Menu
▾
▴
[DL-Learner SVN] SF.net SVN: dl-learner:[2781] trunk/components-ext/src/main
|
From: <chr...@us...> - 2011-05-03 15:27:13
|
Revision: 2781
http://dl-learner.svn.sourceforge.net/dl-learner/?rev=2781&view=rev
Author: christinaunger
Date: 2011-05-03 15:27:07 +0000 (Tue, 03 May 2011)
Log Message:
-----------
Update SPARQL Template Generation (small fixes)
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/templator/SlotBuilder.java
trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex
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 2011-05-03 14:57:22 UTC (rev 2780)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/GrammarFilter.java 2011-05-03 15:27:07 UTC (rev 2781)
@@ -25,7 +25,6 @@
class GrammarFilter {
final static String[] NAMED_Strings = {"named", "called"};
- final static String NAME_PREDICATE = "SLOT.pred:title_name";
static ParseGrammar filter(String taggedinput,LTAGLexicon grammar,List<Integer> temps) {
@@ -100,10 +99,10 @@
try {
- TreeNode tree = c.construct("(DP NUM:'" + token + "' NP[noun])");
+ TreeNode tree = c.construct("NUM:'" + token + "'");
int gid = grammar.addTree(grammar.size(), new Pair<String,TreeNode>(token,tree),
- Collections.singletonList("<x,l1,<<e,t>,t>,[l1:[ x | count(x,c), equal(c," + token + ")]],[(l2,x,noun,<e,t>)],[l2=l1],[]>"));
+ Collections.singletonList("<x,l1,e,[l1:[ x | equal(x," + token + ")]],[],[],[]>"));
add(parseG, tree, gid-1, localID);
localID++;
@@ -230,8 +229,9 @@
rawNames += "DP:'" + split[i] + "' ";
}
semName = semName.substring(1);
- out.add(new Pair<String,String>("(NP NP* ADJ:'"+ w +"' " + rawNames + ")", "<x,l1,<e,t>,[ l1:[ | " + NAME_PREDICATE + "(x,'" + semName + "') ] ], [],[],[]>"));
- out.add(new Pair<String,String>("(DP DP* ADJ:'"+ w +"' " + rawNames + ")", "<x,l1,<<e,t>,t>,[ l1:[ | " + NAME_PREDICATE + "(x,'" + semName + "') ] ], [],[],[]>"));
+ out.add(new Pair<String,String>("(NP NP* ADJ:'"+ w +"' " + rawNames + ")", "<x,l1,<e,t>,[ l1:[ | SLOT_title(x,'" + semName + "') ] ], [],[],[ SLOT_title/PROPERTY/title^name ]>"));
+ out.add(new Pair<String,String>("(DP DP* ADJ:'"+ w +"' " + rawNames + ")", "<x,l1,<<e,t>,t>,[ l1:[ | SLOT_title(x,'" + semName + "') ] ], [],[],[ SLOT_title/PROPERTY/title^name ]>"));
+ out.add(new Pair<String,String>("(ADJ ADJ:'"+ w +"' " + rawNames + ")", "<x,l1,<e,t>,[ l1:[ | SLOT_title(x,'" + semName + "') ] ], [],[],[ SLOT_title/PROPERTY/title^name ]>"));
return out;
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 2011-05-03 14:57:22 UTC (rev 2780)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/ltag/parser/Preprocessor.java 2011-05-03 15:27:07 UTC (rev 2781)
@@ -56,6 +56,7 @@
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)");
m = compAdjPattern.matcher(condensedstring);
@@ -94,6 +95,10 @@
while (m.find()) {
condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2)+"/VPASSIN");
}
+ m = gerundinPattern.matcher(condensedstring);
+ while (m.find()) {
+ condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2)+"/GERUNDIN");
+ }
m = vprepPattern.matcher(condensedstring);
while (m.find()) {
condensedstring = condensedstring.replaceFirst(m.group(1),m.group(2)+"/VPREP");
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 2011-05-03 14:57:22 UTC (rev 2780)
+++ trunk/components-ext/src/main/java/org/dllearner/algorithm/tbsl/templator/SlotBuilder.java 2011-05-03 15:27:07 UTC (rev 2781)
@@ -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","VPREP"};
+ private String[] verb = {"VB","VBD","VBG","VBN","VBP","VBZ","PASSIVE","PASSPART","VPASS","VPASSIN","GERUNDIN","VPREP"};
private String[] preps = {"IN"};
public SlotBuilder() {
@@ -52,8 +52,9 @@
words.add(token);
words.addAll(wordnet.getBestSynonyms(token));
+ String tokenfluent = token.replaceAll(" ","");
String slotX = "x/" + type + "/";
- String slotP = "SLOT_" + token + "/" + type + "/";
+ String slotP = "SLOT_" + tokenfluent + "/" + type + "/";
for (Iterator<String> i = words.iterator(); i.hasNext();) {
String next = i.next().replaceAll(" ","_");
slotX += next; slotP += next;
@@ -72,12 +73,12 @@
/* DP */
String[] dpEntry = {token,
"(DP (NP " + treetoken + "))",
- "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + token + "(x) ] ],[],[],[" + slotP + "]>"};
+ "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + tokenfluent + "(x) ] ],[],[],[" + slotP + "]>"};
result.add(dpEntry);
/* NP */
String[] npEntry = {token,
"(NP " + treetoken + ")",
- "<x,l1,<e,t>,[ l1:[ | SLOT_" + token + "(x) ] ],[],[],[" + slotP + "]>"};
+ "<x,l1,<e,t>,[ l1:[ | SLOT_" + tokenfluent + "(x) ] ],[],[],[" + slotP + "]>"};
result.add(npEntry);
}
else if (pos.equals("NNP") || pos.equals("NNPS")) {
@@ -94,17 +95,17 @@
else if (pos.equals("NPREP")) {
/* DP */
String[] dpEntry1a = {token,
- "(DP (NP " + treetoken + " P:'of' DP[pobj]))",
- "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + token + "(y,x) ] ],[(l2,y,pobj,<<e,t>,t>)],[l2=l1],[" + slotP + "]>"};
+ "(DP (NP " + treetoken + " DP[pobj]))",
+ "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + tokenfluent + "(y,x) ] ],[(l2,y,pobj,<<e,t>,t>)],[l2=l1],[" + slotP + "]>"};
String[] dpEntry1b = {token,
- "(DP (NP " + treetoken + " P:'of' DP[pobj]))",
- "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + token + "(x), SLOT_of(x,y) ] ],[(l2,y,pobj,<<e,t>,t>)],[l2=l1],[" + slotP + "," + "SLOT_of/PROPERTY/" + "]>"};
+ "(DP (NP " + treetoken + " DP[pobj]))",
+ "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + tokenfluent + "(x), SLOT_of(x,y) ] ],[(l2,y,pobj,<<e,t>,t>)],[l2=l1],[" + slotP + "," + "SLOT_of/PROPERTY/" + "]>"};
String[] dpEntry2a = {token,
- "(DP DET[det] (NP " + treetoken + " P:'of' DP[pobj]))",
- "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + token + "(y,x) ] ],[(l2,y,pobj,<<e,t>,t>),(l3,x,det,e)],[l2=l1,l3=l1],[" + slotP + "]>"};
+ "(DP DET[det] (NP " + treetoken + " DP[pobj]))",
+ "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + tokenfluent + "(y,x) ] ],[(l2,y,pobj,<<e,t>,t>),(l3,x,det,e)],[l2=l1,l3=l1],[" + slotP + "]>"};
String[] dpEntry2b = {token,
- "(DP DET[det] (NP " + treetoken + " P:'of' DP[pobj]))",
- "<x,l1,<<e,t>,t>,[ l1:[ x | SLOT_" + token + "(x), SLOT_of(x,y) ] ],[(l2,y,pobj,<<e,t>,t>),(l3,x,det,e)],[l2=l1,l3=l1],[" + slotP + "," + "SLOT_of/PROPERTY/" + "]>"};
+ "(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/" + "]>"};
result.add(dpEntry1a);
result.add(dpEntry1b);
result.add(dpEntry2a);
@@ -147,6 +148,16 @@
"<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("GERUNDIN")) {
+ String[] gerundinEntry1 = {token,
+ "(NP NP* V:'" + token + "' DP[obj]))",
+ "<x,l1,t,[ l1:[ | SLOT_" + token + "(x,y) ] ],[(l2,y,obj,<<e,t>,t>)],[ l2=l1 ],[" + slot + "]>"};
+ String[] gerundinEntry2 = {token,
+ "(ADJ V:'" + token + "' DP[obj]))",
+ "<x,l1,<e,t>,[ l1:[ | SLOT_" + token + "(x,y) ] ],[(l2,y,obj,<<e,t>,t>)],[ l2=l1 ],[" + slot + "]>"};
+ result.add(gerundinEntry1);
+ result.add(gerundinEntry2);
+ }
else if (pos.equals("VPREP")) {
String[] passEntry = {token,
"(S DP[subj] (VP V:'" + token + "' DP[obj]))",
Modified: trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex
===================================================================
--- trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex 2011-05-03 14:57:22 UTC (rev 2780)
+++ trunk/components-ext/src/main/resources/tbsl/lexicon/english.lex 2011-05-03 15:27:07 UTC (rev 2781)
@@ -76,8 +76,10 @@
the most || (DET DET:'the' DET:'most') || <y, l1, e, [ l1:[ | l2:[ y | ] THEMOST y l3:[|] ] ], [], [],[]>
the least || (DET DET:'the' DET:'least') || <y, l1, e, [ l1:[ | l2:[ y | ] THELEAST y l3:[|] ] ], [], [],[]>
+ // COUNT
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
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
