org.apache.avalon.framework.context.ContextException
org.apache.cocoon.components.search.*
org.apache.lucene.document.Document
org.apache.lucene.index.*
org.apache.lucene.document.*
org.apache.lucene.store.*
java.io.*
java.util.*
java.net.*
File workDir = null;
/** Contextualize this class */
public void contextualize(Context context) throws ContextException {
super.contextualize( context );
workDir = (File) context.get(Constants.CONTEXT_WORK_DIR);
}
Directory directory;
void init() throws ProcessingException {
try {
directory = LuceneCocoonHelper.getDirectory( new File( workDir, "index" ), false );
} catch (Exception e) {
throw new ProcessingException( "Exception in init()!", e );
}
}
int numDocs() throws ProcessingException {
try {
IndexReader reader = IndexReader.open( directory );
int num_docs;
num_docs = reader.numDocs();
return num_docs;
} catch (Exception e) {
throw new ProcessingException( "Exception in numDocs()!", e );
}
}
Map allDocuments() throws ProcessingException {
try {
IndexReader reader = IndexReader.open( directory );
HashMap fieldsStatistic = new HashMap();
for (int i = 0; i < reader.maxDoc(); i++ ) {
if (!reader.isDeleted(i)) {
Document document = reader.document(i);
Enumeration fields = document.fields();
while (fields.hasMoreElements()) {
Field f = (Field)fields.nextElement();
String name = f.name();
String value = f.stringValue();
if (value == null) value = "--void--";
String fieldStatistic = name + "/" + value;
if (fieldsStatistic.get( fieldStatistic ) == null) {
fieldsStatistic.put( fieldStatistic, new Integer(1) );
} else {
Integer sum = (Integer)fieldsStatistic.get( fieldStatistic );
int sum_plus = sum.intValue() + 1;
fieldsStatistic.put( fieldStatistic, new Integer( sum_plus ) );
}
}
}
}
return fieldsStatistic;
//map.keySet();
} catch (Exception e) {
throw new ProcessingException( "Exception allDocuments()!", e );
}
}
Map allTerms() throws ProcessingException {
try {
IndexReader reader = IndexReader.open( directory );
TermEnum te = reader.terms();
HashMap all_fields = new HashMap();
while (te.next()) {
Term term = te.term();
int docfreq = te.docFreq();
String field = term.field();
if (field != null) {
if (all_fields.containsKey( field )) {
Integer sum = (Integer)all_fields.get( field );
int sum_plus = sum.intValue() + docfreq;
all_fields.put( field, new Integer( sum_plus ) );
} else {
all_fields.put( field, new Integer( docfreq ) );
}
}
}
te.close();
return all_fields;
} catch (Exception e) {
throw new ProcessingException( "Exception allDocuments()!", e );
}
}
Map sort( Map map ) {
TreeMap treeMap = new TreeMap( map );
return treeMap;
}
init();
Index Statistics
Statistics:
Total Count Of Documents
String.valueOf(numDocs())
| Count Of Terms | Fieldname/Fieldvalue |
Map all_docs = sort(allDocuments());
Iterator it1 = all_docs.keySet().iterator();
while (it1.hasNext()) {
String k = (String)it1.next();
Integer v = (Integer)all_docs.get( k );
| v.toString() |
k |
}
All Terms
| Count Of Terms | Term |
Map all_terms = sort(allTerms());
Iterator it2 = all_terms.keySet().iterator();
while (it2.hasNext()) {
String k = (String)it2.next();
Integer v = (Integer)all_terms.get( k );
| v.toString() |
k |
}