pull #17 changes adopted to implement a whole word check on the entir…

…e keyword, including whitespaces.
robert-bor · Sep 22, 2015 · dc27d6e · dc27d6e
1 parent 76ae822
commit dc27d6e
Show file tree

Hide file tree

Showing 3 changed files with 141 additions and 98 deletions.
diff --git a/src/main/java/org/ahocorasick/trie/Trie.java b/src/main/java/org/ahocorasick/trie/Trie.java
@@ -80,6 +80,10 @@ public Collection<Emit> parseText(CharSequence text) {
             removePartialMatches(text, collectedEmits);
         }
 
+        if (trieConfig.isOnlyWholeWordsWhiteSpaceSeparated()) {
+            removePartialMatchesWhiteSpaceSeparated(text, collectedEmits);
+        }
+
         if (!trieConfig.isAllowOverlaps()) {
             IntervalTree intervalTree = new IntervalTree((List<Intervalable>)(List<?>)collectedEmits);
             intervalTree.removeOverlaps((List<Intervalable>) (List<?>) collectedEmits);
@@ -161,6 +165,21 @@ private void removePartialMatches(CharSequence searchText, List<Emit> collectedE
 		}
 	}
 
+    private void removePartialMatchesWhiteSpaceSeparated(CharSequence searchText, List<Emit> collectedEmits) {
+        long size = searchText.length();
+        List<Emit> removeEmits = new ArrayList<>();
+        for (Emit emit : collectedEmits) {
+            if ((emit.getStart() == 0 || Character.isWhitespace(searchText.charAt(emit.getStart() - 1))) &&
+                (emit.getEnd() + 1 == size || Character.isWhitespace(searchText.charAt(emit.getEnd() + 1)))) {
+                continue;
+            }
+            removeEmits.add(emit);
+        }
+        for (Emit removeEmit : removeEmits) {
+            collectedEmits.remove(removeEmit);
+        }
+    }
+
     private State getState(State currentState, Character character) {
         State newCurrentState = currentState.nextState(character);
         while (newCurrentState == null) {
@@ -237,6 +256,11 @@ public TrieBuilder onlyWholeWords() {
             return this;
         }
 
+        public TrieBuilder onlyWholeWordsWhiteSpaceSeparated() {
+            this.trieConfig.setOnlyWholeWordsWhiteSpaceSeparated(true);
+            return this;
+        }
+
         public TrieBuilder addKeyword(String keyword) {
             trie.addKeyword(keyword);
             return this;

diff --git a/src/main/java/org/ahocorasick/trie/TrieConfig.java b/src/main/java/org/ahocorasick/trie/TrieConfig.java
@@ -6,6 +6,8 @@ public class TrieConfig {
 
     private boolean onlyWholeWords = false;
 
+    private boolean onlyWholeWordsWhiteSpaceSeparated = false;
+
     private boolean caseInsensitive = false;
 
     private boolean stopOnHit = false;
@@ -30,6 +32,12 @@ public void setOnlyWholeWords(boolean onlyWholeWords) {
         this.onlyWholeWords = onlyWholeWords;
     }
 
+    public boolean isOnlyWholeWordsWhiteSpaceSeparated() { return onlyWholeWordsWhiteSpaceSeparated; }
+
+    public void setOnlyWholeWordsWhiteSpaceSeparated(boolean onlyWholeWordsWhiteSpaceSeparated) {
+        this.onlyWholeWordsWhiteSpaceSeparated = onlyWholeWordsWhiteSpaceSeparated;
+    }
+
     public boolean isCaseInsensitive() {
         return caseInsensitive;
     }