/**
* Close the index
*
* @throws IOException
* @throws CorruptIndexException
*/
public void close() throws CorruptIndexException, IOException {
// Signal to the listener that we're done indexing and closing the index (which might take a
// while)
getListener().indexEnd();
getListener().closeStart();
searcher.getIndexStructure().addToTokenCount(getListener().getTokensProcessed());
searcher.getIndexStructure().writeMetadata();
searcher.close();
// Signal that we're completely done now
getListener().closeEnd();
getListener().indexerClosed();
}
/**
* How many more documents should we process?
*
* @return the number of documents
*/
public synchronized int docsToDoLeft() {
try {
if (maxNumberOfDocsToIndex < 0) return maxNumberOfDocsToIndex;
int docsDone = searcher.getWriter().numDocs();
return Math.max(0, maxNumberOfDocsToIndex - docsDone);
} catch (Exception e) {
throw new RuntimeException(e);
}
}
/**
* Index a document from a Reader, using the specified type of DocIndexer
*
* @param documentName some (preferably unique) name for this document (for example, the file name
* or path)
* @param reader where to index from
* @throws Exception
*/
private void indexReader(String documentName, Reader reader) throws Exception {
getListener().fileStarted(documentName);
int docsDoneBefore = searcher.getWriter().numDocs();
long tokensDoneBefore = getListener().getTokensProcessed();
DocIndexer docIndexer = createDocIndexer(documentName, reader);
docIndexer.index();
getListener().fileDone(documentName);
int docsDoneAfter = searcher.getWriter().numDocs();
if (docsDoneAfter == docsDoneBefore) {
System.err.println("*** Warning, couldn't index " + documentName + "; wrong format?");
}
long tokensDoneAfter = getListener().getTokensProcessed();
if (tokensDoneAfter == tokensDoneBefore) {
System.err.println("*** Warning, no words indexed in " + documentName + "; wrong format?");
}
}
@Override
public Response handle() throws BlsException {
Collection<String> indices = searchMan.getAllAvailableIndices(user.getUserId());
DataObjectMapAttribute doIndices = new DataObjectMapAttribute("index", "name");
// DataObjectList doIndices = new DataObjectList("index");
for (String indexName : indices) {
DataObjectMapElement doIndex = new DataObjectMapElement();
Searcher searcher = searchMan.getSearcher(indexName);
IndexStructure struct = searcher.getIndexStructure();
doIndex.put("displayName", struct.getDisplayName());
doIndex.put("status", searchMan.getIndexStatus(indexName));
String documentFormat = struct.getDocumentFormat();
if (documentFormat != null && documentFormat.length() > 0)
doIndex.put("documentFormat", documentFormat);
doIndex.put("timeModified", struct.getTimeModified());
if (struct.getTokenCount() > 0) doIndex.put("tokenCount", struct.getTokenCount());
doIndices.put(indexName, doIndex);
}
DataObjectMapElement doUser = new DataObjectMapElement();
doUser.put("loggedIn", user.isLoggedIn());
if (user.isLoggedIn()) doUser.put("id", user.getUserId());
doUser.put(
"canCreateIndex", user.isLoggedIn() ? searchMan.canCreateIndex(user.getUserId()) : false);
DataObjectMapElement response = new DataObjectMapElement();
response.put("blacklabBuildTime", Searcher.getBlackLabBuildTime());
response.put("indices", doIndices);
response.put("user", doUser);
response.put("helpPageUrl", servlet.getServletContext().getContextPath() + "/help");
if (debugMode) {
response.put("cacheStatus", searchMan.getCacheStatusDataObject());
}
Response responseObj = new Response(response);
responseObj.setCacheAllowed(false); // You can create/delete indices, don't cache the list
return responseObj;
}
/**
* Construct Indexer
*
* @param directory the main BlackLab index directory
* @param create if true, creates a new index; otherwise, appends to existing index
* @param docIndexerClass how to index the files, or null to autodetect
* @param indexTemplateFile JSON file to use as template for index structure / metadata (if
* creating new index)
* @throws DocumentFormatException if no DocIndexer was specified and autodetection failed
* @throws IOException
*/
public Indexer(
File directory,
boolean create,
Class<? extends DocIndexer> docIndexerClass,
File indexTemplateFile)
throws DocumentFormatException, IOException {
this.docIndexerClass = docIndexerClass;
searcher = Searcher.openForWriting(directory, create, indexTemplateFile);
if (!create) searcher.getIndexStructure().setModified();
if (this.docIndexerClass == null) {
// No DocIndexer supplied; try to detect it from the index
// metadata.
String formatId = searcher.getIndexStructure().getDocumentFormat();
if (formatId != null && formatId.length() > 0)
setDocIndexer(DocumentFormats.getIndexerClass(formatId));
else {
throw new DocumentFormatException("Cannot detect document format for index!");
}
}
metadataFieldTypeTokenized = new FieldType();
metadataFieldTypeTokenized.setStored(true);
metadataFieldTypeTokenized.setIndexed(true);
metadataFieldTypeTokenized.setTokenized(true);
metadataFieldTypeTokenized.setOmitNorms(true); // @@@ <-- depending on setting?
metadataFieldTypeTokenized.setStoreTermVectors(true);
metadataFieldTypeTokenized.setStoreTermVectorPositions(true);
metadataFieldTypeTokenized.setStoreTermVectorOffsets(true);
metadataFieldTypeTokenized.freeze();
metadataFieldTypeUntokenized = new FieldType(metadataFieldTypeTokenized);
metadataFieldTypeUntokenized.setTokenized(false);
metadataFieldTypeUntokenized.freeze();
}
/**
* Get the IndexWriter we're using.
*
* <p>Useful if e.g. you want to access FSDirectory.
*
* @return the IndexWriter
*/
protected IndexWriter getWriter() {
return searcher.getWriter();
}
/**
* Get our index directory
*
* @return the index directory
*/
public File getIndexLocation() {
return searcher.getIndexDirectory();
}
public ContentStore getContentStore(String fieldName) {
return searcher.getContentStore(fieldName);
}
/**
* Add a list of tokens to a forward index
*
* @param fieldName what forward index to add this to
* @param tokens the tokens to add
* @param posIncr position increment associated with each token
* @return the id assigned to the content
*/
public int addToForwardIndex(String fieldName, List<String> tokens, List<Integer> posIncr) {
ForwardIndex forwardIndex = searcher.getForwardIndex(fieldName);
if (forwardIndex == null) throw new RuntimeException("No forward index for field " + fieldName);
return forwardIndex.addDocument(tokens, posIncr);
}
/**
* Updates the specified Document in the index.
*
* @param term how to find the document to update
* @param document the updated document
* @throws CorruptIndexException
* @throws IOException
*/
public void update(Term term, Document document) throws CorruptIndexException, IOException {
searcher.getWriter().updateDocument(term, document);
getListener().luceneDocumentAdded();
}
/**
* Add a Lucene document to the index
*
* @param document the document to add
* @throws CorruptIndexException
* @throws IOException
*/
public void add(Document document) throws CorruptIndexException, IOException {
searcher.getWriter().addDocument(document, searcher.getAnalyzer());
getListener().luceneDocumentAdded();
}
/**
* Call this to roll back any changes made to the index this session. Calling close() will
* automatically commit any changes. If you call this method, then call close(), no changes will
* be committed.
*/
public void rollback() {
getListener().rollbackStart();
searcher.rollback();
getListener().rollbackEnd();
}
/**
* Set the collator to use for sorting (passed to ForwardIndex to keep a sorted list of terms).
* Defaults to English collator.
*
* @param collator the collator
* @deprecated use Searcher.setDefaultCollator()
*/
@Deprecated
public static void setCollator(Collator collator) {
Searcher.setDefaultCollator(collator);
}
