// Helper method to transform an IndexTank Document to a Lucene Document
private static org.apache.lucene.document.Document asLuceneDocument(Document itd) {
org.apache.lucene.document.Document doc = new org.apache.lucene.document.Document();
for (String field : itd.getFieldNames()) {
doc.add(new Field(field, itd.getField(field), Field.Store.NO, Field.Index.ANALYZED));
}
return doc;
}
/** @inheritDoc */
public synchronized void add(final String docId, final Document itdoc) {
if (null == docId) {
logger.error("No documentId specified. Ignoring addition.");
return;
}
org.apache.lucene.document.Document doc = asLuceneDocument(itdoc);
org.apache.lucene.document.Field docidPayloadField =
new org.apache.lucene.document.Field(
LsiIndex.PAYLOAD_TERM_FIELD, docId, Field.Store.NO, Field.Index.ANALYZED);
doc.add(docidPayloadField);
doc.add(new Field("documentId", docId, Field.Store.NO, Field.Index.NOT_ANALYZED));
try {
if (logger.isDebugEnabled()) {
logger.debug("Adding document with docId=" + docId + ". Doc is " + itdoc.getFieldNames());
}
writer.updateDocument(docIdTerm(docId), doc);
} catch (IOException e) {
logger.error(e);
}
}
/**
* @see AbstractDocumentSearcher#search(Query query, int start, int limit, int
* scoringFunctionIndex, Map<String, String> extraParameters).
* @param extraParameters: It will process 'fetch_fields', 'snippet_fields' and 'snippet_type'.
* 'fetch_fields' and 'snippet_fields' are comma-separated lists of field names to fetch an
* snippet. 'snippet_type' can be either 'html' or 'lines'. 'html' is the default.
*/
@Override
public SearchResults search(
Query query,
int start,
int limit,
int scoringFunctionIndex,
Map<String, String> extraParameters)
throws InterruptedException {
// call delegate searcher
SearchResults results =
this.delegate.search(query, start, limit, scoringFunctionIndex, extraParameters);
long startTime = System.currentTimeMillis();
String[] fetchFields = parseFields(extraParameters, "fetch");
String[] snippetFields = parseFields(extraParameters, "snippet");
Set<TermQuery> positiveTerms = query.getRoot().getPositiveTerms();
// find out which snippeter type is the right one for this query
String snType = extraParameters.get("snippet_type");
Snippeter sn = null;
if (null == snType || "html".equalsIgnoreCase(snType)) {
sn = this.snippeters.get(SnippeterType.HTML_AWARE);
} else if ("lines".equalsIgnoreCase(snType)) {
sn = this.snippeters.get(SnippeterType.LINE_AWARE);
} else {
throw new IllegalArgumentException("snippet_type has to be either 'html' or 'lines'");
}
if (fetchFields.length + snippetFields.length > 0) {
for (SearchResult result : results.getResults()) {
Document data = storage.getDocument(result.getDocId());
// fetch fields
for (String field : fetchFields) {
// handle '*', as a fetch all
if ("*".equals(field.trim())) {
// assume we get the actual fields, not a copy.
result.getFields().putAll(data.asMap());
break;
}
String text = data.getField(field);
if (null != text) {
result.setField(field, text);
}
}
// snippet fields
for (String field : snippetFields) {
String text = data.getField(field);
if (null != text) {
result.setField("snippet_" + field, sn.snippet(positiveTerms, field, text));
}
}
}
}
long endTime = System.currentTimeMillis();
logger.debug("(search) fetching & snippeting took: " + (endTime - startTime) + " ms.");
return results;
}
