OPENNLP-1888: Add StringUtil.isBlank following the toolkit whitespace definition

A blank check under the toolkit's whitespace definition, which unlike
String.isBlank covers the no-break spaces, so annotators validating labels and
identifiers share one predicate instead of each carrying a private copy. Reads
whole code points; tests pin the no-break and figure spaces, the empty string,
and a supplementary-plane letter.

Author: krickert

opennlp-api/src/main/java/opennlp/tools/util/StringUtil.java

@@ -268,6 +268,27 @@ public static boolean isEmpty(CharSequence theString) {
     return theString.length() == 0;
   }
 
+  /**
+   * Determines whether a {@link CharSequence} is blank: empty, or made up entirely of
+   * code points that {@link #isWhitespace(int)} accepts. Unlike
+   * {@link String#isBlank()}, this follows the toolkit's whitespace definition, which
+   * includes the no-break spaces the JDK predicate leaves out, so a value spelled
+   * entirely from them cannot pass a blank check as content.
+   *
+   * @param theString The {@link CharSequence} to examine. Must not be {@code null}.
+   * @return {@code true} if {@code theString} is empty or all whitespace.
+   */
+  public static boolean isBlank(CharSequence theString) {
+    for (int i = 0; i < theString.length(); ) {
+      final int codePoint = Character.codePointAt(theString, i);
+      if (!isWhitespace(codePoint)) {
+        return false;
+      }
+      i += Character.charCount(codePoint);
+    }
+    return true;
+  }
+
   /**
    * Get the minimum of three values.
    *

opennlp-tools/src/test/java/opennlp/tools/util/StringUtilTest.java

@@ -679,4 +679,23 @@ void testLowercaseBeyondBMP() {
     String lc = StringUtil.toLowerCase(input);
     Assertions.assertArrayEquals(expectedCodePoints, lc.codePoints().toArray());
   }
+
+  /**
+   * Verifies the blank check against the toolkit's whitespace definition: the
+   * no-break space is blank here although the JDK's own check does not cover it,
+   * whitespace-only and empty values are blank, and any non-whitespace code point,
+   * supplementary ones included, makes a value non-blank.
+   */
+  @Test
+  void testIsBlankFollowsTheToolkitWhitespaceDefinition() {
+    Assertions.assertTrue(StringUtil.isBlank(""));
+    Assertions.assertTrue(StringUtil.isBlank(" \t\n"));
+    // U+00A0 no-break space and U+2007 figure space: JDK String.isBlank says false
+    Assertions.assertTrue(StringUtil.isBlank("\u00A0"));
+    Assertions.assertTrue(StringUtil.isBlank(" \u00A0\u2007 "));
+    Assertions.assertFalse(StringUtil.isBlank("a"));
+    Assertions.assertFalse(StringUtil.isBlank(" a "));
+    // U+10428, a supplementary-plane letter read as one code point, not two chars
+    Assertions.assertFalse(StringUtil.isBlank("\uD801\uDC28"));
+  }
 }