org.dlese.dpc.index.writer
Class FileIndexingServiceWriter

java.lang.Object
  org.dlese.dpc.index.writer.FileIndexingServiceWriter

All Implemented Interfaces:

DocWriter

Direct Known Subclasses:

ErrorFileIndexingWriter, XMLFileIndexingWriter

public abstract class FileIndexingServiceWriter
extends Object
implements DocWriter

Abstract class for creating customized Lucene Documents for different file formats such as DLESE-IMS, ADN-item, ADN-collection, etc. Concrete sub-classes may be used with a FileIndexingService to enable automatic updating of the index whenever changes in the source file are made. This class, along with the FileIndexingService, may be used with a SimpleLuceneIndex to provide simple search support over files.

Note: after creating a new concrete FileIndexingServiceWriter, add a switch in RepositoryManager, method putDirInIndex(DirInfo, String) to select it for indexing.

The Lucene fields that are created by this class are:

• doctype - The document format type (e.g. dlese_ims, adn, oai_dc, etc.) defined by concrete classes, with '0' appended to support wildcard searching.
• readerclass - The class which is used to read typed Documents created by the concrete classes, for example "ItemDocReader".
• default - The default field containing content added by concrete classes. Generally this is the field assigned in the Lucene index for default searching.
• docsource - The absolute path to the file, which is used by the FileIndexingService for updating/deleting and may be used by beans or other classes that wish to have access to the source file.
• docdir - The absolute path to the directory where the file resides, which is used by the FileIndexingService for updating/deleting and may be used by beans or other classes.
• modtime - The file modification time, which is used by the FileIndexingService to determine if the file has changed and needs update and may be used by beans or other classes that wish to query the modtime for the record.
• filecontent - The full content of the file, stored but not indexed.
• deleted - Set to 'true' if the file or record for this document has been deleted, otherwise this field does not exist. Stored.
• valid - Set to 'true' if the file or record for this document is valid, otherwise 'false'. This field may also be ommited. Not stored.
• validationreport - Contains a report that provides validation information about the underlying file. This field may be ommited. Not stored.

Author:

John Weatherley

Constructor Summary

FileIndexingServiceWriter()

Method Summary

protected void	abortIndexing() Aborts the indexing process by returning a null index document.
protected abstract void	addCustomFields(org.apache.lucene.document.Document newDoc, org.apache.lucene.document.Document previousRecordDoc, File sourceFile) Adds additional custom fields that are unique the document format being indexed.
protected void	addDocToRemove(String field, String value) Removes a matching item from the index during the FileIndexingService update.
protected void	addToAdminDefaultField(String value) Adds the given String to a text field referenced in the index by the field name 'admindefault'.
protected void	addToDefaultField(String value) Adds the given String to the 'default' and 'stems' fields as text and stemmed text, respectively.
FileIndexingServiceData	create(File sourceFile, org.apache.lucene.document.Document existingLuceneDoc, FileIndexingPlugin plugin, HashMap sessionAttr) Creates the Lucene Document for the given resource or returns null if unable to create.
protected abstract void	destroy() This method is called at the conclusion of processing and may be used for tear-down.
HashMap	getConfigAttributes() Gets the configuration attributes that were set when the writer was created.
org.apache.lucene.document.Document	getDeletedDoc(org.apache.lucene.document.Document previousRecordDoc) Creates a Lucene Document equal to the exsiting FileIndexingService Document except the field "deleted" is to "true" and the field "modtime" has been set to the current time.
abstract String	getDocGroup() Gets the specifier associated with this group of files or null if no group association exists.
String	getDocsource() Gets the absolute path to the file, which is indexed under the 'docsource' field.
abstract String	getDocType() Gets a unique document type key for this kind of record, corresponding to the format type.
String	getFileContent() Gets the full content of the file as a String.
FileIndexingPlugin	getFileIndexingPlugin() Gets the FileIndexingPlugin that has been set for use during indexing, or null if none.
FileIndexingService	getFileIndexingService() Gets the fileIndexingService attribute of the FileIndexingServiceWriter object
org.apache.lucene.document.Document	getLuceneDoc() Gets the Lucene Document that this Writer is building.
org.apache.lucene.document.Document	getPreviousRecordDoc() Gets the previous Document that currently resides in the index for the given resource, or null if none was previously present.
abstract String	getReaderClass() Gets the fully qualified name of the concrete DocReader class that is used to read this type of Document, for example "org.dlese.dpc.index.reader.ItemDocReader".
HashMap	getSessionAttributes() Gets a Map of attributes used in a single indexing session.
File	getSourceDir() Gets the sourceDir that holds the file being indexed.
File	getSourceFile() Gets the sourceFile that is being indexed.
protected String	getValidationReport() Gets a report detailing any errors found in the validation of the file, or null if no error was found.
abstract void	init(File source, org.apache.lucene.document.Document previousRecordDoc) This method is called prior to processing and may be used to for any necessary set-up.
protected boolean	isMakingDeletedDoc() True if the current execution represents a deleted doc is being created.
boolean	isValidationEnabled() Returns true if the files being indexed should be validated, otherwise false.
protected void	prtln(String s) Output a line of text to standard out, with datestamp, if debug is set to true.
protected void	prtlnErr(String s) Output a line of text to error out, with datestamp.
void	setConfigAttributes(HashMap attributes) Sets the configuration attributes - called by the factory method that creates the FileIndexingServiceWriter.
static void	setDebug(boolean db) Sets the debug attribute of the FileIndexingServiceWriter object
void	setFileIndexingPlugin(FileIndexingPlugin plugin) Sets the FileIndexingPlugin that will be used during the indexing process to index additional fields.
void	setFileIndexingService(FileIndexingService fileIndexingService) Sets the fileIndexingService attribute of the FileIndexingServiceWriter object
protected void	setIsMakingDeletedDoc(boolean isMakingDeletedDoc) Sets whether this DocWriter is making a deleted document.
void	setValidationEnabled(boolean validateFiles) Sets whether or not to validate the files being indexed and create a validation report, which is indexed.

Methods inherited from class java.lang.Object

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

Constructor Detail

FileIndexingServiceWriter

public FileIndexingServiceWriter()

Method Detail

getDocType

public abstract String getDocType()
                           throws Exception

Gets a unique document type key for this kind of record, corresponding to the format type. In the DLESE metadata repository, this corresponds to the XML format, for example "oai_dc," "adn," "dlese_ims," or "dlese_anno". The string is parsed using the Lucene StandardAnalyzer so it must be lowercase and should not contain any stop words.

Specified by:

getDocType in interface DocWriter

Returns:

The docType String

Throws:

Exception - This method should throw and Exception with appropriate error message if an error occurs.

getDocGroup

public abstract String getDocGroup()
                            throws Exception

Gets the specifier associated with this group of files or null if no group association exists. In the DLESE metadata repository, this corresponds to the collection key, for example 'dcc', 'comet'.

Returns:

The docGroup specifier

Throws:

Exception - If error occured

getReaderClass

public abstract String getReaderClass()

Gets the fully qualified name of the concrete DocReader class that is used to read this type of Document, for example "org.dlese.dpc.index.reader.ItemDocReader".

Specified by:

getReaderClass in interface DocWriter

Returns:

The name of the DocReader.

init

public abstract void init(File source,
                          org.apache.lucene.document.Document previousRecordDoc)
                   throws Exception

This method is called prior to processing and may be used to for any necessary set-up. This method should throw and exception with appropriate message if an error occurs. The config attributes are set using the FileIndexingService.addDirectory(java.lang.String, java.lang.Class, java.util.HashMap, org.dlese.dpc.index.writer.FileIndexingPlugin, int) method.

Parameters:

source - The source file being indexed

previousRecordDoc - An existing Document that currently resides in the index for the given resource, or null if none was previously present

Throws:

Exception - If an error occured during set-up.

destroy

protected abstract void destroy()

This method is called at the conclusion of processing and may be used for tear-down.

addCustomFields

protected abstract void addCustomFields(org.apache.lucene.document.Document newDoc,
                                        org.apache.lucene.document.Document previousRecordDoc,
                                        File sourceFile)
                                 throws Exception

Adds additional custom fields that are unique the document format being indexed. When implementing this method, use the add method of the Document class to add a Field.

The following Lucene Field types are available for indexing with the Document:
Field.Text(string name, string value) -- tokenized, indexed, stored
Field.UnStored(string name, string value) -- tokenized, indexed, not stored
Field.Keyword(string name, string value) -- not tokenized, indexed, stored
Field.UnIndexed(string name, string value) -- not tokenized, not indexed, stored
Field(String name, String string, boolean store, boolean index, boolean tokenize) -- allows control to do anything you want

Example code:
protected void addCustomFields(Document newDoc, Document previousRecordDoc) throws Exception {
String customContent = "Some content";
newDoc.add(Field.Text("mycustomefield", customContent));
}

Parameters:

newDoc - The new Document that is being created for this resource

previousRecordDoc - An existing Document that currently resides in the index for the given resource, or null if none was previously present

sourceFile - The sourceFile that is being indexed

Throws:

Exception - This method should throw and Exception with appropriate error message if an error occurs.

getFileContent

public String getFileContent()
                      throws IOException

Gets the full content of the file as a String. If the file does not exist or the writer is processing a deleted doc, the content is pulled from the existing Lucene Document rather than the file.

Returns:

The full content of the file

Throws:

IOException - If error

getConfigAttributes

public HashMap getConfigAttributes()

Gets the configuration attributes that were set when the writer was created.

Returns:

The configuration attributes, or null if none were configured

setConfigAttributes

public void setConfigAttributes(HashMap attributes)

Sets the configuration attributes - called by the factory method that creates the FileIndexingServiceWriter.

Parameters:

attributes - The configuration attributes

getSessionAttributes

public HashMap getSessionAttributes()

Gets a Map of attributes used in a single indexing session. A seesion is a portion of indexing for a given directory of records that will be added to the index as a block update. Since records are added to the index at the end of the session, the index can not be used to query information from those records during the session. Thus, these attributes can be used to communitcate information across records being indexed within a given session, such as the record IDs found so far in the session. The attributes are cleared at the end of each session.

Returns:

A Map of records IDs keys, or null

getSourceFile

public File getSourceFile()

Gets the sourceFile that is being indexed. Only available after create() has been called.

Returns:

The sourceFile value

getDocsource

public String getDocsource()

Gets the absolute path to the file, which is indexed under the 'docsource' field.

Returns:

The absolute path to the file

getSourceDir

public File getSourceDir()

Gets the sourceDir that holds the file being indexed. Only available after create() has been called.

Returns:

The sourceDir value

getLuceneDoc

public org.apache.lucene.document.Document getLuceneDoc()

Gets the Lucene Document that this Writer is building.

Returns:

The Lucene Document

getPreviousRecordDoc

public org.apache.lucene.document.Document getPreviousRecordDoc()

Gets the previous Document that currently resides in the index for the given resource, or null if none was previously present.

Returns:

The previousRecordDoc value

setFileIndexingService

public void setFileIndexingService(FileIndexingService fileIndexingService)

Sets the fileIndexingService attribute of the FileIndexingServiceWriter object

Parameters:

fileIndexingService - The new fileIndexingService.

getFileIndexingService

public FileIndexingService getFileIndexingService()

Gets the fileIndexingService attribute of the FileIndexingServiceWriter object

Returns:

The fileIndexingService.

isValidationEnabled

public boolean isValidationEnabled()

Returns true if the files being indexed should be validated, otherwise false. This method may be ignored by concrete classes if not needed.

Returns:

true if validateion is enabled.

setValidationEnabled

public void setValidationEnabled(boolean validateFiles)

Sets whether or not to validate the files being indexed and create a validation report, which is indexed. This value is set by the FileIndexingService prior to indexing. If true, the method getValidationReport() will be called, otherwise it will not.

Parameters:

validateFiles - True to validate, else false.

See Also:

getValidationReport(), FileIndexingService.setValidationEnabled(boolean validateFiles)

getValidationReport

protected String getValidationReport()
                              throws Exception

Gets a report detailing any errors found in the validation of the file, or null if no error was found. This method should be overridden by concrete classes that need to validate the underlying file before indexing. Otherwise, this default method will simply return null. This method is called after all other method calls.

Returns:

Null if no file validation errors were found, otherwise a String that details the nature of the error.

Throws:

Exception - If error.

addToDefaultField

protected void addToDefaultField(String value)

Adds the given String to the 'default' and 'stems' fields as text and stemmed text, respectively. The default and stems fields may be used in queries to quickly search for text across fields. This method should be called from the addCustomFields of implementing classes.

Parameters:

value - A text string to be added to the indexed fields named 'default' and 'stems'

addToAdminDefaultField

protected void addToAdminDefaultField(String value)

Adds the given String to a text field referenced in the index by the field name 'admindefault'. The default field may be used in queries to quickly search for text across fields. This method should be called from the addCustomFields of implementing classes.

Parameters:

value - A text string to be added to the indexed field named 'admindefault.'

getDeletedDoc

public org.apache.lucene.document.Document getDeletedDoc(org.apache.lucene.document.Document previousRecordDoc)
                                                  throws Throwable

Creates a Lucene Document equal to the exsiting FileIndexingService Document except the field "deleted" is to "true" and the field "modtime" has been set to the current time.

Design note: This method should be overwritten by subclasses that require more envolved logic for deletes, and this super method should be called first and then subclassed should check #getIsMakingDeletedDoc to execute as appropriate.

Parameters:

previousRecordDoc - An existing FileIndexingService Document that currently resides in the index for the given file

Returns:

A Lucene FileIndexingService Document with appropriate fields updated

Throws:

Throwable - Thrown if error occurs

setIsMakingDeletedDoc

protected void setIsMakingDeletedDoc(boolean isMakingDeletedDoc)

Sets whether this DocWriter is making a deleted document. Used by subclassed that crate a DocWriter in their getDeletedDoc method.

Parameters:

isMakingDeletedDoc - Sets the making deleted doc status

isMakingDeletedDoc

protected final boolean isMakingDeletedDoc()

True if the current execution represents a deleted doc is being created.

Returns:

True if a deleted doc is being created

abortIndexing

protected void abortIndexing()

Aborts the indexing process by returning a null index document.

addDocToRemove

protected void addDocToRemove(String field,
                              String value)

Removes a matching item from the index during the FileIndexingService update. This method should be called to instruct the indexer to remove documents that should no longer be in the index.

Parameters:

field - The field to search in.

value - The matching value for the item to remove.

create

public FileIndexingServiceData create(File sourceFile,
                                      org.apache.lucene.document.Document existingLuceneDoc,
                                      FileIndexingPlugin plugin,
                                      HashMap sessionAttr)
                               throws Throwable

Creates the Lucene Document for the given resource or returns null if unable to create. This method is called by class FileIndexingService.

Parameters:

sourceFile - The source file to be indexed

existingLuceneDoc - An existing Document that currently resides in the index for the given resource, or null if none was previously present

plugin - The FileIndexingPlugin being used, or null

sessionAttr - Attributes used in a given indexing session

Returns:

A Lucene Document with it's fields populated, or null.

Throws:

Throwable - Thrown if error occurs

setFileIndexingPlugin

public void setFileIndexingPlugin(FileIndexingPlugin plugin)

Sets the FileIndexingPlugin that will be used during the indexing process to index additional fields. Set to null to remove.

Parameters:

plugin - A FileIndexingPlugin to use during indexing.

getFileIndexingPlugin

public FileIndexingPlugin getFileIndexingPlugin()

Gets the FileIndexingPlugin that has been set for use during indexing, or null if none.

Returns:

The FileIndexingPlugin configured for use used, or null.

prtlnErr

protected final void prtlnErr(String s)

Output a line of text to error out, with datestamp.

Parameters:

s - The text that will be output to error out.

prtln

protected final void prtln(String s)

Output a line of text to standard out, with datestamp, if debug is set to true.

Parameters:

s - The String that will be output.

setDebug

public static final void setDebug(boolean db)

Sets the debug attribute of the FileIndexingServiceWriter object

Parameters:

db - The new debug value