Author: catholicon
Date: Thu Jan 18 16:14:07 2018
New Revision: 1821521
URL: http://svn.apache.org/viewvc?rev=1821521&view=rev
Log:
OAK-4401: Excerpt Highlighting for a property is not correct (backport
r1821325,r1821516) from trunk)
This is just about ensuring that SimpleExcerptProvider would use non-letter or
non-digit as delimeter marker before highlighting
Modified:
jackrabbit/oak/branches/1.8/ (props changed)
jackrabbit/oak/branches/1.8/oak-core/src/main/java/org/apache/jackrabbit/oak/query/SimpleExcerptProvider.java
jackrabbit/oak/branches/1.8/oak-core/src/test/java/org/apache/jackrabbit/oak/query/SimpleExcerptProviderTest.java
Propchange: jackrabbit/oak/branches/1.8/
------------------------------------------------------------------------------
--- svn:mergeinfo (original)
+++ svn:mergeinfo Thu Jan 18 16:14:07 2018
@@ -1,3 +1,3 @@
/jackrabbit/oak/branches/1.0:1665962
-/jackrabbit/oak/trunk:1820660-1820661,1820859,1820861,1820878,1820888,1820947,1821130,1821140-1821141,1821240,1821258,1821358,1821362,1821370,1821375,1821477,1821487
+/jackrabbit/oak/trunk:1820660-1820661,1820859,1820861,1820878,1820888,1820947,1821130,1821140-1821141,1821240,1821258,1821325,1821358,1821362,1821370,1821375,1821477,1821487,1821516
/jackrabbit/trunk:1345480
Modified:
jackrabbit/oak/branches/1.8/oak-core/src/main/java/org/apache/jackrabbit/oak/query/SimpleExcerptProvider.java
URL:
http://svn.apache.org/viewvc/jackrabbit/oak/branches/1.8/oak-core/src/main/java/org/apache/jackrabbit/oak/query/SimpleExcerptProvider.java?rev=1821521&r1=1821520&r2=1821521&view=diff
==============================================================================
---
jackrabbit/oak/branches/1.8/oak-core/src/main/java/org/apache/jackrabbit/oak/query/SimpleExcerptProvider.java
(original)
+++
jackrabbit/oak/branches/1.8/oak-core/src/main/java/org/apache/jackrabbit/oak/query/SimpleExcerptProvider.java
Thu Jan 18 16:14:07 2018
@@ -34,6 +34,7 @@ import org.apache.jackrabbit.oak.query.a
import org.apache.jackrabbit.oak.query.ast.OrImpl;
import org.apache.jackrabbit.oak.plugins.memory.PropertyValues;
+import static java.lang.Character.isLetterOrDigit;
import static org.apache.jackrabbit.util.Text.encodeIllegalXMLCharacters;
/**
@@ -239,19 +240,38 @@ class SimpleExcerptProvider {
}
int endIndex = index + token.length();
if (isLike) {
- int nextSpace = text.indexOf(" ", endIndex);
- if (nextSpace != -1) {
+ int nextSpace = endIndex;
+
+ while (nextSpace < text.length() &&
!isDelimeter(text.codePointAt(nextSpace))) {
+ nextSpace++;
+ }
+
+ if (nextSpace != text.length()) {
endIndex = nextSpace;
} else {
endIndex = text.length();
}
}
- while (index < endIndex) {
- highlightBits.set(index++);
+
+ boolean isStartOk = (index == 0) || //allow for highlighting for
token at the beginning
+ isDelimeter(text.codePointAt(index - 1)); //else token
must follow a delimeter
+ boolean isEndOk = (endIndex == text.length()) || //token is at the
end of string
+ isDelimeter(text.codePointAt(endIndex)); //else token must
precede a delimeter
+
+ if (isStartOk && isEndOk) {
+ while (index < endIndex) {
+ highlightBits.set(index++);
+ }
+ } else {
+ index = endIndex;
}
}
}
+ static boolean isDelimeter(int codePoint) {
+ return !isLetterOrDigit(codePoint);
+ }
+
static PropertyValue getExcerpt(PropertyValue value) {
Splitter listSplitter =
Splitter.on(',').trimResults().omitEmptyStrings();
StringBuilder excerpt = new StringBuilder(EXCERPT_BEGIN);
Modified:
jackrabbit/oak/branches/1.8/oak-core/src/test/java/org/apache/jackrabbit/oak/query/SimpleExcerptProviderTest.java
URL:
http://svn.apache.org/viewvc/jackrabbit/oak/branches/1.8/oak-core/src/test/java/org/apache/jackrabbit/oak/query/SimpleExcerptProviderTest.java?rev=1821521&r1=1821520&r2=1821521&view=diff
==============================================================================
---
jackrabbit/oak/branches/1.8/oak-core/src/test/java/org/apache/jackrabbit/oak/query/SimpleExcerptProviderTest.java
(original)
+++
jackrabbit/oak/branches/1.8/oak-core/src/test/java/org/apache/jackrabbit/oak/query/SimpleExcerptProviderTest.java
Thu Jan 18 16:14:07 2018
@@ -23,8 +23,10 @@ import static com.google.common.collect.
import static org.apache.jackrabbit.oak.query.SimpleExcerptProvider.highlight;
import static org.junit.Assert.assertEquals;
+import java.util.Map;
import java.util.Random;
+import com.google.common.collect.Maps;
import org.junit.Test;
public class SimpleExcerptProviderTest {
@@ -59,6 +61,61 @@ public class SimpleExcerptProviderTest {
}
}
+ @Test
+ public void hightlightCompleteWordOnly() {
+ // using 2 non-simple spaces as mentioned in
http://jkorpela.fi/chars/spaces.html
+ String[] delimiters = new String[] {" ", "\t", "\n", ":", "\u1680",
"\u00A0"};
+ Map<String, String> simpleCheck = Maps.newHashMap(); // highlight "of"
+
+ // simple ones
+ simpleCheck.put("official conflict of interest",
+ "<div><span>official conflict <strong>of</strong>
interest</span></div>");
+ simpleCheck.put("of to new city",
+ "<div><span><strong>of</strong> to new city</span></div>");
+ simpleCheck.put("out of the roof",
+ "<div><span>out <strong>of</strong> the roof</span></div>");
+ simpleCheck.put("well this is of",
+ "<div><span>well this is <strong>of</strong></span></div>");
+
+ for (Map.Entry<String, String> simple : simpleCheck.entrySet()) {
+ for (String delimiter : delimiters) {
+ String text = simple.getKey().replaceAll(" ", delimiter);
+ String expect = simple.getValue().replaceAll(" ", delimiter);
+ assertEquals("highlighting '" + text + "' for 'of' (delimiter
- '" + delimiter + "')",
+ expect, highlight(sb(text), of("of")));
+ }
+ }
+
+ Map<String, String> wildcardCheck = Maps.newHashMap(); // highlight
"of*"
+ wildcardCheck.put("office room",
+ "<div><span><strong>office</strong> room</span></div>");
+ wildcardCheck.put("office room off",
+ "<div><span><strong>office</strong> room
<strong>off</strong></span></div>");
+ wildcardCheck.put("big office room",
+ "<div><span>big <strong>office</strong> room</span></div>");
+
+ for (Map.Entry<String, String> wildcard : wildcardCheck.entrySet()) {
+ for (String delimiter : delimiters) {
+ String text = wildcard.getKey().replaceAll(" ", delimiter);
+ String expect = wildcard.getValue().replaceAll(" ", delimiter);
+ assertEquals("highlighting '" + text + "' for 'of*' (delimiter
- '" + delimiter + "')",
+ expect, highlight(sb(text), of("of*")));
+ }
+ }
+ }
+
+ @Test
+ public void multipleSearchTokens() {
+ String text = "To be, or not to be. That is the question!";
+ String expected = "<div><span>To <strong>be</strong>, " +
+ "or not to <strong>be</strong>. " +
+ "That is the <strong>question</strong>!</span></div>";
+
+ assertEquals(expected, highlight(sb(text), of("question", "be")));
+ assertEquals(expected, highlight(sb(text), of("quest*", "be")));
+ assertEquals(expected, highlight(sb(text), of("quest*", "b*")));
+ }
+
private static String randomString(Random r, String set) {
int len = r.nextInt(10);
StringBuilder buff = new StringBuilder();