Revision: 7268
http://languagetool.svn.sourceforge.net/languagetool/?rev=7268&view=rev
Author: milek_pl
Date: 2012-06-06 19:19:36 +0000 (Wed, 06 Jun 2012)
Log Message:
-----------
another change in the unifier, it should be now consistent
Modified Paths:
--------------
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/AbstractPatternRule.java
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Element.java
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/PatternRuleHandler.java
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Unifier.java
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
trunk/JLanguageTool/src/rules/xx/grammar.xml
trunk/JLanguageTool/src/test/org/languagetool/rules/patterns/UnifierTest.java
Modified:
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/AbstractPatternRule.java
===================================================================
---
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/AbstractPatternRule.java
2012-06-06 19:08:18 UTC (rev 7267)
+++
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/AbstractPatternRule.java
2012-06-06 19:19:36 UTC (rev 7268)
@@ -192,31 +192,37 @@
}
protected boolean testUnificationAndGroups(final boolean matched,
- final boolean lastReading, final AnalyzedToken matchToken,
- final Element elem) {
- boolean thisMatched = matched;
- if (testUnification) {
- if (matched && elem.isUnified()) {
- if (elem.isUniNegated()) {
- thisMatched = !(thisMatched && unifier.isUnified(matchToken,
elem.getUniFeatures(),
- lastReading));
- } else {
- thisMatched = thisMatched && unifier.isUnified(matchToken,
elem.getUniFeatures(),
- lastReading);
- }
- }
- if (thisMatched && getUnified) {
- unifiedTokens = unifier.getFinalUnified();
- }
- if (!elem.isUnified()) {
- unifier.reset();
- }
- }
- elem.addMemberAndGroup(matchToken);
- if (lastReading) {
- thisMatched &= elem.checkAndGroup(thisMatched);
- }
- return thisMatched;
+ final boolean lastReading, final AnalyzedToken matchToken,
+ final Element elem) {
+ boolean thisMatched = matched;
+ if (testUnification) {
+ if (matched && elem.isUnified()) {
+ if (elem.isUniNegated()) {
+ thisMatched = !(thisMatched && unifier.isUnified(matchToken,
elem.getUniFeatures(),
+ lastReading));
+ } else {
+ if (elem.isLastInUnification()) {
+ thisMatched = thisMatched &&
unifier.isUnified(matchToken, elem.getUniFeatures(),
+ lastReading);
+ } else { //we don't care about the truth value, let it run
+ unifier.isUnified(matchToken, elem.getUniFeatures(),
+ lastReading);
+ }
+
+ }
+ }
+ if (thisMatched && getUnified) {
+ unifiedTokens = unifier.getFinalUnified();
+ }
+ if (!elem.isUnified()) {
+ unifier.reset();
+ }
+ }
+ elem.addMemberAndGroup(matchToken);
+ if (lastReading) {
+ thisMatched &= elem.checkAndGroup(thisMatched);
+ }
+ return thisMatched;
}
Modified:
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Element.java
===================================================================
--- trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Element.java
2012-06-06 19:08:18 UTC (rev 7267)
+++ trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Element.java
2012-06-06 19:19:36 UTC (rev 7268)
@@ -140,7 +140,13 @@
private boolean posUnknown;
+
/**
+ * Set to true on tokens that close the unification block.
+ */
+ private boolean isLastUnified;
+
+ /**
* Creates Element that is used to match tokens in the text.
*
* @param token
@@ -784,6 +790,14 @@
return uniNegation;
}
+ public final boolean isLastInUnification() {
+ return isLastUnified;
+ }
+
+ public final void setLastInUnification() {
+ isLastUnified = true;
+ }
+
public final void setWhitespaceBefore(final boolean isWhite) {
whitespaceBefore = isWhite;
testWhitespace = true;
Modified:
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/PatternRuleHandler.java
===================================================================
---
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/PatternRuleHandler.java
2012-06-06 19:08:18 UTC (rev 7267)
+++
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/PatternRuleHandler.java
2012-06-06 19:19:36 UTC (rev 7268)
@@ -277,8 +277,10 @@
//clear the features...
equivalenceFeatures = new HashMap<String, List<String>>();
//set negation on the last token only!
- if (uniNegation) {
- elementList.get(elementList.size() - 1).setUniNegation();
+ final int lastElement = elementList.size() - 1;
+ elementList.get(lastElement).setLastInUnification();
+ if (uniNegation) {
+ elementList.get(lastElement).setUniNegation();
}
}
}
Modified:
trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Unifier.java
===================================================================
--- trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Unifier.java
2012-06-06 19:08:18 UTC (rev 7267)
+++ trunk/JLanguageTool/src/java/org/languagetool/rules/patterns/Unifier.java
2012-06-06 19:19:36 UTC (rev 7268)
@@ -36,12 +36,6 @@
*/
public class Unifier {
- //TODO: add a possibility to negate some features but not all
- /**
- * Negates the meaning of unification just like negation in Element tokens.
- */
- private boolean negation;
-
private boolean allFeatsIn;
private int tokCnt;
@@ -261,14 +255,6 @@
}
tmpFeaturesFound = new ArrayList<Boolean>(featuresFound);
}
-
- public final void setNegation(final boolean neg) {
- negation = neg;
- }
-
- public final boolean getNegation() {
- return negation;
- }
/**
* Resets after use of unification. Required.
@@ -276,7 +262,6 @@
public final void reset() {
equivalencesMatched.clear();
allFeatsIn = false;
- negation = false;
tokCnt = -1;
featuresFound.clear();
tmpFeaturesFound.clear();
Modified:
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
===================================================================
---
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
2012-06-06 19:08:18 UTC (rev 7267)
+++
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
2012-06-06 19:19:36 UTC (rev 7268)
@@ -375,8 +375,10 @@
inUnification = false;
equivalenceFeatures = new HashMap<String, List<String>>();
//set negation on the last token only!
- if (uniNegation) {
- elementList.get(elementList.size() - 1).setUniNegation();
+ final int lastElement = elementList.size() - 1;
+ elementList.get(lastElement).setLastInUnification();
+ if (uniNegation) {
+ elementList.get(lastElement).setUniNegation();
}
} else if (qName.equals(WD)) {
addNewWord(wd.toString(), wdLemma, wdPos);
Modified: trunk/JLanguageTool/src/rules/xx/grammar.xml
===================================================================
--- trunk/JLanguageTool/src/rules/xx/grammar.xml 2012-06-06 19:08:18 UTC
(rev 7267)
+++ trunk/JLanguageTool/src/rules/xx/grammar.xml 2012-06-06 19:19:36 UTC
(rev 7268)
@@ -16,6 +16,22 @@
<token regexp="yes">\p{Ll}+</token>
</equivalence>
</unification>
+ <unification feature="gender">
+ <equivalence type="masc">
+ <token regexp="yes">.*_masc</token>
+ </equivalence>
+ <equivalence type="fem">
+ <token regexp="yes">.*_fem</token>
+ </equivalence>
+ </unification>
+ <unification feature="number">
+ <equivalence type="plur">
+ <token regexp="yes">.*_plur_.*</token>
+ </equivalence>
+ <equivalence type="sing">
+ <token regexp="yes">.*_sing_.*</token>
+ </equivalence>
+ </unification>
<phrases>
<phrase id="UNIFICATION_PHRASE">
<unify>
@@ -340,5 +356,24 @@
<example type="incorrect">Dogs <marker>blah blah</marker></example>
</rule>
</rulegroup>
+ <rulegroup id="test_unification_with_negation" name="Test unification with
negation">
+ <rule>
+ <pattern>
+ <unify negate="yes">
+ <feature id="gender"/>
+ <feature id="number"/>
+ <token><exception postag="SENT_START"/></token>
+ <token/>
+ <token/>
+ </unify>
+ </pattern>
+ <message>blabla</message>
+ <example type="correct">det_sing_masc</example>
+ <example type="incorrect"><marker>det_sing_masc adj_sing_masc
subst_plur_fem</marker></example>
+ <example type="incorrect"><marker>det_sing_fem adj_sing_masc
subst_plur_fem</marker></example>
+ <example type="incorrect"><marker>det_sing_fem adj_sing_fem
subst_plur_fem</marker></example>
+ <example type="incorrect"><marker>det_plur_fem adj_sing_fem
subst_plur_fem</marker></example>
+ </rule>
+ </rulegroup>
</category>
</rules>
\ No newline at end of file
Modified:
trunk/JLanguageTool/src/test/org/languagetool/rules/patterns/UnifierTest.java
===================================================================
---
trunk/JLanguageTool/src/test/org/languagetool/rules/patterns/UnifierTest.java
2012-06-06 19:08:18 UTC (rev 7267)
+++
trunk/JLanguageTool/src/test/org/languagetool/rules/patterns/UnifierTest.java
2012-06-06 19:19:36 UTC (rev 7268)
@@ -342,7 +342,6 @@
//now let's negate this
//traditional way
- uni.setNegation(true);
satisfied = uni.isSatisfied(det_sing_masc, equiv);
uni.startUnify();
satisfied &= uni.isSatisfied(sing_masc, equiv);
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Languagetool-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/languagetool-cvs
