View Javadoc
1   ///////////////////////////////////////////////////////////////////////////////////////////////
2   // checkstyle: Checks Java source code and other text files for adherence to a set of rules.
3   // Copyright (C) 2001-2025 the original author or authors.
4   //
5   // This library is free software; you can redistribute it and/or
6   // modify it under the terms of the GNU Lesser General Public
7   // License as published by the Free Software Foundation; either
8   // version 2.1 of the License, or (at your option) any later version.
9   //
10  // This library is distributed in the hope that it will be useful,
11  // but WITHOUT ANY WARRANTY; without even the implied warranty of
12  // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
13  // Lesser General Public License for more details.
14  //
15  // You should have received a copy of the GNU Lesser General Public
16  // License along with this library; if not, write to the Free Software
17  // Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
18  ///////////////////////////////////////////////////////////////////////////////////////////////
19  
20  package com.puppycrawl.tools.checkstyle.checks.javadoc;
21  
22  import java.util.ArrayList;
23  import java.util.Arrays;
24  import java.util.BitSet;
25  import java.util.List;
26  import java.util.Optional;
27  import java.util.regex.Pattern;
28  import java.util.stream.Stream;
29  
30  import com.puppycrawl.tools.checkstyle.StatelessCheck;
31  import com.puppycrawl.tools.checkstyle.api.DetailNode;
32  import com.puppycrawl.tools.checkstyle.api.JavadocTokenTypes;
33  import com.puppycrawl.tools.checkstyle.utils.CommonUtil;
34  import com.puppycrawl.tools.checkstyle.utils.JavadocUtil;
35  import com.puppycrawl.tools.checkstyle.utils.TokenUtil;
36  
37  /**
38   * <div>
39   * Checks that
40   * <a href="https://www.oracle.com/technical-resources/articles/java/javadoc-tool.html#firstsentence">
41   * Javadoc summary sentence</a> does not contain phrases that are not recommended to use.
42   * Summaries that contain only the {@code {@inheritDoc}} tag are skipped.
43   * Summaries that contain a non-empty {@code {@return}} are allowed.
44   * Check also violate Javadoc that does not contain first sentence, though with {@code {@return}} a
45   * period is not required as the Javadoc tool adds it.
46   * </div>
47   *
48   * <p>
49   * Note: For defining a summary, both the first sentence and the @summary tag approaches
50   * are supported.
51   * </p>
52   *
53   * <ul>
54   * <li>
55   * Property {@code forbiddenSummaryFragments} - Specify the regexp for forbidden summary fragments.
56   * Type is {@code java.util.regex.Pattern}.
57   * Default value is {@code "^$"}.
58   * </li>
59   * <li>
60   * Property {@code period} - Specify the period symbol. Used to check the first sentence ends with a
61   * period. Periods that are not followed by a whitespace character are ignored (eg. the period in
62   * v1.0). Because some periods include whitespace built into the character, if this is set to a
63   * non-default value any period will end the sentence, whether it is followed by whitespace or not.
64   * Type is {@code java.lang.String}.
65   * Default value is {@code "."}.
66   * </li>
67   * <li>
68   * Property {@code violateExecutionOnNonTightHtml} - Control when to print violations
69   * if the Javadoc being examined by this check violates the tight html rules defined at
70   * <a href="https://checkstyle.org/writingjavadocchecks.html#Tight-HTML_rules">Tight-HTML Rules</a>.
71   * Type is {@code boolean}.
72   * Default value is {@code false}.
73   * </li>
74   * </ul>
75   *
76   * <p>
77   * Parent is {@code com.puppycrawl.tools.checkstyle.TreeWalker}
78   * </p>
79   *
80   * <p>
81   * Violation Message Keys:
82   * </p>
83   * <ul>
84   * <li>
85   * {@code javadoc.missed.html.close}
86   * </li>
87   * <li>
88   * {@code javadoc.parse.rule.error}
89   * </li>
90   * <li>
91   * {@code javadoc.unclosedHtml}
92   * </li>
93   * <li>
94   * {@code javadoc.wrong.singleton.html.tag}
95   * </li>
96   * <li>
97   * {@code summary.first.sentence}
98   * </li>
99   * <li>
100  * {@code summary.javaDoc}
101  * </li>
102  * <li>
103  * {@code summary.javaDoc.missing}
104  * </li>
105  * <li>
106  * {@code summary.javaDoc.missing.period}
107  * </li>
108  * </ul>
109  *
110  * @since 6.0
111  */
112 @StatelessCheck
113 public class SummaryJavadocCheck extends AbstractJavadocCheck {
114 
115     /**
116      * A key is pointing to the warning message text in "messages.properties"
117      * file.
118      */
119     public static final String MSG_SUMMARY_FIRST_SENTENCE = "summary.first.sentence";
120 
121     /**
122      * A key is pointing to the warning message text in "messages.properties"
123      * file.
124      */
125     public static final String MSG_SUMMARY_JAVADOC = "summary.javaDoc";
126 
127     /**
128      * A key is pointing to the warning message text in "messages.properties"
129      * file.
130      */
131     public static final String MSG_SUMMARY_JAVADOC_MISSING = "summary.javaDoc.missing";
132 
133     /**
134      * A key is pointing to the warning message text in "messages.properties" file.
135      */
136     public static final String MSG_SUMMARY_MISSING_PERIOD = "summary.javaDoc.missing.period";
137 
138     /**
139      * This regexp is used to convert multiline javadoc to single-line without stars.
140      */
141     private static final Pattern JAVADOC_MULTILINE_TO_SINGLELINE_PATTERN =
142             Pattern.compile("\n +(\\*)|^ +(\\*)");
143 
144     /**
145      * This regexp is used to remove html tags, whitespace, and asterisks from a string.
146      */
147     private static final Pattern HTML_ELEMENTS =
148             Pattern.compile("<[^>]*>");
149 
150     /** Default period literal. */
151     private static final String DEFAULT_PERIOD = ".";
152 
153     /** Summary tag text. */
154     private static final String SUMMARY_TEXT = "@summary";
155 
156     /** Return tag text. */
157     private static final String RETURN_TEXT = "@return";
158 
159     /** Set of allowed Tokens tags in summary java doc. */
160     private static final BitSet ALLOWED_TYPES = TokenUtil.asBitSet(
161                     JavadocTokenTypes.WS,
162                     JavadocTokenTypes.DESCRIPTION,
163                     JavadocTokenTypes.TEXT);
164 
165     /**
166      * Specify the regexp for forbidden summary fragments.
167      */
168     private Pattern forbiddenSummaryFragments = CommonUtil.createPattern("^$");
169 
170     /**
171      * Specify the period symbol. Used to check the first sentence ends with a period. Periods that
172      * are not followed by a whitespace character are ignored (eg. the period in v1.0). Because some
173      * periods include whitespace built into the character, if this is set to a non-default value
174      * any period will end the sentence, whether it is followed by whitespace or not.
175      */
176     private String period = DEFAULT_PERIOD;
177 
178     /**
179      * Setter to specify the regexp for forbidden summary fragments.
180      *
181      * @param pattern a pattern.
182      * @since 6.0
183      */
184     public void setForbiddenSummaryFragments(Pattern pattern) {
185         forbiddenSummaryFragments = pattern;
186     }
187 
188     /**
189      * Setter to specify the period symbol. Used to check the first sentence ends with a period.
190      * Periods that are not followed by a whitespace character are ignored (eg. the period in v1.0).
191      * Because some periods include whitespace built into the character, if this is set to a
192      * non-default value any period will end the sentence, whether it is followed by whitespace or
193      * not.
194      *
195      * @param period period's value.
196      * @since 6.2
197      */
198     public void setPeriod(String period) {
199         this.period = period;
200     }
201 
202     @Override
203     public int[] getDefaultJavadocTokens() {
204         return new int[] {
205             JavadocTokenTypes.JAVADOC,
206         };
207     }
208 
209     @Override
210     public int[] getRequiredJavadocTokens() {
211         return getAcceptableJavadocTokens();
212     }
213 
214     @Override
215     public void visitJavadocToken(DetailNode ast) {
216         final Optional<DetailNode> inlineTagNode = getInlineTagNode(ast);
217         boolean shouldValidateUntaggedSummary = true;
218         if (inlineTagNode.isPresent()) {
219             final DetailNode node = inlineTagNode.get();
220             if (isSummaryTag(node) && isDefinedFirst(node)) {
221                 shouldValidateUntaggedSummary = false;
222                 validateSummaryTag(node);
223             }
224             else if (isInlineReturnTag(node)) {
225                 shouldValidateUntaggedSummary = false;
226                 validateInlineReturnTag(node);
227             }
228         }
229         if (shouldValidateUntaggedSummary && !startsWithInheritDoc(ast)) {
230             validateUntaggedSummary(ast);
231         }
232     }
233 
234     /**
235      * Checks the javadoc text for {@code period} at end and forbidden fragments.
236      *
237      * @param ast the javadoc text node
238      */
239     private void validateUntaggedSummary(DetailNode ast) {
240         final String summaryDoc = getSummarySentence(ast);
241         if (summaryDoc.isEmpty()) {
242             log(ast.getLineNumber(), MSG_SUMMARY_JAVADOC_MISSING);
243         }
244         else if (!period.isEmpty()) {
245             if (summaryDoc.contains(period)) {
246                 final Optional<String> firstSentence = getFirstSentence(ast, period);
247 
248                 if (firstSentence.isPresent()) {
249                     if (containsForbiddenFragment(firstSentence.get())) {
250                         log(ast.getLineNumber(), MSG_SUMMARY_JAVADOC);
251                     }
252                 }
253                 else {
254                     log(ast.getLineNumber(), MSG_SUMMARY_FIRST_SENTENCE);
255                 }
256             }
257             else {
258                 log(ast.getLineNumber(), MSG_SUMMARY_FIRST_SENTENCE);
259             }
260         }
261     }
262 
263     /**
264      * Gets the node for the inline tag if present.
265      *
266      * @param javadoc javadoc root node.
267      * @return the node for the inline tag if present.
268      */
269     private static Optional<DetailNode> getInlineTagNode(DetailNode javadoc) {
270         return Arrays.stream(javadoc.getChildren())
271             .filter(SummaryJavadocCheck::isInlineTagPresent)
272             .findFirst()
273             .map(SummaryJavadocCheck::getInlineTagNodeForAst);
274     }
275 
276     /**
277      * Whether the {@code {@summary}} tag is defined first in the javadoc.
278      *
279      * @param inlineSummaryTag node of type {@link JavadocTokenTypes#JAVADOC_INLINE_TAG}
280      * @return {@code true} if the {@code {@summary}} tag is defined first in the javadoc
281      */
282     private static boolean isDefinedFirst(DetailNode inlineSummaryTag) {
283         boolean isDefinedFirst = true;
284         DetailNode currentAst = inlineSummaryTag;
285         while (currentAst != null && isDefinedFirst) {
286             switch (currentAst.getType()) {
287                 case JavadocTokenTypes.TEXT:
288                     isDefinedFirst = currentAst.getText().isBlank();
289                     break;
290                 case JavadocTokenTypes.HTML_ELEMENT:
291                     isDefinedFirst = !isTextPresentInsideHtmlTag(currentAst);
292                     break;
293                 default:
294                     break;
295             }
296             currentAst = JavadocUtil.getPreviousSibling(currentAst);
297         }
298         return isDefinedFirst;
299     }
300 
301     /**
302      * Whether some text is present inside the HTML element or tag.
303      *
304      * @param node DetailNode of type {@link JavadocTokenTypes#HTML_TAG}
305      *             or {@link JavadocTokenTypes#HTML_ELEMENT}
306      * @return {@code true} if some text is present inside the HTML element or tag
307      */
308     public static boolean isTextPresentInsideHtmlTag(DetailNode node) {
309         DetailNode nestedChild = JavadocUtil.getFirstChild(node);
310         if (node.getType() == JavadocTokenTypes.HTML_ELEMENT) {
311             nestedChild = JavadocUtil.getFirstChild(nestedChild);
312         }
313         boolean isTextPresentInsideHtmlTag = false;
314         while (nestedChild != null && !isTextPresentInsideHtmlTag) {
315             switch (nestedChild.getType()) {
316                 case JavadocTokenTypes.TEXT:
317                     isTextPresentInsideHtmlTag = !nestedChild.getText().isBlank();
318                     break;
319                 case JavadocTokenTypes.HTML_TAG:
320                 case JavadocTokenTypes.HTML_ELEMENT:
321                     isTextPresentInsideHtmlTag = isTextPresentInsideHtmlTag(nestedChild);
322                     break;
323                 default:
324                     break;
325             }
326             nestedChild = JavadocUtil.getNextSibling(nestedChild);
327         }
328         return isTextPresentInsideHtmlTag;
329     }
330 
331     /**
332      * Checks if the inline tag node is present.
333      *
334      * @param ast ast node to check.
335      * @return true, if the inline tag node is present.
336      */
337     private static boolean isInlineTagPresent(DetailNode ast) {
338         return getInlineTagNodeForAst(ast) != null;
339     }
340 
341     /**
342      * Returns an inline javadoc tag node that is within a html tag.
343      *
344      * @param ast html tag node.
345      * @return inline summary javadoc tag node or null if no node is found.
346      */
347     private static DetailNode getInlineTagNodeForAst(DetailNode ast) {
348         DetailNode node = ast;
349         DetailNode result = null;
350         // node can never be null as this method is called when there is a HTML_ELEMENT
351         if (node.getType() == JavadocTokenTypes.JAVADOC_INLINE_TAG) {
352             result = node;
353         }
354         else if (node.getType() == JavadocTokenTypes.HTML_TAG) {
355             // HTML_TAG always has more than 2 children.
356             node = node.getChildren()[1];
357             result = getInlineTagNodeForAst(node);
358         }
359         else if (node.getType() == JavadocTokenTypes.HTML_ELEMENT
360                 // Condition for SINGLETON html element which cannot contain summary node
361                 && node.getChildren()[0].getChildren().length > 1) {
362             // Html elements have one tested tag before actual content inside it
363             node = node.getChildren()[0].getChildren()[1];
364             result = getInlineTagNodeForAst(node);
365         }
366         return result;
367     }
368 
369     /**
370      * Checks if the javadoc inline tag is {@code {@summary}} tag.
371      *
372      * @param javadocInlineTag node of type {@link JavadocTokenTypes#JAVADOC_INLINE_TAG}
373      * @return {@code true} if inline tag is summary tag.
374      */
375     private static boolean isSummaryTag(DetailNode javadocInlineTag) {
376         return isInlineTagWithName(javadocInlineTag, SUMMARY_TEXT);
377     }
378 
379     /**
380      * Checks if the first tag inside ast is {@code {@return}} tag.
381      *
382      * @param javadocInlineTag node of type {@link JavadocTokenTypes#JAVADOC_INLINE_TAG}
383      * @return {@code true} if first tag is return tag.
384      */
385     private static boolean isInlineReturnTag(DetailNode javadocInlineTag) {
386         return isInlineTagWithName(javadocInlineTag, RETURN_TEXT);
387     }
388 
389     /**
390      * Checks if the first tag inside ast is a tag with the given name.
391      *
392      * @param javadocInlineTag node of type {@link JavadocTokenTypes#JAVADOC_INLINE_TAG}
393      * @param name name of inline tag.
394      *
395      * @return {@code true} if first tag is a tag with the given name.
396      */
397     private static boolean isInlineTagWithName(DetailNode javadocInlineTag, String name) {
398         final DetailNode[] child = javadocInlineTag.getChildren();
399 
400         // Checking size of ast is not required, since ast contains
401         // children of Inline Tag, as at least 2 children will be present which are
402         // RCURLY and LCURLY.
403         return name.equals(child[1].getText());
404     }
405 
406     /**
407      * Checks the inline summary (if present) for {@code period} at end and forbidden fragments.
408      *
409      * @param inlineSummaryTag node of type {@link JavadocTokenTypes#JAVADOC_INLINE_TAG}
410      */
411     private void validateSummaryTag(DetailNode inlineSummaryTag) {
412         final String inlineSummary = getContentOfInlineCustomTag(inlineSummaryTag);
413         final String summaryVisible = getVisibleContent(inlineSummary);
414         if (summaryVisible.isEmpty()) {
415             log(inlineSummaryTag.getLineNumber(), MSG_SUMMARY_JAVADOC_MISSING);
416         }
417         else if (!period.isEmpty()) {
418             final boolean isPeriodNotAtEnd =
419                     summaryVisible.lastIndexOf(period) != summaryVisible.length() - 1;
420             if (isPeriodNotAtEnd) {
421                 log(inlineSummaryTag.getLineNumber(), MSG_SUMMARY_MISSING_PERIOD);
422             }
423             else if (containsForbiddenFragment(inlineSummary)) {
424                 log(inlineSummaryTag.getLineNumber(), MSG_SUMMARY_JAVADOC);
425             }
426         }
427     }
428 
429     /**
430      * Checks the inline return for forbidden fragments.
431      *
432      * @param inlineReturnTag node of type {@link JavadocTokenTypes#JAVADOC_INLINE_TAG}
433      */
434     private void validateInlineReturnTag(DetailNode inlineReturnTag) {
435         final String inlineReturn = getContentOfInlineCustomTag(inlineReturnTag);
436         final String returnVisible = getVisibleContent(inlineReturn);
437         if (returnVisible.isEmpty()) {
438             log(inlineReturnTag.getLineNumber(), MSG_SUMMARY_JAVADOC_MISSING);
439         }
440         else if (containsForbiddenFragment(inlineReturn)) {
441             log(inlineReturnTag.getLineNumber(), MSG_SUMMARY_JAVADOC);
442         }
443     }
444 
445     /**
446      * Gets the content of inline custom tag.
447      *
448      * @param inlineTag inline tag node.
449      * @return String consisting of the content of inline custom tag.
450      */
451     public static String getContentOfInlineCustomTag(DetailNode inlineTag) {
452         final DetailNode[] childrenOfInlineTag = inlineTag.getChildren();
453         final StringBuilder customTagContent = new StringBuilder(256);
454         final int indexOfContentOfSummaryTag = 3;
455         if (childrenOfInlineTag.length != indexOfContentOfSummaryTag) {
456             DetailNode currentNode = childrenOfInlineTag[indexOfContentOfSummaryTag];
457             while (currentNode.getType() != JavadocTokenTypes.JAVADOC_INLINE_TAG_END) {
458                 extractInlineTagContent(currentNode, customTagContent);
459                 currentNode = JavadocUtil.getNextSibling(currentNode);
460             }
461         }
462         return customTagContent.toString();
463     }
464 
465     /**
466      * Extracts the content of inline custom tag recursively.
467      *
468      * @param node DetailNode
469      * @param customTagContent content of custom tag
470      */
471     private static void extractInlineTagContent(DetailNode node,
472         StringBuilder customTagContent) {
473         final DetailNode[] children = node.getChildren();
474         if (children.length == 0) {
475             customTagContent.append(node.getText());
476         }
477         else {
478             for (DetailNode child : children) {
479                 if (child.getType() != JavadocTokenTypes.LEADING_ASTERISK) {
480                     extractInlineTagContent(child, customTagContent);
481                 }
482             }
483         }
484     }
485 
486     /**
487      * Gets the string that is visible to user in javadoc.
488      *
489      * @param summary entire content of summary javadoc.
490      * @return string that is visible to user in javadoc.
491      */
492     private static String getVisibleContent(String summary) {
493         final String visibleSummary = HTML_ELEMENTS.matcher(summary).replaceAll("");
494         return visibleSummary.trim();
495     }
496 
497     /**
498      * Tests if first sentence contains forbidden summary fragment.
499      *
500      * @param firstSentence string with first sentence.
501      * @return {@code true} if first sentence contains forbidden summary fragment.
502      */
503     private boolean containsForbiddenFragment(String firstSentence) {
504         final String javadocText = JAVADOC_MULTILINE_TO_SINGLELINE_PATTERN
505                 .matcher(firstSentence).replaceAll(" ");
506         return forbiddenSummaryFragments.matcher(trimExcessWhitespaces(javadocText)).find();
507     }
508 
509     /**
510      * Trims the given {@code text} of duplicate whitespaces.
511      *
512      * @param text the text to transform.
513      * @return the finalized form of the text.
514      */
515     private static String trimExcessWhitespaces(String text) {
516         final StringBuilder result = new StringBuilder(256);
517         boolean previousWhitespace = true;
518 
519         for (char letter : text.toCharArray()) {
520             final char print;
521             if (Character.isWhitespace(letter)) {
522                 if (previousWhitespace) {
523                     continue;
524                 }
525 
526                 previousWhitespace = true;
527                 print = ' ';
528             }
529             else {
530                 previousWhitespace = false;
531                 print = letter;
532             }
533 
534             result.append(print);
535         }
536 
537         return result.toString();
538     }
539 
540     /**
541      * Checks if the node starts with an {&#64;inheritDoc}.
542      *
543      * @param root the root node to examine.
544      * @return {@code true} if the javadoc starts with an {&#64;inheritDoc}.
545      */
546     private static boolean startsWithInheritDoc(DetailNode root) {
547         boolean found = false;
548 
549         for (DetailNode child : root.getChildren()) {
550             if (child.getType() == JavadocTokenTypes.JAVADOC_INLINE_TAG
551                     && child.getChildren()[1].getType() == JavadocTokenTypes.INHERIT_DOC_LITERAL) {
552                 found = true;
553             }
554             if ((child.getType() == JavadocTokenTypes.TEXT
555                     || child.getType() == JavadocTokenTypes.HTML_ELEMENT)
556                     && !CommonUtil.isBlank(child.getText())) {
557                 break;
558             }
559         }
560 
561         return found;
562     }
563 
564     /**
565      * Finds and returns summary sentence.
566      *
567      * @param ast javadoc root node.
568      * @return violation string.
569      */
570     private static String getSummarySentence(DetailNode ast) {
571         final StringBuilder result = new StringBuilder(256);
572         for (DetailNode child : ast.getChildren()) {
573             if (child.getType() != JavadocTokenTypes.EOF
574                     && ALLOWED_TYPES.get(child.getType())) {
575                 result.append(child.getText());
576             }
577             else {
578                 final String summary = result.toString();
579                 if (child.getType() == JavadocTokenTypes.HTML_ELEMENT
580                         && CommonUtil.isBlank(summary)) {
581                     result.append(getStringInsideTag(summary,
582                             child.getChildren()[0].getChildren()[0]));
583                 }
584             }
585         }
586         return result.toString().trim();
587     }
588 
589     /**
590      * Get concatenated string within text of html tags.
591      *
592      * @param result javadoc string
593      * @param detailNode javadoc tag node
594      * @return java doc tag content appended in result
595      */
596     private static String getStringInsideTag(String result, DetailNode detailNode) {
597         final StringBuilder contents = new StringBuilder(result);
598         DetailNode tempNode = detailNode;
599         while (tempNode != null) {
600             if (tempNode.getType() == JavadocTokenTypes.TEXT) {
601                 contents.append(tempNode.getText());
602             }
603             tempNode = JavadocUtil.getNextSibling(tempNode);
604         }
605         return contents.toString();
606     }
607 
608     /**
609      * Finds the first sentence.
610      *
611      * @param ast The Javadoc root node.
612      * @param period The configured period symbol.
613      * @return An Optional containing the first sentence
614      *     up to and excluding the period, or an empty
615      *     Optional if no ending was found.
616      */
617     private static Optional<String> getFirstSentence(DetailNode ast, String period) {
618         final List<String> sentenceParts = new ArrayList<>();
619         Optional<String> result = Optional.empty();
620         for (String text : (Iterable<String>) streamTextParts(ast)::iterator) {
621             final Optional<String> sentenceEnding = findSentenceEnding(text, period);
622 
623             if (sentenceEnding.isPresent()) {
624                 sentenceParts.add(sentenceEnding.get());
625                 result = Optional.of(String.join("", sentenceParts));
626                 break;
627             }
628             else {
629                 sentenceParts.add(text);
630             }
631         }
632         return result;
633     }
634 
635     /**
636      * Streams through all the text under the given node.
637      *
638      * @param node The Javadoc node to examine.
639      * @return All the text in all nodes that have no child nodes.
640      */
641     private static Stream<String> streamTextParts(DetailNode node) {
642         final Stream<String> stream;
643         if (node.getChildren().length == 0) {
644             stream = Stream.of(node.getText());
645         }
646         else {
647             stream = Stream.of(node.getChildren())
648                 .flatMap(SummaryJavadocCheck::streamTextParts);
649         }
650         return stream;
651     }
652 
653     /**
654      * Finds the end of a sentence. The end of sentence detection here could be replaced in the
655      * future by Java's built-in BreakIterator class.
656      *
657      * @param text The string to search.
658      * @param period The period character to find.
659      * @return An Optional containing the string up to and excluding the period,
660      *     or empty Optional if no ending was found.
661      */
662     private static Optional<String> findSentenceEnding(String text, String period) {
663         int periodIndex = text.indexOf(period);
664         Optional<String> result = Optional.empty();
665         while (periodIndex >= 0) {
666             final int afterPeriodIndex = periodIndex + period.length();
667 
668             // Handle western period separately as it is only the end of a sentence if followed
669             // by whitespace. Other period characters often include whitespace in the character.
670             if (!DEFAULT_PERIOD.equals(period)
671                 || afterPeriodIndex >= text.length()
672                 || Character.isWhitespace(text.charAt(afterPeriodIndex))) {
673                 final String resultStr = text.substring(0, periodIndex);
674                 result = Optional.of(resultStr);
675                 break;
676             }
677             else {
678                 periodIndex = text.indexOf(period, afterPeriodIndex);
679             }
680         }
681         return result;
682     }
683 }