Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[2787] trunk/components-ext/src
|
From: <chr...@us...> - 2011-05-09 13:37:41
|
Revision: 2787
http://dl-learner.svn.sourceforge.net/dl-learner/?rev=2787&view=rev
Author: christinaunger
Date: 2011-05-09 13:37:35 +0000 (Mon, 09 May 2011)
Log Message:
-----------
Update SPARQL Template Generation
Modified Paths:
--------------
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/GrammarFilter.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParser.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParserTokenManager.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAG_Parser.jj
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/ParseException.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/SimpleCharStream.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/Token.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/TokenMgrError.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/SlotBuilder.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/Templator.java
trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex
trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/PatternMatchingTest.java
trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/TemplateGenerationTest.java
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/GrammarFilter.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/GrammarFilter.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/GrammarFilter.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -27,10 +27,10 @@
final static String[] NAMED_Strings = {"named", "called"};
static ParseGrammar filter(String taggedinput,LTAGLexicon grammar,List<Integer> temps) {
-
+
SlotBuilder slotbuilder = new SlotBuilder();
- List<String> input = getWordList(taggedinput);
+ List<String> input = getWordList(taggedinput.trim());
input.add(0,"#"); // This is important. Don't mess with the parser!
ParseGrammar parseG = new ParseGrammar(input.size());
@@ -75,7 +75,7 @@
for (Pair<String,String> p : named) {
try {
- TreeNode tree = c.construct(p.getFirst());
+ TreeNode tree = c.construct(p.getFirst().replaceAll("_"," "));
int gid = grammar.addTree(grammar.size(), new Pair<String,TreeNode>(token,tree), Collections.singletonList(p.getSecond()));
add(parseG, tree, gid-1, localID);
@@ -176,7 +176,7 @@
newtaggedstring += part + " ";
}
}
- newtaggedstring = newtaggedstring.trim();
+
// build token-POStag-pairs
String[] newparts = newtaggedstring.trim().split(" ");
for (String s : newparts) {
@@ -185,8 +185,7 @@
} else {
System.out.println("Oh no, " + s + " has no POS tag!"); // DEBUG
}
- }
- buildSlotFor = Preprocessor.condenseNominalPhrases(buildSlotFor);
+ }
System.out.println("build slot for: " + buildSlotFor + "\n");
List<String[]> entries = slotbuilder.build(taggedinput,buildSlotFor);
@@ -282,9 +281,11 @@
List<String> result = new ArrayList<String>();
for (String s : string.split(" ")) {
- result.add(s.substring(0,s.indexOf("/")));
+ result.add(s.substring(0,s.indexOf("/")));
}
+ System.out.println("Word list: " + result);
+
return result;
}
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -132,37 +132,32 @@
return condensedstring;
}
- public static List<Pair<String,String>> condenseNominalPhrases(List<Pair<String,String>> tokenPOSpairs) {
+ public static String condenseNominals(String s) {
- List<Pair<String,String>> out = new ArrayList<Pair<String,String>>();
-
- String flat = "";
- for (Pair<String,String> p : tokenPOSpairs) {
- flat += " " + p.fst.trim() + "/" + p.snd.trim();
- }
- flat = flat.trim();
+ String flat = s;
Matcher m;
- Pattern nnpPattern = Pattern.compile("\\s?((\\w+)/NNP[S]?\\s(\\w+))/NNP[S]?");
- Pattern nnPattern = Pattern.compile("\\s?((\\w+)/NN[S]?\\s(\\w+))/NN[S]?");
+ Pattern nnpPattern = Pattern.compile("\\s?((\\w+)/NNP[S]?\\s(\\w+))/NNP[S]?(\\W|$)");
+ Pattern nnPattern = Pattern.compile("\\s?((\\w+)/NN[S]?\\s(\\w+))/NN[S]?(\\W|$)");
+ Pattern nnnnpPattern = Pattern.compile("\\s?((\\w+)/NNP[S]?)\\s(\\w+)/NN[S]?(\\W|$)");
m = nnpPattern.matcher(flat);
while (m.find()) {
flat = flat.replaceFirst(m.group(1),m.group(2) + "_" + m.group(3));
}
+ m = nnpPattern.matcher(flat);
+ while (m.find()) {
+ flat = flat.replaceFirst(m.group(1),m.group(2) + "_" + m.group(3));
+ }
m = nnPattern.matcher(flat);
while (m.find()) {
flat = flat.replaceFirst(m.group(1),m.group(2) + "_" + m.group(3));
}
-
- System.out.println("NNP stuff: " + flat);
-
- String[] flatParts = flat.split(" ");
- for (String part : flatParts) {
- System.out.println(part);
- out.add(new Pair<String,String>(part.substring(0,part.indexOf("/")).replaceAll("_"," "), part.substring(part.indexOf("/")+1)));
+ m = nnnnpPattern.matcher(flat);
+ while (m.find()) {
+ flat = flat.replaceFirst(m.group(1),m.group(2) + "/JJ");
}
-
- return out;
+
+ return flat;
}
}
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParser.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParser.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParser.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -2,16 +2,11 @@
package org.dllearner.algorithm.tbsl.ltag.reader;
import java.util.ArrayList;
+import java.util.Arrays;
import java.util.List;
+import org.dllearner.algorithm.tbsl.ltag.data.*;
+import org.dllearner.algorithm.tbsl.ltag.agreement.*;
-import org.dllearner.algorithm.tbsl.ltag.agreement.Feature;
-import org.dllearner.algorithm.tbsl.ltag.data.Category;
-import org.dllearner.algorithm.tbsl.ltag.data.FootNode;
-import org.dllearner.algorithm.tbsl.ltag.data.SubstNode;
-import org.dllearner.algorithm.tbsl.ltag.data.TerminalNode;
-import org.dllearner.algorithm.tbsl.ltag.data.Tree;
-import org.dllearner.algorithm.tbsl.ltag.data.TreeNode;
-
public class LTAGTreeParser implements LTAGTreeParserConstants {
/** Main entry point. */
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParserTokenManager.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParserTokenManager.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAGTreeParserTokenManager.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -1,5 +1,10 @@
/* Generated By:JavaCC: Do not edit this line. LTAGTreeParserTokenManager.java */
package org.dllearner.algorithm.tbsl.ltag.reader;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.List;
+import org.dllearner.algorithm.tbsl.ltag.data.*;
+import org.dllearner.algorithm.tbsl.ltag.agreement.*;
/** Token Manager. */
public class LTAGTreeParserTokenManager implements LTAGTreeParserConstants
@@ -100,21 +105,21 @@
switch(jjstateSet[--i])
{
case 2:
- if ((0x7fffffeL & l) != 0L)
+ if ((0x7fffffe80000000L & l) != 0L)
{
+ if (kind > 11)
+ kind = 11;
+ jjCheckNAdd(0);
+ }
+ else if ((0x7fffffeL & l) != 0L)
+ {
if (kind > 12)
kind = 12;
jjCheckNAdd(1);
}
- else if ((0x7fffffe00000000L & l) != 0L)
- {
- if (kind > 11)
- kind = 11;
- jjCheckNAdd(0);
- }
break;
case 0:
- if ((0x7fffffe00000000L & l) == 0L)
+ if ((0x7fffffe80000000L & l) == 0L)
break;
kind = 11;
jjCheckNAdd(0);
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAG_Parser.jj
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAG_Parser.jj 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/LTAG_Parser.jj 2011-05-09 13:37:35 UTC (rev 2787)
@@ -21,13 +21,13 @@
PARSER_BEGIN(LTAGTreeParser)
-package ltag.reader;
+package org.dllearner.algorithm.tbsl.ltag.reader;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
-import ltag.data.*;
-import ltag.agreement.*;
+import org.dllearner.algorithm.tbsl.ltag.data.*;
+import org.dllearner.algorithm.tbsl.ltag.agreement.*;
public class LTAGTreeParser {
@@ -202,7 +202,7 @@
}
}
-TOKEN: {<WORD: (["a"-"z"]|["0"-"9"]|["?"]|["-"]|["!"]|[","]|[";"]|["."]|[":"]|["/"])+>}
+TOKEN: {<WORD: (["a"-"z"]|["0"-"9"]|["?"]|["-"]|["_"]|["!"]|[","]|[";"]|["."]|[":"]|["/"])+>}
TOKEN: {<CATEGORY: (["A"-"Z"])+>}
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/ParseException.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/ParseException.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/ParseException.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -184,4 +184,4 @@
}
}
-/* JavaCC - OriginalChecksum=83ef9865be98df5303271061c019e2cd (do not edit this line) */
+/* JavaCC - OriginalChecksum=3c249cda5771570567da479d4816c9e8 (do not edit this line) */
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/SimpleCharStream.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/SimpleCharStream.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/SimpleCharStream.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -468,4 +468,4 @@
}
}
-/* JavaCC - OriginalChecksum=a11c4feda26a580a0ed87cfec6fef57f (do not edit this line) */
+/* JavaCC - OriginalChecksum=6f78e8f501021ab7dfa602bb3724055c (do not edit this line) */
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/Token.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/Token.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/Token.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -128,4 +128,4 @@
}
}
-/* JavaCC - OriginalChecksum=86ab0d8d1ea8fe24b02fb84bc0089ecf (do not edit this line) */
+/* JavaCC - OriginalChecksum=67903b6a4b68c2296c1fa28eb89303c2 (do not edit this line) */
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/TokenMgrError.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/TokenMgrError.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/reader/TokenMgrError.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -144,4 +144,4 @@
this(LexicalError(EOFSeen, lexState, errorLine, errorColumn, errorAfter, curChar), reason);
}
}
-/* JavaCC - OriginalChecksum=a1db3550bf73c10ce8a15134b3c45c2a (do not edit this line) */
+/* JavaCC - OriginalChecksum=c1534bb74977aaf9daef2e021f7f41e9 (do not edit this line) */
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/SlotBuilder.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/SlotBuilder.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/SlotBuilder.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -108,10 +108,18 @@
String[] dpEntry2b = {token,
"(DP DET[det] (NP " + treetoken + " DP[pobj]))",
"<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + tokenfluent + "(x), SLOT_of(x,y) ] ],[(l2,y,pobj,<<e,t>,t>),(l3,x,det,e)],[l2=l1,l3=l1],[" + slotP + "," + "SLOT_of/PROPERTY/" + "]>"};
+ String[] npEntry1 = {token,
+ "(NP " + treetoken + " DP[pobj])",
+ "<x,l1,<e,t>,[ l1:[ | SLOT_" + tokenfluent + "(y,x) ] ],[(l2,y,pobj,<<e,t>,t>)],[l2=l1],[" + slotP + "]>"};
+ String[] npEntry2 = {token,
+ "(NP " + treetoken + " DP[pobj])",
+ "<x,l1,<e,t>,[ l1:[ | SLOT_" + tokenfluent + "(x), SLOT_of(x,y) ] ],[(l2,y,pobj,<<e,t>,t>)],[l2=l1],[" + slotP + "," + "SLOT_of/PROPERTY/" + "]>"};
result.add(dpEntry1a);
result.add(dpEntry1b);
result.add(dpEntry2a);
result.add(dpEntry2b);
+ result.add(npEntry1);
+ result.add(npEntry2);
}
}
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/Templator.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/Templator.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/Templator.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -59,11 +59,13 @@
s = Preprocessor.normalize(s);
String tagged = tagger.tag(s);
System.out.println("Tagged input: " + tagged);
- tagged = Preprocessor.condense(tagged);
- System.out.println("Preprocessed: " + tagged);
- p.parse(tagged,g);
+ String newtagged = Preprocessor.condenseNominals(tagged);
+ newtagged = Preprocessor.condense(newtagged);
+ System.out.println("Preprocessed: " + newtagged);
+ p.parse(newtagged,g);
+
if (p.getDerivationTrees().isEmpty()) {
p.clear(g,p.getTemps());
clearAgain = false;
Modified: trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex
===================================================================
--- trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex 2011-05-09 13:37:35 UTC (rev 2787)
@@ -20,8 +20,8 @@
// TO BE: YES/NO QUESTIONS
- is || (S V:'is' DP[subject] DP[object]) || <x, l1, t, [ l1:[ | ], l2:[ | x=y ] ], [ (l3,x,subject,<<e,t>,t>), (l4,y,object,<<e,t>,t>) ], [ l3<l1, l4<l1, l2<scope(l3), l2<scope(l4) ],[]>
- is || (S V:'is' DP[subject] ADJ[comp]) || <x, l1, t, [ l1:[ | x=y ]], [ (l2,x,subject,<<e,t>,t>), (l3,y,comp,<e,t>) ], [ l2=l1, l3=l2 ],[]>
+ is || (S (VP V:'is' DP[subject] DP[object])) || <x, l1, t, [ l1:[ | ], l2:[ | x=y ] ], [ (l3,x,subject,<<e,t>,t>), (l4,y,object,<<e,t>,t>) ], [ l3<l1, l4<l1, l2<scope(l3), l2<scope(l4) ],[]>
+ is || (S (VP V:'is' DP[subject] ADJ[comp])) || <x, l1, t, [ l1:[ | x=y ]], [ (l2,x,subject,<<e,t>,t>), (l3,y,comp,<e,t>) ], [ l2=l1, l3=l2 ],[]>
was || (S V:'was' DP[subject] DP[object]) || <x, l1, t, [ l1:[ | ], l2:[ | x=y ] ], [ (l3,x,subject,<<e,t>,t>), (l4,y,object,<<e,t>,t>) ], [ l3<l1, l4<l1, l2<scope(l3), l2<scope(l4) ],[]>
was || (S V:'was' DP[subject] ADJ[comp]) || <x, l1, t, [ l1:[ | x=y ]], [ (l2,x,subject,<<e,t>,t>), (l3,y,comp,<e,t>) ], [ l2=l1, l3=l2 ],[]>
are || (S V:'are' DP[subject] DP[object]) || <x, l1, t, [ l1:[ | ], l2:[ | x=y ] ], [ (l3,x,subject,<<e,t>,t>), (l4,y,object,<<e,t>,t>) ], [ l3<l1, l4<l1, l2<scope(l3), l2<scope(l4) ],[]>
@@ -86,12 +86,15 @@
more than || (DP DET:'more' DET:'than' NUM[num] NP[np]) || <x,l1,<<e,t>,t>,[ l1:[ c | count(y,c), greater(c,z) ] ],[(l2,y,np,<e,t>),(l3,z,num,e)],[l2=l1,l3=l1],[]>
less than || (DP DET:'less' DET:'than' NUM[num] NP[np]) || <x,l1,<<e,t>,t>,[ l1:[ c | count(y,c), less(c,z) ] ],[(l2,y,np,<e,t>),(l3,z,num,e)],[l2=l1,l3=l1],[]>
+ // HOW
+ how || (DP DET:'how' ADJ[adj]) || <x,l1,<<e,t>,t>,[ l1:[?x|] ],[ (x,l2,adj,<e,t>) ],[l2=l1],[]>
// EMPTY STUFF
// ------------
also || (VP ADV:'also' VP*) || <x,l1,t,[ l1:[|] ],[],[],[]>
+ also || (DP ADV:'also' DP*) || <x,l1,<<e,t>,t>,[ l1:[|] ],[],[],[]>
// WH WORDS
Modified: trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/PatternMatchingTest.java
===================================================================
--- trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/PatternMatchingTest.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/PatternMatchingTest.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -11,13 +11,13 @@
public static void main(String[] args) {
// TODO Auto-generated method stub
- String s = "how/WRB many/JJ software/NN companies/NN are/VBP located/VBN in/IN New/NNP York/NNP";
+ String s = "New/NNP York/NNP City/NNP is/VBZ a/DT US/NNP state/NN";
- Pattern nprepPattern = Pattern.compile("\\s((\\w+)/NN[S]?\\s(\\w+))/NN[S]?");
+ Pattern nprepPattern = Pattern.compile("\\s?((\\w+)/NNP[S]?)\\s(\\w+)/NN[S]?(\\W|$)");
Matcher m = nprepPattern.matcher(s);
while (m.find()) {
System.out.println("Found!");
- s = s.replaceFirst(m.group(1),m.group(2) + "_" + m.group(3));
+ s = s.replaceFirst(m.group(1),m.group(2) + "/JJ");
}
System.out.println(s);
Modified: trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/TemplateGenerationTest.java
===================================================================
--- trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/TemplateGenerationTest.java 2011-05-09 04:35:54 UTC (rev 2786)
+++ trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/TemplateGenerationTest.java 2011-05-09 13:37:35 UTC (rev 2787)
@@ -54,7 +54,7 @@
* @param args
*/
public static void main(String[] args) {
- File file = new File("src/main/resources/tbsl/evaluation/dbpedia-test-questions.xml");
+ File file = new File("src/main/resources/tbsl/evaluation/dbpedia-train.xml");
List<String> questions = readQuestions(file);
Templator templateGenerator = new Templator();
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
