[SPARK-37026][ML][BUILD] Ensure the element type of ResolvedRFormula.terms is scala.Seq for Scala 2.13

### What changes were proposed in this pull request?

This PR fixes the issue that `scala.Seq[scala.collection.mutable.ArraySeq$ofRef]` will be passed to `ResolvedRFormula.terms` though it expects `scala.Seq[scala.Seq[String]]` with Scala 2.13.
As of Scala 2.13, `scala.Seq` is `scala.collection.immutable.Seq`, so this issue happens.

### Why are the changes needed?

Bug fix.
Due to this issue, `ResolvedRFormula.toString` throws `ClassCastException`.
```
java.lang.ClassCastException: scala.collection.mutable.ArraySeq$ofRef cannot be cast to scala.collection.immutable.Seq
        at scala.collection.immutable.List.map(List.scala:246)
        at scala.collection.immutable.List.map(List.scala:79)
        at org.apache.spark.ml.feature.ResolvedRFormula.toString(RFormulaParser.scala:143)
        at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
        at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
        at java.lang.reflect.Method.invoke(Method.java:498)
        at py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)
        at py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)
        at py4j.Gateway.invoke(Gateway.java:282)
        at py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
        at py4j.commands.CallCommand.execute(CallCommand.java:79)
        at py4j.ClientServerConnection.waitForCommands(ClientServerConnection.java:182)
        at py4j.ClientServerConnection.run(ClientServerConnection.java:106)
        at java.lang.Thread.run(Thread.java:748)
```

### Does this PR introduce _any_ user-facing change?

No.

### How was this patch tested?

New test is added and `build/sbt -Pscala-2.13 "testOnly org.apache.spark.ml.feature.RFormulaSuite"` passes.
CIs should ensure that this change works with Scala 2.12 too.

Closes apache#34301 from sarutak/fix-rformula-scala-2.13.

Authored-by: Kousuke Saruta <[email protected]>
Signed-off-by: Hyukjin Kwon <[email protected]>

Author: sarutak

mllib/src/main/scala/org/apache/spark/ml/feature/RFormula.scala

@@ -449,7 +449,7 @@ object RFormulaModel extends MLReadable[RFormulaModel] {
       val dataPath = new Path(path, "data").toString
       val data = sparkSession.read.parquet(dataPath).select("label", "terms", "hasIntercept").head()
       val label = data.getString(0)
-      val terms = data.getSeq[Seq[String]](1)
+      val terms = data.getSeq[scala.collection.Seq[String]](1).map(_.toSeq)
       val hasIntercept = data.getBoolean(2)
       val resolvedRFormula = ResolvedRFormula(label, terms, hasIntercept)
 

mllib/src/test/scala/org/apache/spark/ml/feature/RFormulaSuite.scala

@@ -627,4 +627,15 @@ class RFormulaSuite extends MLTest with DefaultReadWriteTest {
     assert(get_output("keep").count() == 6)
   }
 
+  test("SPARK-37026: Ensure the element type of ResolvedRFormula.terms is " +
+    "scala.Seq for Scala 2.13") {
+    withTempPath { path =>
+      val dataset = Seq((1, "foo", "zq"), (2, "bar", "zq"), (3, "bar", "zz")).toDF("id", "a", "b")
+      val rFormula = new RFormula().setFormula("id ~ a:b")
+      val model = rFormula.fit(dataset)
+      model.save(path.getCanonicalPath)
+      val newModel = RFormulaModel.load(path.getCanonicalPath)
+      newModel.resolvedFormula.toString
+    }
+  }
 }