From e5ed37a4834b8bd2abfbdac159edab81629e26fc Mon Sep 17 00:00:00 2001
From: Lucas Satabin <lucas.satabin@gnieh.org>
Date: Tue, 30 Jan 2024 23:44:33 +0100
Subject: [PATCH] Add basic util to split big texts

---
 .../scala/fs2/data/text/render/Renderer.scala | 22 ++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

diff --git a/text/shared/src/main/scala/fs2/data/text/render/Renderer.scala b/text/shared/src/main/scala/fs2/data/text/render/Renderer.scala
index 3e087ee4e..e3e2d5deb 100644
--- a/text/shared/src/main/scala/fs2/data/text/render/Renderer.scala
+++ b/text/shared/src/main/scala/fs2/data/text/render/Renderer.scala
@@ -16,10 +16,30 @@
 
 package fs2.data.text.render
 
-import fs2.{Stream, Pure}
+import fs2.{Chunk, Pure, Stream}
 
 trait Renderer[Event] {
 
+  /**
+    * Splits words in the given text into a stream of document events.
+    * This is a utility method, when you want to reformat a text using
+    * the pretty printer. The pretty printing algorithm assumes that each
+    * `DocEvent.Text` is an atomic value, so if it contains new lines it
+    * can break the computations.
+    *
+    * Between 2 words, it adds a `DocEvent.Line` event. Empty lines are not
+    * Generated.
+    *
+    * @param text The text to split
+    * @param wordBoundary The regular expression on which to split words
+    */
+  def words(text: String, wordBoundary: String = raw"\s+"): Stream[Pure, DocEvent] =
+    Stream
+      .chunk(Chunk.array(text.split(wordBoundary)))
+      .filter(_.nonEmpty)
+      .map(DocEvent.Text(_))
+      .intersperse(DocEvent.Line)
+
   /** Transforms the event into a stream of document events.
     * The stream may be partial (e.g. opening a group when the event describes a new tree node).
     */