Adding SpellCheckerFactory

src/main/java/org/luwrain/nlp/SpellCheckerFactory.java

@@ -0,0 +1,38 @@
+/*
+ Copyright 2012-2022 Michael Pozhidaev <[email protected]>
+
+ This file is part of LUWRAIN.
+
+ LUWRAIN is free software; you can redistribute it and/or
+ modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation; either
+ version 3 of the License, or (at your option) any later version.
+
+ LUWRAIN is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ General Public License for more details.
+*/
+
+package org.luwrain.nlp;
+
+import org.luwrain.nlp.ru.spell.*;
+
+public final class SpellCheckerFactory
+{
+ static private RuSpellChecker ru = null;
+
+ public SpellChecker newChecker(String lang)
+ {
+ switch(lang.toUpperCase())
+ {
+ case "RU": {
+ if (ru == null)
+ ru = new RuSpellChecker();
+ return ru;
+ }
+ default:
+ throw new IllegalArgumentException("Unknown language: " + lang);
+ }
+ }
+ }

src/main/java/org/luwrain/nlp/SpellProblem.java

@@ -16,8 +16,12 @@
 
 package org.luwrain.nlp;
 
-public interface SpellProblem
+import org.luwrain.core.*;
+
+public interface SpellProblem extends LineMarks.MarkObject
 {
  String getComment();
  String getShortComment();
+ int getStart();
+ int getEnd();
 }

src/main/java/org/luwrain/nlp/SpellText.java

@@ -18,6 +18,11 @@
 
 import java.util.*;
 
+import org.luwrain.core.*;
+import org.luwrain.controls.*;
+import org.luwrain.controls.DefaultLineMarks.*;
+import static org.luwrain.util.RangeUtils.*;
+
 public class SpellText
 {
  final String text;
@@ -51,6 +56,24 @@ public SpellText(String[] text, SpellChecker checker)
  this.problems = checker.check(this.text);
  }
 
+ public List<List<LineMarks.Mark>> buildMarks()
+ {
+ final List<List<LineMarks.Mark>> res = new ArrayList<>();
+ for(Fragment f: fragments)
+ {
+ final List<LineMarks.Mark> a = new ArrayList();
+ for(SpellProblem p: problems)
+ {
+ final int[] range = commonRangeByBounds(p.getStart(), p.getEnd(), f.posFrom, f.posTo);
+ if (range == null)
+ continue;
+ a.add(new MarkImpl(LineMarks.Mark.Type.WEAK, range[0] - f.posFrom, range[1] - f.posFrom, p));
+ }
+ res.add(a);
+ }
+ return res;
+ }
+
  static public final class Fragment
  {
  final int posFrom, posTo;

src/main/java/org/luwrain/nlp/ru/spell/Problem.java

@@ -0,0 +1,47 @@
+/*
+ Copyright 2012-2022 Michael Pozhidaev <[email protected]>
+
+ This file is part of LUWRAIN.
+
+ LUWRAIN is free software; you can redistribute it and/or
+ modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation; either
+ version 3 of the License, or (at your option) any later version.
+
+ LUWRAIN is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ General Public License for more details.
+*/
+
+package org.luwrain.nlp.ru.spell;
+
+import org.languagetool.rules.*;
+
+import org.luwrain.nlp.*;
+
+final class Problem implements SpellProblem
+{
+private final String message, shortMessage;
+
+ Problem(RuleMatch match)
+ {
+ this.message = match.getMessage();
+ this.shortMessage = match.getShortMessage();
+ /*
+ List<SuggestedReplacement> repl = r.getSuggestedReplacementObjects();
+ for(SuggestedReplacement rr: repl)
+ System.out.println(rr);
+ */
+ }
+
+ @Override public String getComment() { return message; }
+ @Override public String getShortComment(){ return shortMessage; }
+ @Override public int getStart(){ return 0; }
+ @Override public int getEnd(){ return 0;}
+
+ @Override public String toString()
+ {
+ return message;
+ }
+}

src/main/java/org/luwrain/nlp/ru/spell/RuSpellChecker.java

@@ -0,0 +1,54 @@
+/*
+ Copyright 2012-2022 Michael Pozhidaev <[email protected]>
+
+ This file is part of LUWRAIN.
+
+ LUWRAIN is free software; you can redistribute it and/or
+ modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation; either
+ version 3 of the License, or (at your option) any later version.
+
+ LUWRAIN is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ General Public License for more details.
+*/
+
+package org.luwrain.nlp.ru.spell;
+
+import java.util.*;
+import java.io.*;
+
+import org.languagetool.*;
+import org.languagetool.language.*;
+import org.languagetool.rules.*;
+import org.languagetool.rules.spelling.*;
+
+import org.luwrain.core.*;
+import org.luwrain.nlp.*;
+
+public final class RuSpellChecker implements SpellChecker
+{
+ final JLanguageTool langTool;
+
+ public RuSpellChecker()
+ {
+ this.langTool = new JLanguageTool(new Russian());
+ }
+
+ @Override public List<SpellProblem> check(String text)
+ {
+ final List<SpellProblem> res = new ArrayList<>();
+ try {
+ final List<RuleMatch> m = langTool.check(text);
+ System.out.println("match " + m.size());
+ for(RuleMatch mm: m)
+ res.add(new Problem(mm));
+ return res;
+ }
+ catch(IOException e)
+ {
+ throw new RuntimeException(e);
+ }
+ }
+}

src/test/java/org/luwrain/io/nlp/SpellTextTest.java

@@ -0,0 +1,35 @@
+/*
+ Copyright 2012-2022 Michael Pozhidaev <[email protected]>
+
+ This file is part of LUWRAIN.
+
+ LUWRAIN is free software; you can redistribute it and/or
+ modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation; either
+ version 3 of the License, or (at your option) any later version.
+
+ LUWRAIN is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ General Public License for more details.
+*/
+
+package org.luwrain.nlp;
+
+import org.junit.*;
+
+import org.luwrain.core.*;
+
+public class SpellTextTest extends Assert
+{
+ private SpellChecker checker = null;
+
+ @Test public void main()
+ {
+ }
+
+ @Before public void create()
+ {
+ checker = new SpellCheckerFactory().newChecker("ru");
+ }
+}

src/test/java/org/luwrain/io/nlp/ru/spell/RuSpellCheckerTest.java

@@ -0,0 +1,42 @@
+/*
+ Copyright 2012-2022 Michael Pozhidaev <[email protected]>
+
+ This file is part of LUWRAIN.
+
+ LUWRAIN is free software; you can redistribute it and/or
+ modify it under the terms of the GNU General Public
+ License as published by the Free Software Foundation; either
+ version 3 of the License, or (at your option) any later version.
+
+ LUWRAIN is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ General Public License for more details.
+*/
+
+
+package org.luwrain.nlp.ru.spell;
+
+//https://languagetool.org/development/api/org/languagetool/rules/spelling/SpellingCheckRule.html
+
+
+import java.io.*;
+import java.net.*;
+
+import org.junit.*;
+
+import org.luwrain.core.*;
+
+public class RuSpellCheckerTest extends Assert
+{
+ private RuSpellChecker c = null;
+
+ @Test public void main()
+ {
+ }
+
+ @Before public void create()
+ {
+ c = new RuSpellChecker();
+ }
+}