Revision: 7326
http://languagetool.svn.sourceforge.net/languagetool/?rev=7326&view=rev
Author: milek_pl
Date: 2012-06-12 17:28:42 +0000 (Tue, 12 Jun 2012)
Log Message:
-----------
[pl] fix false alarms
Modified Paths:
--------------
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
trunk/JLanguageTool/src/resource/pl/disambiguation.xml
trunk/JLanguageTool/src/rules/pl/grammar.xml
Modified:
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
===================================================================
---
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
2012-06-12 16:44:56 UTC (rev 7325)
+++
trunk/JLanguageTool/src/java/org/languagetool/tagging/disambiguation/rules/DisambiguationRuleLoader.java
2012-06-12 17:28:42 UTC (rev 7326)
@@ -129,6 +129,9 @@
} else if (qName.equals(AND)) {
inAndGroup = true;
tokenCountForMarker++;
+ if (inUnification) {
+ uniCounter++;
+ }
} else if (qName.equals(UNIFY)) {
inUnification = true;
uniNegation = YES.equals(attrs.getValue(NEGATE));
@@ -337,11 +340,13 @@
elementList.add(tokenElement);
}
if (inAndGroup) {
- andGroupCounter++;
+ andGroupCounter++;
}
if (inUnification) {
- tokenElement.setUnification(equivalenceFeatures);
- uniCounter++;
+ tokenElement.setUnification(equivalenceFeatures);
+ if (!inAndGroup) {
+ uniCounter++;
+ }
}
if (inUnificationDef) {
language.getDisambiguationUnifier().setEquivalence(uFeature, uType,
tokenElement);
Modified: trunk/JLanguageTool/src/resource/pl/disambiguation.xml
===================================================================
--- trunk/JLanguageTool/src/resource/pl/disambiguation.xml 2012-06-12
16:44:56 UTC (rev 7325)
+++ trunk/JLanguageTool/src/resource/pl/disambiguation.xml 2012-06-12
17:28:42 UTC (rev 7326)
@@ -125,6 +125,33 @@
</pattern>
<disambig action="unify"/>
</rule>
+
+ <rulegroup name="co (i) rusz" id="CO_RUSZ">
+ <rule>
+ <pattern >
+ <token regexp="yes">co|ani</token>
+ <marker>
+ <token>rusz</token>
+ </marker>
+ </pattern>
+ <disambig postag="qub"/>
+ <example type="untouched">Rusz tylko tę książkę!</example>
+ <example type="ambiguous"
inputform="rusz[rusza/subst:pl:gen:f,ruszyć/verb:impt:sg:sec:perf]"
outputform="rusz[rusza/qub]">Co <marker>rusz</marker> pyta o godzinę.</example>
+ </rule>
+ <rule>
+ <pattern>
+ <token>co</token>
+ <token>i</token>
+ <marker>
+ <token>rusz</token>
+ </marker>
+ </pattern>
+ <disambig postag="qub"/>
+ <example type="untouched">Rusz tylko tę książkę!</example>
+ <example type="ambiguous"
inputform="rusz[rusza/subst:pl:gen:f,ruszyć/verb:impt:sg:sec:perf]"
outputform="rusz[rusza/qub]">Co i <marker>rusz</marker> pyta o
godzinę.</example>
+ </rule>
+ </rulegroup>
+
<rulegroup name="prep + czasownik -> ~ czasownik" id="prep_verb">
<!-- przyimki zawsze z tym samym przypadkiem -->
@@ -163,6 +190,19 @@
<example type="untouched">Tu mamy piękny przykład sztuki
barokowej.</example>
<example type="ambiguous"
inputform="mamy[mama/subst:pl:acc:f,mama/subst:pl:nom:f,mama/subst:pl:voc:f,mama/subst:sg:gen:f,mieć/verb:fin:pl:pri:imperf]"
outputform="mamy[mama/subst:pl:acc:f,mama/subst:pl:nom:f,mama/subst:pl:voc:f,mama/subst:sg:gen:f]">Stanąłem
naprzeciw <marker>mamy</marker>.</example>
</rule>
+
+ <rule>
+ <pattern>
+ <unify>
+ <feature id="case"/>
+ <token postag="prep:.*"
postag_regexp="yes"><exception postag="</?[\p{Lu}:]+>"
postag_regexp="yes"/><exception postag="prep.*" negate_pos="yes"
postag_regexp="yes"/></token>
+ <token postag_regexp="yes"
postag="(adj|subst|ppron).*"><exception postag="</?[\p{Lu}:]+>"
postag_regexp="yes"/><exception>pewne</exception></token>
+ </unify>
+ </pattern>
+ <disambig action="unify"/>
+ <example type="untouched">Oni mną papiery.</example>
+ <example type="ambiguous"
inputform="mną[ja/ppron12:sg:inst:f.m.n.n1.n2:pri,miąć/verb:fin:pl:ter:imperf]"
outputform="mną[ja/ppron12:sg:inst:f.m.n.n1.n2:pri]">Jedź ze
<marker>mną</marker>!</example>
+ </rule>
</rulegroup>
<rule name="do to nie subst" id="do_non_subst">
@@ -177,7 +217,7 @@
<pattern>
<unify> <feature id="case"/>
<token postag="prep.*" postag_regexp="yes"><exception
postag="prep.*" negate_pos="yes"
postag_regexp="yes"/><exception>co</exception><exception
postag="</?[\p{Lu}:]+>" postag_regexp="yes"/></token>
- <token postag="(adj|subst|ppron|ppas):.*"
postag_regexp="yes"><exception>co</exception><exception
postag="</?[\p{Lu}:]+>|qub" postag_regexp="yes"/></token>
+ <token postag="(adj|subst|ppron.*|ppas):.*"
postag_regexp="yes"><exception>co</exception><exception
postag="</?[\p{Lu}:]+>|qub" postag_regexp="yes"/></token>
</unify>
</pattern>
<disambig action="unify"></disambig>
@@ -507,32 +547,6 @@
</rule>
</rulegroup>
- <rulegroup name="co (i) rusz" id="CO_RUSZ">
- <rule>
- <pattern >
- <token regexp="yes">co|ani</token>
- <marker>
- <token>rusz</token>
- </marker>
- </pattern>
- <disambig postag="qub"/>
- <example type="untouched">Rusz tylko tę książkę!</example>
- <example type="ambiguous"
inputform="rusz[rusza/subst:pl:gen:f,ruszyć/verb:impt:sg:sec:perf]"
outputform="rusz[rusza/qub]">Co <marker>rusz</marker> pyta o godzinę.</example>
- </rule>
- <rule>
- <pattern>
- <token>co</token>
- <token>i</token>
- <marker>
- <token>rusz</token>
- </marker>
- </pattern>
- <disambig postag="qub"/>
- <example type="untouched">Rusz tylko tę książkę!</example>
- <example type="ambiguous"
inputform="rusz[rusza/subst:pl:gen:f,ruszyć/verb:impt:sg:sec:perf]"
outputform="rusz[rusza/qub]">Co i <marker>rusz</marker> pyta o
godzinę.</example>
- </rule>
- </rulegroup>
-
<!-- TODO: czy to nadal potrzebne po PoliMorfie? -->
<rule name="skrótowce jako nieodmienne" id="SKROTOWCE_IRREG">
<pattern>
Modified: trunk/JLanguageTool/src/rules/pl/grammar.xml
===================================================================
--- trunk/JLanguageTool/src/rules/pl/grammar.xml 2012-06-12 16:44:56 UTC
(rev 7325)
+++ trunk/JLanguageTool/src/rules/pl/grammar.xml 2012-06-12 17:28:42 UTC
(rev 7326)
@@ -949,7 +949,7 @@
<short>Błąd interpunkcyjny</short>
<example type="correct">To tyle, co nic.</example>
<example correction="tyle, co" type="incorrect">To
<marker>tyle co</marker> nic.</example>
- </rule>
+ </rule>
<rule>
<pattern>
<marker>
@@ -968,7 +968,7 @@
<example type="correct">Naprawdę nie masz za co
dziękować.</example>
<example correction="Oto, do" type="incorrect"><marker>Oto
do</marker> czego doprowadziły rządy ludzi niespełna rozumu.</example>
<example type="correct">Spotykamy się w co drugi
piątek.</example>
- </rule>
+ </rule>
<rule>
<pattern>
<marker>
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
