edkuksa/russianmorphology
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

adding clean for strings

Browse Source 
git-svn-id: https://russianmorphology.googlecode.com/svn/trunk@50 d817d54c-26ab-11de-abc9-2f7d1455ff7a
This commit is contained in:
alexander.a.kuznetsov
2009-10-02 17:02:52 +00:00
parent 710384987c
commit ca1a9be6b7
3 changed files with 15 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
russian/src/main/java/org/apache/lucene/morphology/russian/HeuristicBuilder.java
View File

@ -31,7 +31,8 @@ public class HeuristicBuilder {
GrammaReader grammaInfo = new GrammaReader("dictonary/Dicts/Morph/rgramtab.tab");
DictonaryReader dictonaryReader = new DictonaryReader("dictonary/Dicts/SrcMorph/RusSrc/morphs.mrd", form);
StatiticsCollector statiticsCollector = new StatiticsCollector(grammaInfo);
RussianSuffixDecoderEncoder decoderEncoder = new RussianSuffixDecoderEncoder();
StatiticsCollector statiticsCollector = new StatiticsCollector(grammaInfo, decoderEncoder);
dictonaryReader.proccess(statiticsCollector);
statiticsCollector.saveHeuristic();

2
russian/src/main/java/org/apache/lucene/morphology/russian/RussianSuffixDecoderEncoder.java
View File

@ -107,6 +107,6 @@ public class RussianSuffixDecoderEncoder implements LetterDecoderEncoder {
}
public String cleanString(String s) {
return s;
return s.replace((char) (34 + RussianSuffixDecoderEncoder.RUSSIAN_SMALL_LETTER_OFFSET), (char) (6 + RussianSuffixDecoderEncoder.RUSSIAN_SMALL_LETTER_OFFSET));
}
}