From 3928386766b5b64f7f5e5dda6664589f3f23dcc7 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E6=A2=A6=E5=A2=83=E8=BF=B7=E7=A6=BB?=
 <dreamylost@outlook.com>
Date: Mon, 2 May 2022 11:13:30 +0800
Subject: [PATCH] add StringUtils and test

---
 build.sbt                                     |  1 -
 .../org/bitlap/csv/core/StringUtils.scala     | 28 +++++++++++++++++++
 .../core/test/CsvableAndScalableTest.scala    | 21 ++++++++++++++
 .../csv/core/test/StringUtilsTest.scala       |  7 +++++
 4 files changed, 56 insertions(+), 1 deletion(-)

diff --git a/build.sbt b/build.sbt
index 3c088de..759e53d 100644
--- a/build.sbt
+++ b/build.sbt
@@ -1,4 +1,3 @@
-import sbt.librarymanagement.InclExclRule
 import sbt.{ Def, Test }
 import sbtrelease.ReleaseStateTransformations._
 
diff --git a/csv-core/src/main/scala/org/bitlap/csv/core/StringUtils.scala b/csv-core/src/main/scala/org/bitlap/csv/core/StringUtils.scala
index 1b693f4..901569f 100644
--- a/csv-core/src/main/scala/org/bitlap/csv/core/StringUtils.scala
+++ b/csv-core/src/main/scala/org/bitlap/csv/core/StringUtils.scala
@@ -22,6 +22,8 @@
 package org.bitlap.csv.core
 
 import scala.collection.mutable.ListBuffer
+import java.util.regex.Pattern
+import scala.util.matching.Regex
 
 /**
  * split csv column value by columnSeparator.
@@ -31,6 +33,32 @@ import scala.collection.mutable.ListBuffer
  */
 object StringUtils {
 
+  private val regex: Regex = "\\{(.*?)\\}".r
+  private val kvr: Regex = "(.*):(.*)".r
+  private val pattern: Pattern = Pattern.compile(regex.toString())
+
+  def extraJsonPairs(input: String): String = {
+    val matcher = pattern.matcher(input)
+    while (matcher.find) {
+      val tail = matcher.group().tail.init
+      if (tail != null && tail.nonEmpty) {
+        return tail
+      } else return null
+    }
+
+    null
+  }
+
+  def extraJsonValues[T <: Product](jsonString: String)(func: (String, String) => T): List[T] = {
+    val pairs = extraJsonPairs(jsonString)
+    if (pairs == null) return Nil
+    val jsonElements = pairs.split(",")
+    val kvs = jsonElements.collect {
+      case kvr(k, v) if k.length > 2 && v.length > 2 => k.init.tail -> v.init.tail
+    }
+    kvs.toList.map(f => func(f._1, f._2))
+  }
+
   def splitColumns(line: String, columnSeparator: Char): List[String] = {
     val listBuffer = ListBuffer[String]()
     val columnBuffer = ListBuffer[Char]()
diff --git a/csv-core/src/test/scala/org/bitlap/csv/core/test/CsvableAndScalableTest.scala b/csv-core/src/test/scala/org/bitlap/csv/core/test/CsvableAndScalableTest.scala
index 25980da..d88b10e 100644
--- a/csv-core/src/test/scala/org/bitlap/csv/core/test/CsvableAndScalableTest.scala
+++ b/csv-core/src/test/scala/org/bitlap/csv/core/test/CsvableAndScalableTest.scala
@@ -21,6 +21,8 @@
 
 package org.bitlap.csv.core.test
 
+import org.bitlap.csv.core.StringUtils
+
 import org.scalatest.flatspec.AnyFlatSpec
 import org.scalatest.matchers.should.Matchers
 import org.bitlap.csv.core.ScalableBuilder
@@ -122,4 +124,23 @@ class CsvableAndScalableTest extends AnyFlatSpec with Matchers {
     assert(metrics.head.get.dimensions.head.key == "city")
     assert(metrics.head.get.dimensions.head.value == "北京")
   }
+
+  "CsvableAndScalable3" should "ok when using StringUtils" in {
+    val metrics = csvData
+      .split("\n")
+      .map(csv =>
+        ScalableBuilder[Metric2]
+          .setField[Seq[Dimension3]](
+            _.dimensions,
+            dims => StringUtils.extraJsonValues[Dimension3](dims)((k, v) => Dimension3(k, v))
+          )
+          .build(csv)
+          .toScala
+      )
+
+    println(metrics.toList)
+
+    assert(metrics.head.get.dimensions.head.key == "city")
+    assert(metrics.head.get.dimensions.head.value == "北京")
+  }
 }
diff --git a/csv-core/src/test/scala/org/bitlap/csv/core/test/StringUtilsTest.scala b/csv-core/src/test/scala/org/bitlap/csv/core/test/StringUtilsTest.scala
index bf580e2..fe02723 100644
--- a/csv-core/src/test/scala/org/bitlap/csv/core/test/StringUtilsTest.scala
+++ b/csv-core/src/test/scala/org/bitlap/csv/core/test/StringUtilsTest.scala
@@ -37,4 +37,11 @@ class StringUtilsTest extends AnyFlatSpec with Matchers {
     println(csv)
     assert(csv.size == 8)
   }
+
+  "StringUtilsTest2" should "ok" in {
+    val line = """abc,"{""a"":""b"",""c"":""d""}",d,12,2,false,0.1,0.23333"""
+    val csv = StringUtils.extraJsonValues[Dimension3](line)((k, v) => Dimension3(k, v))
+    println(csv)
+    assert(csv.toString() == "List(Dimension3(a,b), Dimension3(c,d))")
+  }
 }
-- 
GitLab