added recognition of multi-word terms in vocabulary matching

this makes the PND usable: it is now possible to recognize persons and
navigate with a 'Persons' facet.

source/net/yacy/document/Autotagging.java

@@ -149,18 +149,31 @@ public class Autotagging {
         final WordTokenizer tokens = new WordTokenizer(new ByteArrayInputStream(UTF8.getBytes(text)), LibraryProvider.dymLib);
         String tag;
         while (tokens.hasMoreElements()) {
-            tag = getTagFromWord(tokens.nextElement().toString()).toString();
+            tag = getTagFromTerm(tokens.nextElement().toString()).toString();
             if (tag != null) as.add(tag);
         }
         return as;
     }
+    
+    public int size() {
+    	return this.vocabularies.size();
+    }
+    
+    /**
+     * maximum number of compound tags (number of words in one tag)
+     * @return
+     */
+    public int getMaxWordsInTerm() {
+    	//TODO: calculate from database
+    	return 4;
+    }
 
-    public Tagging.Metatag getTagFromWord(String word) {
+    public Tagging.Metatag getTagFromTerm(String term) {
         if (this.vocabularies.isEmpty()) return null;
         Tagging.Metatag tag;
-        word = Tagging.normalizeWord(word);
+        term = Tagging.normalizeWord(term);
         for (Map.Entry<String, Tagging> v: this.vocabularies.entrySet()) {
-            tag = v.getValue().getMetatagFromSynonym(this.prefixChar, word);
+            tag = v.getValue().getMetatagFromSynonym(this.prefixChar, term);
             if (tag != null) return tag;
         }
         return null;

source/net/yacy/document/Condenser.java

@@ -299,6 +299,8 @@ public final class Condenser {
         assert is != null;
         final Set<String> currsentwords = new HashSet<String>();
         String word = "";
+        String[] wordcache = new String[LibraryProvider.autotagging.getMaxWordsInTerm() - 1];
+        for (int i = 0; i < wordcache.length; i++) wordcache[i] = "";
         String k;
         Tagging.Metatag tag;
         int wordlen;
@@ -312,7 +314,8 @@ public final class Condenser {
         int wordInSentenceCounter = 1;
         boolean comb_indexof = false, last_last = false, last_index = false;
         final Map<StringBuilder, Phrase> sentences = new HashMap<StringBuilder, Phrase>(100);
-
+        if (LibraryProvider.autotagging.size() == 0) doAutotagging = false;
+        
         // read source
         final WordTokenizer wordenum = new WordTokenizer(is, meaningLib);
         try {
@@ -323,18 +326,33 @@ public final class Condenser {
 
 	            // get tags from autotagging
 	            if (doAutotagging) {
-	                tag = LibraryProvider.autotagging.getTagFromWord(word);
+	            	for (int wordc = 1; wordc <= wordcache.length + 1; wordc++) {
-	                if (tag != null) {
+	            		// wordc is number of words that are tested
-	                    Set<Tagging.Metatag> tagset = this.tags.get(tag.getVocabularyName());
+	            		StringBuilder sb = new StringBuilder();
-	                    if (tagset == null) {
+	            		if (wordc == 1) {
-	                        tagset = new HashSet<Tagging.Metatag>();
+	            			sb.append(word);
-	                        tagset.add(tag);
+	            		} else {
-	                        this.tags.put(tag.getVocabularyName(), tagset);
+	            			for (int w = 0; w < wordc - 1; w++) {
-	                    } else {
+	            				sb.append(wordcache[wordcache.length - wordc + w + 1]).append(' ');
+	            			}
+	            			sb.append(word);
+	            		}
+	            		String testterm = sb.toString().trim();
+	            		//System.out.println("Testing: " + testterm);
+		                tag = LibraryProvider.autotagging.getTagFromTerm(testterm);
+		                if (tag != null) {
+		                    Set<Tagging.Metatag> tagset = this.tags.get(tag.getVocabularyName());
+		                    if (tagset == null) {
+		                        tagset = new HashSet<Tagging.Metatag>();
+		                        this.tags.put(tag.getVocabularyName(), tagset);
+		                    }
 	                        tagset.add(tag);
-	                    }
+		                }
-	                }
+	            	}
 	            }
+	            // shift wordcache
+	            System.arraycopy(wordcache, 1, wordcache, 0, wordcache.length - 1);
+	            wordcache[wordcache.length - 1] = word;
 
 	            // distinguish punctuation and words
 	            wordlen = word.length();