Added more documentation

src/main/scala/workflow/graph/Cacher.scala

@@ -12,13 +12,13 @@ import scala.reflect.ClassTag
  * @tparam T Type of the input to cache.
  */
 case class Cacher[T: ClassTag](name: Option[String] = None) extends Transformer[T,T] with Logging {
-  override protected def apply(in: RDD[T]): RDD[T] = {
+  override def apply(in: RDD[T]): RDD[T] = {
     logInfo(s"CACHING ${name.getOrElse(in.id)}")
     name match {
       case Some(x) => in.cache().setName(x)
       case None => in.cache()
     }
   }
 
-  override protected def apply(in: T): T = in
+  override def apply(in: T): T = in
 }

src/main/scala/workflow/graph/ExtractSaveablePrefixes.scala

@@ -1,7 +1,8 @@
 package workflow.graph
 
 /**
- * Extract the prefixes whose state we want to save (aka for any cacher or estimator node)
+ * Extract the prefixes of all Nodes whose state we want to save for reuse by other Pipeline apply and fit calls.
+ * This is all nodes that either have a Cacher or an EstimatorOperator as the internal operator.
  */
 object ExtractSaveablePrefixes extends Rule {
   override def apply(plan: Graph, prefixes: Map[NodeId, Prefix]): (Graph, Map[NodeId, Prefix]) = {

src/main/scala/workflow/graph/GraphExecutor.scala

@@ -62,7 +62,7 @@ private[graph] class GraphExecutor(val graph: Graph, val optimize: Boolean = tru
           val operator = optimizedGraph.getOperator(node)
           val expression = operator.execute(depExpressions)
 
-          // Save state if prefix exists!
+          // Save state if the prefix is being considered for saving/loading state!
           // Note: This whole process isn't threadsafe & things could get messed up if
           // multiple pipelines are executing and/or optimizing at once
           if (prefixes.contains(node)) {

src/main/scala/workflow/graph/Identity.scala

@@ -10,6 +10,6 @@ import scala.reflect.ClassTag
  * @tparam T Type of the input and, by definition, output.
  */
 case class Identity[T : ClassTag]() extends Transformer[T,T] {
-  override protected def apply(in: T): T = in
-  override protected def apply(in: RDD[T]): RDD[T] = in
+  override def apply(in: T): T = in
+  override def apply(in: RDD[T]): RDD[T] = in
 }

src/main/scala/workflow/graph/Prefix.scala

@@ -22,4 +22,9 @@ private[graph] object Prefix {
   }
 }
 
+/**
+ * This case class represents the logical prefix of a node in a Pipeline.
+ * @param operator The operator stored at the node
+ * @param deps The prefixes of the operator's dependencies
+ */
 private[graph] case class Prefix(operator: Operator, deps: Seq[Prefix])

src/main/scala/workflow/graph/SavedStateLoadRule.scala

@@ -1,7 +1,8 @@
 package workflow.graph
 
 /**
- * A rule to load any saved state for the nodes w/ prefixes attached
+ * A rule to load any saved state for the global [[Pipeline.state]] prefix state table
+ * for nodes we want to consider either loading or saving the results of.
  */
 object SavedStateLoadRule extends Rule {
   override def apply(plan: Graph, prefixes: Map[NodeId, Prefix]): (Graph, Map[NodeId, Prefix]) = {

src/main/scala/workflow/graph/UnusedBranchRemovalRule.scala

@@ -1,7 +1,8 @@
 package workflow.graph
 
 /**
- * A rule to remove all nodes & sources in a graph that don't lead to any sink.
+ * A rule to remove all nodes & sources in a graph that don't lead to any sink,
+ * and are effectively unused.
  */
 object UnusedBranchRemovalRule extends Rule {
   override def apply(plan: Graph, prefixes: Map[NodeId, Prefix]): (Graph, Map[NodeId, Prefix]) = {