.toUpperCase, r.getAs("id").toString)
      }).collectAsMap())

      spark.udf.register("filterByPackage", filterByPackage _)
      var package_name = ""
      GsonUtil.String2JsonArray(jobProfileEntity.package_name).foreach(j => {
        package_name += "" + bMap.value.getOrElse(j.getAsString.toUpperCase, "-1") + ","
      })
      package_name = package_name.substring(0, package_name.length - 1)
      user_sql = user_sql + s" AND filterByPackage(CONCAT_WS(',',install),'${package_name}')"
    }
    spark.udf.register("check_deviceId", check_deviceId _)

    user_sql = user_sql + " AND @check_deviceId"

    println("user_sql ==>> " + user_sql)
    val user_df = spark.sql(user_sql.replace("@check_deviceId", "check_deviceId(device_id)"))
    //  .rdd.map(r => {DeviceFrequencyEntity(r.getAs("device_id"), r.getAs("frequency"))})
    //  .persist(StorageLevel.MEMORY_AND_DISK_SER)

    active_sql = active_sql + " AND @check_deviceId"
    println("active_sql ==>> " + active_sql)
    val active_df = spark.sql(active_sql.replace("@check_deviceId", "check_deviceId(device_id)"))

    //  .rdd.map(r => {DeviceActiveEntity(r.getAs("device_id"), r.getAs("tags"), r.getAs("part"))})

    val df = if (StringUtils.isNotBlank(jobProfileEntity.dimension) && GsonUtil.String2JsonArray(jobProfileEntity.dimension).size() > 0) {
      val code_sql: String =
        """
          |SELECT new_second_id, tag_id FROM
          | dwh.dm_old2new_tag
        """.stripMargin
      val bMap = spark.sparkContext.broadcast(spark.sql(code_sql).rdd.map(r => {
        (r.getAs("new_second_id").toString, r.getAs("tag_id").toString)
      }).collectAsMap())
      val dimensionEntity = parseDimension(GsonUtil.String2JsonArray(jobProfileEntity.dimension).get(0).getAsJsonArray.get(0).getAsJsonObject)
      val part = dimensionEntity.active_dimension.toLowerCase

      val tmp_df = active_df.toDF.filter(s"part = '${part}'").select("device_id", "tags")
      tmp_df.createOrReplaceTempView("active")
      user_df.toDF.createOrReplaceTempView("user")

      println("user_df.count ==>> " + user_df.count())
      println("tmp_df.count ==>> " + tmp_df.count())
      val join_sql =
        """
          |SELECT a.device_id device_id, a.frequency frequency, b.tags tags
          | FROM user a JOIN active b
          | ON a.device_id = b.device_id
        """.stripMargin

      val join_df = spark.sql(join_sql)
        .rdd.map(r => {
        DeviceCntCount(r.getAs("device_id"), r.getAs("frequency"), r.getAs("tags"))
      }).persist(StorageLevel.MEMORY_ONLY_SER)
      println("join_df.count ==>> " + join_df.count())

      val rdd = dimensionForeach(join_df, GsonUtil.String2JsonArray(jobProfileEntity.dimension), bMap, spark).dropDuplicates
      rdd
    } else {
      val user_rdd = user_df.map(r => {
        r.getAs("device_id").toString
      }).toDF
      user_rdd.dropDuplicates
    }
    df.cache
    val limit: Int = if (jobProfileEntity.limit != null && jobProfileEntity.limit <= 10000000) {
      jobProfileEntity.limit
    } else {
      10000000
    }

    (df.limit(limit), df.count.toInt)
  }

  def check_deviceId(device_id: String): Boolean = {
    device_id.matches(didPtn) && !device_id.equals(allZero)
  }

  //  UDF
  def filterByPackage(packages: String, pkgs: String): Boolean = {
    val set = pkgs.split(",").toSet
    var flag = false
    val itr = packages.split(",").iterator
    //  val itr = packages.iterator
    while (itr.hasNext && !flag) {
      val rs = itr.next
      if (set.contains(rs)) {
        flag = true
      }
    }
    flag
  }

  // dimension 一致优化
  def dimensionForeach(df: RDD[DeviceCntCount], jsonArray: JsonArray, bMap: Broadcast[scala.collection.Map[String, String]], spark: SparkSession): DataFrame = {
    import spark.implicits._
    var outer_df = spark.emptyDataFrame
    jsonArray.foreach(json => {
      var inter_df = spark.emptyDataFrame
      json.getAsJsonArray.foreach(j => {
        val dimensionEntity = parseDimension(j.getAsJsonObject)
        //  val interest = bMap.value.getOrElse(dimensionEntity.interest, "0")
        val install_cnt = dimensionEntity.install_cnt
        val active_count = dimensionEntity.active_count
        inter_df = if (inter_df.rdd.isEmpty()) {
          parseCnt_Count(df, dimensionEntity.interest, bMap, install_cnt, active_count).toDF
        } else {
          inter_df.union(parseCnt_Count(df, dimensionEntity.interest, bMap, install_cnt, active_count).toDF)
        }
      })
      outer_df = if (outer_df.rdd.isEmpty()) {
        inter_df
      } else {
        outer_df.intersect(inter_df)
      }
    })
    outer_df
  }

  //  dimension 计算
  def dimensionForeach(user_df: RDD[DeviceFrequencyEntity], active_df: RDD[DeviceActiveEntity], jsonArray: JsonArray, bMap: Broadcast[scala.collection.Map[String, String]], spark: SparkSession): DataFrame = {
    import spark.implicits._
    var outer_df = spark.emptyDataFrame
    jsonArray.foreach(json => {
      var inter_df = spark.emptyDataFrame
      json.getAsJsonArray.foreach(j => {
        val dimensionEntity = parseDimension(j.getAsJsonObject)
        val interest = bMap.value.getOrElse(dimensionEntity.interest, "0")
        val install_cnt = dimensionEntity.install_cnt
        val active_count = dimensionEntity.active_count
        val active_dimension = dimensionEntity.active_dimension
        inter_df = if (inter_df.rdd.isEmpty()) {
          parseInterestCnt(user_df, interest, install_cnt).toDF.intersect(parseInterestActive(active_df, dimensionEntity.interest, active_count, active_dimension).toDF)
        } else {
          inter_df.union(parseInterestCnt(user_df, interest, install_cnt).toDF.intersect(parseInterestActive(active_df, dimensionEntity.interest, active_count, active_dimension).toDF))
        }
      })
      outer_df = if (outer_df.rdd.isEmpty()) {
        inter_df
      } else {
        outer_df.intersect(inter_df)
      }
    })
    outer_df
  }

  //  install_cnt 安装频度
  def parseCnt_Count(df: RDD[DeviceCntCount], interest: String, bMap: Broadcast[scala.collection.Map[String, String]], install_cnt: String, active_count: String): RDD[String] = {
    val cnter =
      if (install_cnt.split("-").size == 2) {
        (Integer.parseInt(install_cnt.split("-")(0)), Integer.parseInt(install_cnt.split("-")(1)))
      } else {
        (Integer.parseInt(install_cnt.split("-")(0)), -1)
      }

    val active_cnter =
      if (active_count.split("-").size == 2) {
        (Integer.parseInt(active_count.split("-")(0)), Integer.parseInt(active_count.split("-")(1)))
      } else {
        (Integer.parseInt(active_count.split("-")(0)), -1)
      }
    //  val query_tag = bMap.value.getOrElse(interest, "0")
    val result = df.filter(d => {
      var freq_flag = false
      if (d.frequency != null) {
        for (i <- d.frequency.indices if !freq_flag) {
          val tag = d.frequency.get(i).asInstanceOf[GenericRowWithSchema].getAs("tag").toString
          val cnt = Integer.parseInt(d.frequency.get(i).asInstanceOf[GenericRowWithSchema].getAs("cnt").toString)
          freq_flag = tag.equals(interest) && (if (cnter._2 == -1) cnt >= cnter._1 else cnt >= cnter._1 && cnt <= cnter._2)
        }
      }
      var active_flag = false
      val tags = GsonUtil.String2JsonArray(d.tags)
      for (tag <- tags if !active_flag) {
        val active = GsonUtil.fromJson(tag.getAsJsonObject, classOf[ActiveTagEntity])
        val tag_id = active.tag_id
        val cnt = active.cnt
        active_flag = tag_id.equals(interest) && (if (active_cnter._2 == -1) cnt >= active_cnter._1 else cnt >= active_cnter._1 && cnt <= active_cnter._2)
      }
      freq_flag && active_flag
    }).map(r => {
      r.deviceId
    })
    result
  }

  //  install_cnt 安装频度
  def parseInterestCnt(df: RDD[DeviceFrequencyEntity], interest: String, install_cnt: String): RDD[String] = {
    val cnter =
      if (install_cnt.split("-").size == 2) {
        (Integer.parseInt(install_cnt.split("-")(0)), Integer.parseInt(install_cnt.split("-")(1)))
      } else {
        (Integer.parseInt(install_cnt.split("-")(0)), -1)
      }
    val result = df.filter(d => {
      var flag = false
      if (d.frequency != null) {
        for (i <- d.frequency.indices if !flag) {
          val tag = d.frequency.get(i).asInstanceOf[GenericRowWithSchema].getAs("tag").toString
          val cnt = Integer.parseInt(d.frequency.get(i).asInstanceOf[GenericRowWithSchema].getAs("cnt").toString)
          flag = tag.equals(interest) && (if (cnter._2 == -1) cnt >= cnter._1 else cnt >= cnter._1 && cnt <= cnter._2)
        }
      }
      flag
    }).map(r => {
      r.device_id
    })
    result
  }

  //  active_count 活跃天数
  def parseInterestActive(df: RDD[DeviceActiveEntity], interest: String, active_count: String, active_dimension: String): RDD[String] = {
    val cnter =
      if (active_count.split("-").size == 2) {
        (Integer.parseInt(active_count.split("-")(0)), Integer.parseInt(active_count.split("-")(1)))
      } else {
        (Integer.parseInt(active_count.split("-")(0)), -1)
      }
    val result = df.filter(d => {
      d.part.equals(active_dimension)
    }).filter(d => {
      var flag = false
      val tags = GsonUtil.String2JsonArray(d.tags)
      val part = d.part
      for (tag <- tags if !flag) {
        val active = GsonUtil.fromJson(tag.getAsJsonObject, classOf[ActiveTagEntity])
        val tag_id = active.tag_id
        val cnt = active.cnt
        flag = tag_id.equals(interest) && (if (cnter._2 == -1) cnt >= cnter._1 else cnt >= cnter._1 && cnt <= cnter._2)
      }
      flag
    }).map(r => {
      r.device_id
    })
    result
  }

}