Author: pkluegl Date: Tue Jun 18 12:32:48 2013 New Revision: 1494117 URL: http://svn.apache.org/r1494117 Log: UIMA-3005 - wildcard is actually slower than the token-based wildcard because of the CONTAINS condition - reverted changes
Modified: uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/AnnotationRule.java uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ExpansionRule.java uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ShiftAllRule.java Modified: uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/AnnotationRule.java URL: http://svn.apache.org/viewvc/uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/AnnotationRule.java?rev=1494117&r1=1494116&r2=1494117&view=diff ============================================================================== --- uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/AnnotationRule.java (original) +++ uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/AnnotationRule.java Tue Jun 18 12:32:48 2013 @@ -100,8 +100,8 @@ public class AnnotationRule extends Trab ruleString += frontBoundaryItem + after + " ("; // We include all tokens between the boundaries. - //ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster - ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ... but faster + ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster +// ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ... but faster // Check, if the rear item should be included and mark all tokens between the brackets as the Modified: uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ExpansionRule.java URL: http://svn.apache.org/viewvc/uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ExpansionRule.java?rev=1494117&r1=1494116&r2=1494117&view=diff ============================================================================== --- uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ExpansionRule.java (original) +++ uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ExpansionRule.java Tue Jun 18 12:32:48 2013 @@ -89,9 +89,9 @@ public class ExpansionRule extends Shift ruleString += frontBoundaryItem + after + " ("; // We include all tokens between the boundaries. - //ruleString += "ANY*{-PARTOF(" + annotation.getType().getShortName() + ")} "; // like ANY*? but + ruleString += "ANY*{-PARTOF(" + annotation.getType().getShortName() + ")} "; // like ANY*? but // faster - ruleString += "#{-CONTAINS(" + annotation.getType().getShortName() + ")} "; +// ruleString += "#{-CONTAINS(" + annotation.getType().getShortName() + ")} "; // these strings will be attached to the brackets String mark; @@ -134,8 +134,8 @@ public class ExpansionRule extends Shift ruleString += "(" + annotation.getType().getShortName() + unmark + " "; // We include all tokens between the boundaries. -// ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster - ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ... but faster + ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster +// ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ... but faster // these strings will be attached to the brackets String mark; Modified: uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ShiftAllRule.java URL: http://svn.apache.org/viewvc/uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ShiftAllRule.java?rev=1494117&r1=1494116&r2=1494117&view=diff ============================================================================== --- uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ShiftAllRule.java (original) +++ uima/sandbox/ruta/trunk/ruta-ep-textruler/src/main/java/org/apache/uima/ruta/textruler/learner/trabal/ShiftAllRule.java Tue Jun 18 12:32:48 2013 @@ -96,10 +96,10 @@ public class ShiftAllRule extends Shifti ruleString += frontBoundaryItem + after + " ("; // We include all tokens between the boundaries. -// ruleString += "ANY*{-PARTOF(" + errorBoundaryItem + ")} " + unmark + " ANY*{-PARTOF(" -// + rearBoundaryItem + ")}"; // like ANY*? but faster - ruleString += "#{-CONTAINS(" + errorBoundaryItem + ")} " + unmark + " #{-CONTAINS(" + ruleString += "ANY*{-PARTOF(" + errorBoundaryItem + ")} " + unmark + " ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster +// ruleString += "#{-CONTAINS(" + errorBoundaryItem + ")} " + unmark + " #{-CONTAINS(" +// + rearBoundaryItem + ")}"; // like ANY*? but faster // Check, if the rear item should be included and mark all tokens between the brackets as @@ -114,8 +114,8 @@ public class ShiftAllRule extends Shifti private void compileShiftToRight(boolean frontItemInBorders, boolean rearItemInBorders, String before, String after, String action, String unmark) { // The old annotation begins before the new annotation -// ruleString += unmark + " ANY*{-PARTOF(" + frontBoundaryItem + ")} "; - ruleString += unmark + " #{-CONTAINS(" + frontBoundaryItem + ")} "; + ruleString += unmark + " ANY*{-PARTOF(" + frontBoundaryItem + ")} "; +// ruleString += unmark + " #{-CONTAINS(" + frontBoundaryItem + ")} "; // If the front item is part of the future annotation, it has to be included in the // brackets. @@ -125,8 +125,8 @@ public class ShiftAllRule extends Shifti ruleString += frontBoundaryItem + after + " ("; // We include all tokens between the boundaries. -// ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster - ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ANY*? but faster + ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster +// ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ANY*? but faster @@ -155,8 +155,8 @@ public class ShiftAllRule extends Shifti else ruleString += frontBoundaryItem + unmark + " ("; -// ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster - ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ANY*? but faster + ruleString += "ANY*{-PARTOF(" + rearBoundaryItem + ")}"; // like ANY*? but faster +// ruleString += "#{-CONTAINS(" + rearBoundaryItem + ")}"; // like ANY*? but faster // Check, if the rear item should be included and mark all tokens between the brackets as the // new annotation.