org.dspace.search
Class DSIndexer

java.lang.Object
  org.dspace.search.DSIndexer

public class DSIndexer
extends Object

DSIndexer contains the methods that index Items and their metadata, collections, communities, etc. It is meant to either be invoked from the command line (see dspace/bin/index-all) or via the indexContent() methods within DSpace. As of 1.4.2 this class has new incremental update of index functionality and better detection of locked state thanks to Lucene 2.1 moving write.lock. It will attempt to attain a lock on the index in the event that an update is requested and will wait a maximum of 30 seconds (a worst case scenario) to attain the lock before giving up and logging the failure to log4j and to the DSpace administrator email account. The Administrator can choose to run DSIndexer in a cron that repeats regularly, a failed attempt to index from the UI will be "caught" up on in that cron.

Author:

Mark Diggory, Graham Triggs

Constructor Summary

DSIndexer()

Method Summary

static void	cleanIndex(Context context) Iterates over all documents in the Lucene index and verifies they are in database, if not, they are removed.
static void	createIndex(Context c) create full index - wiping old index
static void	indexContent(Context context, DSpaceObject dso) If the handle for the "dso" already exists in the index, and the "dso" has a lastModified timestamp that is newer than the document in the index then it is updated, otherwise a new document is added.
static void	indexContent(Context context, DSpaceObject dso, boolean force) If the handle for the "dso" already exists in the index, and the "dso" has a lastModified timestamp that is newer than the document in the index then it is updated, otherwise a new document is added.
static void	main(String[] args) When invoked as a command-line tool, creates, updates, removes content from the whole index
static void	optimizeIndex(Context c) Optimize the existing index.
static void	reIndexContent(Context context, DSpaceObject dso) reIndexContent removes something from the index, then re-indexes it
static void	unIndexContent(Context context, DSpaceObject dso) unIndex removes an Item, Collection, or Community only works if the DSpaceObject has a handle (uses the handle for its unique ID)
static void	unIndexContent(Context context, String handle) Unindex a Docment in the Lucene Index.
static void	updateIndex(Context context) Iterates over all Items, Collections and Communities.
static void	updateIndex(Context context, boolean force) Iterates over all Items, Collections and Communities.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DSIndexer

public DSIndexer()

Method Detail

indexContent

public static void indexContent(Context context,
                                DSpaceObject dso)
                         throws SQLException

If the handle for the "dso" already exists in the index, and the "dso" has a lastModified timestamp that is newer than the document in the index then it is updated, otherwise a new document is added.

Parameters:

context - Users Context

dso - DSpace Object (Item, Collection or Community

Throws:

SQLException

IOException

indexContent

public static void indexContent(Context context,
                                DSpaceObject dso,
                                boolean force)
                         throws SQLException

If the handle for the "dso" already exists in the index, and the "dso" has a lastModified timestamp that is newer than the document in the index then it is updated, otherwise a new document is added.

Parameters:

context - Users Context

dso - DSpace Object (Item, Collection or Community

force - Force update even if not stale.

Throws:

SQLException

IOException

unIndexContent

public static void unIndexContent(Context context,
                                  DSpaceObject dso)
                           throws SQLException,
                                  IOException

unIndex removes an Item, Collection, or Community only works if the DSpaceObject has a handle (uses the handle for its unique ID)

Parameters:

context -

dso - DSpace Object, can be Community, Item, or Collection

Throws:

SQLException

IOException

unIndexContent

public static void unIndexContent(Context context,
                                  String handle)
                           throws SQLException,
                                  IOException

Unindex a Docment in the Lucene Index.

Parameters:

context -

handle -

Throws:

SQLException

IOException

reIndexContent

public static void reIndexContent(Context context,
                                  DSpaceObject dso)
                           throws SQLException,
                                  IOException

reIndexContent removes something from the index, then re-indexes it

Parameters:

context - context object

dso - object to re-index

Throws:

SQLException

IOException

createIndex

public static void createIndex(Context c)
                        throws SQLException,
                               IOException

create full index - wiping old index

Parameters:

c - context to use

Throws:

SQLException

IOException

optimizeIndex

public static void optimizeIndex(Context c)
                          throws SQLException,
                                 IOException

Optimize the existing index. Iimportant to do regularly to reduce filehandle usage and keep performance fast!

Parameters:

c - Users Context

Throws:

SQLException

IOException

main

public static void main(String[] args)
                 throws SQLException,
                        IOException

When invoked as a command-line tool, creates, updates, removes content from the whole index

Parameters:

args - the command-line arguments, none used

Throws:

IOException

SQLException

updateIndex

public static void updateIndex(Context context)

Iterates over all Items, Collections and Communities. And updates them in the index. Uses decaching to control memory footprint. Uses indexContent and isStale ot check state of item in index.

Parameters:

context -

updateIndex

public static void updateIndex(Context context,
                               boolean force)

Iterates over all Items, Collections and Communities. And updates them in the index. Uses decaching to control memory footprint. Uses indexContent and isStale ot check state of item in index. At first it may appear counterintuitive to have an IndexWriter/Reader opened and closed on each DSO. But this allows the UI processes to step in and attain a lock and write to the index even if other processes/jvms are running a reindex.

Parameters:

context -

force -

cleanIndex

public static void cleanIndex(Context context)
                       throws IOException,
                              SQLException

Iterates over all documents in the Lucene index and verifies they are in database, if not, they are removed.

Parameters:

context -

Throws:

IOException

SQLException