fix: add cross-version parallel collections support via compat package

Implement version-specific compat package objects to handle parallel
collections differences between Scala 2.12 and 2.13:

- Scala 2.12: Empty compat package (parallel collections built-in)
- Scala 2.13: Provide .par extension method via implicit class that
  delegates to CollectionConverters

Also fix .distinct call on ParIterable by converting to Seq first.

This enables .par to work across both Scala versions in KMeansPlusPlus.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

Author: derrickburns

src/main/scala-2.12/com/massivedatascience/clusterer/compat/package.scala

@@ -1,4 +1,7 @@
 package com.massivedatascience.clusterer
 
-// Scala 2.12: Parallel collections are built-in, no imports needed
-package object compat
+// Scala 2.12: Parallel collections are built-in, no additional imports needed
+// This package object exists for cross-version compatibility
+package object compat {
+  // Empty - parallel collections work out of the box in 2.12
+}

src/main/scala-2.13/com/massivedatascience/clusterer/compat/package.scala

@@ -1,13 +1,11 @@
 package com.massivedatascience.clusterer
 
-// Scala 2.13: Re-export CollectionConverters implicits for .par support
-package object compat {
-  implicit def asParIterable[A](iterable: Iterable[A]): scala.collection.parallel.CollectionConverters.IterableIsParallelizable[A] =
-    scala.collection.parallel.CollectionConverters.IterableIsParallelizable(iterable)
-
-  implicit def asParSeq[A](seq: scala.collection.Seq[A]): scala.collection.parallel.CollectionConverters.SeqIsParallelizable[A] =
-    scala.collection.parallel.CollectionConverters.SeqIsParallelizable(seq)
+import scala.language.implicitConversions
 
-  implicit def asParMap[K, V](map: scala.collection.Map[K, V]): scala.collection.parallel.CollectionConverters.MapIsParallelizable[K, V] =
-    scala.collection.parallel.CollectionConverters.MapIsParallelizable(map)
+// Scala 2.13: Provide .par extension method via compat package
+package object compat {
+  implicit class ParOps[A, CC[X] <: Iterable[X]](private val coll: CC[A]) extends AnyVal {
+    @inline def par: scala.collection.parallel.ParIterable[A] =
+      scala.collection.parallel.CollectionConverters.IterableIsParallelizable(coll).par
+  }
 }

src/main/scala/com/massivedatascience/clusterer/KMeansPlusPlus.scala

@@ -25,8 +25,7 @@ import com.massivedatascience.util.XORShiftRandom
 import scala.annotation.tailrec
 import scala.collection.mutable.ArrayBuffer
 
-// Scala 2.13 needs this import for .par; Scala 2.12 will ignore it (built-in)
-// The ParCollectionCompat trait brings this into scope
+// Cross-version parallel collections support via compat package
 import com.massivedatascience.clusterer.compat._
 
 /** This implements the <a href="http://ilpubs.stanford.edu:8090/778/1/2006-13.pdf">KMeans++ initialization
@@ -155,7 +154,7 @@ class KMeansPlusPlus(ops: BregmanPointOps) extends Serializable with Logging {
             pickWeighted(rand, cumulative).iterator
           }
 
-          val uniqueSelected = selected.distinct
+          val uniqueSelected = selected.seq.toSeq.distinct
           logger.debug(
             s"Selected ${uniqueSelected.size} unique centers from ${selected.size} samples"
           )