Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[2784] trunk/components-ext/src
|
From: <chr...@us...> - 2011-05-06 16:02:36
|
Revision: 2784
http://dl-learner.svn.sourceforge.net/dl-learner/?rev=2784&view=rev
Author: christinaunger
Date: 2011-05-06 16:02:30 +0000 (Fri, 06 May 2011)
Log Message:
-----------
Update: SPARQL Template Generation
Modified Paths:
--------------
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/POStagger.java
trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/SlotBuilder.java
trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex
trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/POStest.java
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/Preprocessor.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java 2011-05-04 10:00:34 UTC (rev 2783)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java 2011-05-06 16:02:30 UTC (rev 2784)
@@ -47,17 +47,22 @@
String condensedstring = taggedstring;
Matcher m;
- Pattern compAdjPattern = Pattern.compile("\\s(\\w+/RBR.([a-zA-Z_0-9]+)/JJ)");
- Pattern superAdjPattern = Pattern.compile("\\s(\\w+/RBS.([a-zA-Z_0-9]+)/JJ)");
- Pattern howAdjPattern = Pattern.compile("\\s(\\w+/WRB.([a-zA-Z_0-9]+)/JJ)");
- Pattern nprepPattern = Pattern.compile("\\s((\\w+)/NNS?.of/IN)");
- Pattern passivePattern1 = Pattern.compile("(((has)|(have)|(had))/VB[A-Z]?.been/VBN.(\\w+)/VBN.by/IN)");
- Pattern passivePattern2 = Pattern.compile("(((is)|(are)|(was)|(were))/VB[A-Z]?.(\\w+)/VBN.by/IN)");
- Pattern passpartPattern = Pattern.compile("\\s((\\w+)/VBN.by/IN)");
- Pattern vpassPattern = Pattern.compile("\\s(\\w+/VBD.(\\w+)/VBN)");
- Pattern vpassinPattern = Pattern.compile("\\s((\\w+)/VPASS.\\w+/IN)");
- Pattern gerundinPattern = Pattern.compile("\\s((\\w+)/((VBG)|(VBN)).\\w+/IN)");
- Pattern vprepPattern = Pattern.compile("\\s((\\w+)/V[A-Z]+\\s\\w+/IN)");
+ Pattern compAdjPattern = Pattern.compile("\\s(\\w+/RBR.([a-zA-Z_0-9]+)/JJ)");
+ Pattern superAdjPattern = Pattern.compile("\\s(\\w+/RBS.([a-zA-Z_0-9]+)/JJ)");
+ Pattern howAdjPattern = Pattern.compile("\\s(\\w+/WRB.([a-zA-Z_0-9]+)/JJ)");
+ Pattern nprepPattern = Pattern.compile("\\s((\\w+)/NNS?.of/IN)");
+ Pattern didPattern = Pattern.compile("(?i)(\\s((did)|(do)|(does))/VB.?)\\s");
+ Pattern passivePattern1a = Pattern.compile("(((has)|(have)|(had))/VB[A-Z]?.been/VBN.(\\w+)/VBN.by/IN)");
+ Pattern passivePattern1b = Pattern.compile("(\\s((has)|(have)|(had))/VB[A-Z]?(.+\\s)been/VBN\\s(\\w+)/VB(N|D))");
+ Pattern passivePattern2a = Pattern.compile("(((is)|(are)|(was)|(were))/VB[A-Z]?.(\\w+)/VBN.by/IN)");
+ Pattern passivePattern2b = Pattern.compile("(((is)|(are)|(was)|(were))/VB[A-Z]?.(.+)(\\s\\w+)/VB(N|D))");
+ Pattern passpartPattern = Pattern.compile("\\s((\\w+)/VBN.by/IN)");
+ Pattern vpassPattern = Pattern.compile("\\s(\\w+/VBD.(\\w+)/VBN)");
+ Pattern vpassinPattern = Pattern.compile("\\s((\\w+)/VPASS.\\w+/IN)");
+ Pattern gerundinPattern = Pattern.compile("\\s((\\w+)/((VBG)|(VBN)).\\w+/IN)");
+ Pattern vprepPattern = Pattern.compile("\\s((\\w+)/V[A-Z]+\\s\\w+/(IN|TO))");
+ Pattern whenPattern = Pattern.compile("(?i)(when/WRB\\s(.+\\s)(\\w+)/((V[A-Z]+)|(PASS[A-Z]+)))");
+ Pattern wherePattern = Pattern.compile("(?i)(where/WRB\\s(.+\\s)(\\w+)/((V[A-Z]+)|(PASS[A-Z]+)))");
m = compAdjPattern.matcher(condensedstring);
while (m.find()) {
@@ -75,14 +80,26 @@
while (m.find()) {
condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2)+"/NPREP");
}
- m = passivePattern1.matcher(condensedstring);
+ m = didPattern.matcher(condensedstring);
while (m.find()) {
+ condensedstring = condensedstring.replaceFirst(m.group(1),"");
+ }
+ m = passivePattern1a.matcher(condensedstring);
+ while (m.find()) {
condensedstring = condensedstring.replaceFirst(m.group(1),m.group(6)+"/PASSIVE");
}
- m = passivePattern2.matcher(condensedstring);
+ m = passivePattern1b.matcher(condensedstring);
while (m.find()) {
+ condensedstring = condensedstring.replaceFirst(m.group(1),m.group(6) + m.group(7)+"/PASSIVE");
+ }
+ m = passivePattern2a.matcher(condensedstring);
+ while (m.find()) {
condensedstring = condensedstring.replaceFirst(m.group(1),m.group(7)+"/PASSIVE");
}
+ m = passivePattern2b.matcher(condensedstring);
+ while (m.find()) {
+ condensedstring = condensedstring.replaceFirst(m.group(1),m.group(7) + m.group(8)+"/PASSIVE");
+ }
m = passpartPattern.matcher(condensedstring);
while (m.find()) {
condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2)+"/PASSPART");
@@ -103,47 +120,49 @@
while (m.find()) {
condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2)+"/VPREP");
}
+ m = whenPattern.matcher(condensedstring);
+ while (m.find()) {
+ condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2) + m.group(3)+"/WHEN");
+ }
+ m = wherePattern.matcher(condensedstring);
+ while (m.find()) {
+ condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2) + m.group(3)+"/WHERE");
+ }
return condensedstring;
}
- public static List<Pair<String,String>> condenseNominalPhrases(List<Pair<String,String>> tokenPOSpairs){
- List<Pair<String,String>> test = new ArrayList<Pair<String,String>>();
+ public static List<Pair<String,String>> condenseNominalPhrases(List<Pair<String,String>> tokenPOSpairs) {
- String nounPhrase = "";
- String phraseTag = "";
- for(Pair<String,String> pair : tokenPOSpairs){
- if(pair.snd.startsWith("NNP")){
- if(phraseTag.equals("NN")){
- if(!nounPhrase.isEmpty()){
- test.add(new Pair<String, String>(phraseTag.trim(), "NN"));
- nounPhrase = "";
- }
- }
- phraseTag = "NNP";
- nounPhrase += " " + pair.fst;
- } else if(pair.snd.startsWith("NN")){
- if(phraseTag.equals("NNP")){
- if(!nounPhrase.isEmpty()){
- test.add(new Pair<String, String>(phraseTag.trim(), "NNP"));
- nounPhrase = "";
- }
- }
- phraseTag = "NN";
- nounPhrase += " " + pair.fst;
- } else {
- if(!nounPhrase.isEmpty()){
- test.add(new Pair<String, String>(nounPhrase.trim(), phraseTag));
- nounPhrase = "";
- }
- test.add(pair);
- }
+ 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();
}
- if(!nounPhrase.isEmpty()){
- test.add(new Pair<String, String>(nounPhrase.trim(), phraseTag));
- nounPhrase = "";
+ flat = flat.trim();
+
+ 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]?");
+
+ 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));
+ }
- return test;
+ 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)));
+ }
+
+ return out;
}
}
Modified: trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/POStagger.java
===================================================================
--- trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/POStagger.java 2011-05-04 10:00:34 UTC (rev 2783)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/POStagger.java 2011-05-06 16:02:30 UTC (rev 2784)
@@ -20,6 +20,7 @@
}
public POStagger() throws IOException, ClassNotFoundException {
taggermodel = "src/main/resources/tbsl/models/bidirectional-distsim-wsj-0-18.tagger";
+ //taggermodel = "src/main/resources/tbsl/models/left3words-wsj-0-18.tagger";
tagger = new MaxentTagger(taggermodel);
}
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 2011-05-04 10:00:34 UTC (rev 2783)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/SlotBuilder.java 2011-05-06 16:02:30 UTC (rev 2784)
@@ -11,7 +11,7 @@
private WordNet wordnet;
private String[] noun = {"NN","NNS","NNP","NNPS","NPREP"};
private String[] adjective = {"JJ","JJR","JJS","JJH"};
- private String[] verb = {"VB","VBD","VBG","VBN","VBP","VBZ","PASSIVE","PASSPART","VPASS","VPASSIN","GERUNDIN","VPREP"};
+ private String[] verb = {"VB","VBD","VBG","VBN","VBP","VBZ","PASSIVE","PASSPART","VPASS","VPASSIN","GERUNDIN","VPREP","WHEN","WHERE"};
private String[] preps = {"IN"};
public SlotBuilder() {
@@ -50,7 +50,9 @@
}
List<String> words = new ArrayList<String>();
words.add(token);
- words.addAll(wordnet.getBestSynonyms(token));
+ if (!pos.equals("NNP") && !pos.equals("NNPS")) {
+ words.addAll(wordnet.getBestSynonyms(token));
+ }
String tokenfluent = token.replaceAll(" ","");
String slotX = "x/" + type + "/";
@@ -125,10 +127,14 @@
}
}
if (pos.equals("PASSIVE")) {
- String[] passEntry = {token,
+ String[] passEntry1 = {token,
"(S DP[subj] (VP V:'" + token + "' DP[obj]))",
"<x,l1,t,[ l1:[|], l4:[ | SLOT_" + token + "(y,x) ] ],[(l2,x,subj,<<e,t>,t>),(l3,y,obj,<<e,t>,t>)],[ l2<l1,l3<l1,l4<scope(l2),l4<scope(l3) ],[" + slot + "]>"};
- result.add(passEntry);
+ String[] passEntry2 = {token,
+ "(S DP[wh] (VP DP[dp] V:'" + token + "'))",
+ "<x,l1,t,[ l1:[|], l4:[ | SLOT_" + token + "(y,x) ] ],[(l2,x,wh,<<e,t>,t>),(l3,y,dp,<<e,t>,t>)],[ l2<l1,l3<l1,l4<scope(l2),l4<scope(l3) ],[" + slot + "]>"};
+ result.add(passEntry1);
+ result.add(passEntry2);
}
else if (pos.equals("PASSPART")) {
String[] passpartEntry = {token,
@@ -164,7 +170,7 @@
"<x,l1,t,[ l1:[|], l4:[ | SLOT_" + token + "(x,y) ] ],[(l2,x,subj,<<e,t>,t>),(l3,y,obj,<<e,t>,t>)],[ l2<l1,l3<l1,l4<scope(l2),l4<scope(l3) ],[" + slot + "]>"};
result.add(passEntry);
}
- else if (pos.equals("VBD") || pos.equals("VBZ") || pos.equals("VBP")) {
+ else if (pos.equals("VBD") || pos.equals("VBZ") || pos.equals("VBP") || pos.equals("VB")) {
String[] vEntry = {token,
"(S DP[subj] (VP V:'" + token + "' DP[obj]))",
"<x,l1,t,[ l1:[|], l4:[ | SLOT_" + token + "(x,y) ] ],[(l2,x,subj,<<e,t>,t>),(l3,y,obj,<<e,t>,t>)],[ l2<l1,l3<l1,l4<scope(l2),l4<scope(l3) ],[" + slot + "]>"};
@@ -176,6 +182,32 @@
"<x,l1,t,[ l1:[ | SLOT_" + token + "(x,y) ] ],[(l2,y,dp,<<e,t>,t>)],[ l2=l1 ],[" + slot + "]>"};
result.add(gerEntry);
}
+ else if (pos.equals("WHEN")) {
+ String dateSlot = "SLOT_" + token + "/PROPERTY/" + token + "Date";
+ String tokenSlot = "SLOT_" + token + "/PROPERTY/" + token;
+ String[] whenEntry1 = {token,
+ "(S DP[subj] (VP V:'" + token + "'))",
+ "<x,l1,t,[ l1:[ ?y | SLOT_" + token + "(x,y) ] ],[(l2,x,subj,<<e,t>,t>)],[ l2=l1 ],[ " + dateSlot + " ]>"};
+ String[] whenEntry2 = {token,
+ "(S DP[subj] (VP V:'" + token + "' DP[obj]))",
+ "<x,l1,t,[ l1:[|], l4:[ ?z | SLOT_" + token + "(x,y), SLOT_date(x,z) ] ],[(l2,x,subj,<<e,t>,t>),(l3,y,obj,<<e,t>,t>)],[ l2<l1,l3<l1,l4<scope(l2),l4<scope(l3) ]," +
+ "[" + tokenSlot + ", SLOT_date/PROPERTY/date ]>"};
+ result.add(whenEntry1);
+ result.add(whenEntry2);
+ }
+ else if (pos.equals("WHERE")) {
+ String placeSlot = "SLOT_" + token + "/PROPERTY/" + token + "Place";
+ String tokenSlot = "SLOT_" + token + "/PROPERTY/" + token;
+ String[] whereEntry1 = {token,
+ "(S DP[subj] (VP V:'" + token + "'))",
+ "<x,l1,t,[ l1:[ ?y | SLOT_" + token + "(x,y) ] ],[(l2,x,subj,<<e,t>,t>)],[ l2=l1 ],[ " + placeSlot + " ]>"};
+ String[] whereEntry2 = {token,
+ "(S DP[subj] (VP V:'" + token + "' DP[obj]))",
+ "<x,l1,t,[ l1:[|], l4:[ ?z | SLOT_" + token + "(x,y), SLOT_place(x,z) ] ],[(l2,x,subj,<<e,t>,t>),(l3,y,obj,<<e,t>,t>)],[ l2<l1,l3<l1,l4<scope(l2),l4<scope(l3) ]," +
+ "[" + tokenSlot + ", SLOT_place/PROPERTY/place ]>"};
+ result.add(whereEntry1);
+ result.add(whereEntry2);
+ }
}
/* ADJECTIVES */
Modified: trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex
===================================================================
--- trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex 2011-05-04 10:00:34 UTC (rev 2783)
+++ trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex 2011-05-06 16:02:30 UTC (rev 2784)
@@ -1,12 +1,6 @@
-// TEST
-//------
-
- mushrooms || (NP N:'mushrooms') || <x,l1,<e,t>,[ l1:[ | mushroom(x) ] ], [],[],[]>
-
-
// TO BE
-// --------
+// ------
is || (S DP[subject] (VP V:'is' 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 DP[subject] (VP V:'is' ADJ[comp])) || <x, l1, t, [ l1:[ | x=y ]], [ (l2,x,subject,<<e,t>,t>), (l3,y,comp,<e,t>) ], [ l2=l1, l3=l2 ],[]>
@@ -24,7 +18,19 @@
is there || (S V:'is' C:'there' DP[dp]) || <x, l1, t, [ l1:[ | ] ], [ (l2,x,dp,<<e,t>,t>) ], [ l2=l1 ],[]>
+// 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 ],[]>
+ 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) ],[]>
+ are || (S V:'are' 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 ],[]>
+ were || (S V:'were' 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) ],[]>
+ were || (S V:'were' 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 ],[]>
+
+ did || (S V:'did' S*) || <x,l1,t,[ l1:[|] ],[],[],[]>
+
// IMPERATIVES
// ---------------------
@@ -80,18 +86,23 @@
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],[]>
+
// EMPTY STUFF
+// ------------
also || (VP ADV:'also' VP*) || <x,l1,t,[ l1:[|] ],[],[],[]>
+
// WH WORDS
// --------
- what || (DP WH:'what') || <x, l1, <<e,t>,t>, [ l1:[ ?x | ] ], [], [], []>
- which || (DP WH:'which') || <x, l1, <<e,t>,t>, [ l1:[ ?x | ] ], [], [], []>
+ what || (DP WH:'what') || <x, l1, <<e,t>,t>, [ l1:[ ?x | ] ], [], [], []>
+ which || (DP WH:'which') || <x, l1, <<e,t>,t>, [ l1:[ ?x | ] ], [], [], []>
how many || (DP WH:'how' ADJ:'many' NP[noun]) || <y, l1, <<e,t>,t>, [ l1:[ | l2:[ y | ] HOWMANY y l3:[|] ] ], [ (l4,y,noun,<e,t>) ], [ l4=l2 ],[]>
- who || (DP WH:'who') || <x, l1, <<e,t>,t>, [ l1:[ ?x | ] ], [], [], []>
+ who || (DP WH:'who') || <x, l1, <<e,t>,t>, [ l1:[ ?x | ] ], [], [], []>
+ when || (S WH:'when' S[s]) || <x, l1, <<e,t>,t>, [ l1:[ ?x | SLOT_p(y,x) ] ], [(l2,y,s,t)], [l2=l1], [ SLOT_p/PROPERTY/date ]>
+ where || (S WH:'where' S[s]) || <x, l1, <<e,t>,t>, [ l1:[ ?x | SLOT_p(y,x) ] ], [(l2,y,s,t)], [l2=l1], [ SLOT_p/PROPERTY/place ]>
// NEGATION
@@ -104,7 +115,7 @@
do not || (VP V:'do' NEG:'not' VP*) || <x,l2,t,[ l1:[ | NOT l2:[|] ] ],[],[],[]>
-// CONJUNCTION
+// COORDINATION
// ------------
and || (S S* CC:'and' S[s]) || <x,l1,t,[l1:[|]],[(l2,y,s,t)],[l1=l2],[]>
@@ -113,6 +124,8 @@
and || (VP VP* CC:'and' VP[vp]) || -
and || (ADJ ADJ* CC:'and' ADJ[adj]) || -
+ as well as || (NP NP* CC:'as' CC:'well' CC:'as' NP[np]) || <x,l1,<e,t>,[l1:[|]],[(l2,y,np,<e,t>)],[l1=l2],[]>
+
or || (S S* CC:'or' S[2]) || -
or || (DP DP* CC:'or' DP[2]) || -
or || (NP NP* CC:'or' NP[2]) || -
Modified: trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/POStest.java
===================================================================
--- trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/POStest.java 2011-05-04 10:00:34 UTC (rev 2783)
+++ trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/POStest.java 2011-05-06 16:02:30 UTC (rev 2784)
@@ -8,13 +8,27 @@
public static void main(String[] args) throws IOException, ClassNotFoundException {
+
POStagger tagger = new POStagger();
- String sentence = "give me all cities in Germany";
+ String sentence = "When did Nirvana record Nevermind?";
String tagged = tagger.tag(sentence);
System.out.println(tagged);
+
+// Tagger tagger = new Tagger("en");
+//
+// String s = "";
+//
+// String[] words = s.split(" ");
+// String[] tagged;
+//
+// tagged = tagger.tag(words);
+//
+// for (String string : tagged) {
+// System.out.println(string);
+// }
}
}
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 2011-05-04 10:00:34 UTC (rev 2783)
+++ trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/PatternMatchingTest.java 2011-05-06 16:02:30 UTC (rev 2784)
@@ -11,13 +11,13 @@
public static void main(String[] args) {
// TODO Auto-generated method stub
- String s = "was/VBD developed/VBN by/IN";
+ String s = "how/WRB many/JJ software/NN companies/NN are/VBP located/VBN in/IN New/NNP York/NNP";
- Pattern nprepPattern = Pattern.compile("(((is)|(are)|(was)|(were))/VB[A-Z]?.(\\w+)/VBN.by/IN)");
+ Pattern nprepPattern = Pattern.compile("\\s((\\w+)/NN[S]?\\s(\\w+))/NN[S]?");
Matcher m = nprepPattern.matcher(s);
while (m.find()) {
System.out.println("Found!");
- s = s.replaceFirst(m.group(1),m.group(7)+"/PASSIVE");
+ s = s.replaceFirst(m.group(1),m.group(2) + "_" + m.group(3));
}
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 2011-05-04 10:00:34 UTC (rev 2783)
+++ trunk/components-ext/src/test/java/org/dllearner/algorithm/tbsl/TemplateGenerationTest.java 2011-05-06 16:02:30 UTC (rev 2784)
@@ -54,7 +54,7 @@
* @param args
*/
public static void main(String[] args) {
- File file = new File("src/main/resources/tbsl/evaluation/dbpedia-train.xml");
+ File file = new File("src/main/resources/tbsl/evaluation/dbpedia-test-questions.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.
|
