org.terrier.indexing

Class SimpleFileCollection

java.lang.Object

org.terrier.indexing.SimpleFileCollection

All Implemented Interfaces:

Closeable, AutoCloseable, Collection

public class SimpleFileCollection
extends Object
implements Collection

Implements a collection that can read arbitrary files on disk. It will use the file list given to it in the constructor, or it will read the file specified by the property collection.spec. Properties:

• indexing.simplefilecollection.extensionsparsers - a comma delimited lists of tuples, in the form "extension:DocumentClass". For instance, one tuple could be "txt:FileDocument". The default txt:FileDocument,text:FileDocument,tex:FileDocument,bib:FileDocument,pdf:PDFDocument,html:TaggedDocument,htm:TaggedDocument,xhtml:TaggedDocument,xml:TaggedDocument,doc:MSWordDocument,ppt:MSPowerpointDocument,xls:MSExcelDocument.
• indexing.simplefilecollection.defaultparser - the default parser for any unknown extensions. If this property is empty, then such documents will not be opened.
• indexing.simplefilecollection.recurse - whether directories should be opened looking for files.

Author:

Craig Macdonald & Vassilis Plachouras

Field Summary

Fields

Modifier and Type	Field and Description
protected InputStream	currentStream The InputStream of the most recently opened document.
protected int	Docid The identifier of a document in the collection.
protected Map<String,Class<? extends Document>>	extension_DocumentClass Maps filename extensions to Document classes.
protected LinkedList<String>	FileList The list of files to index.
protected List<String>	firstList Contains the list of files first handed to the SimpleFileCollection, allowing the SimpleFileCollection instance to be simply reset.
protected List<String>	indexedFiles This is filled during traversal, so document IDs can be matched with filenames
protected static org.slf4j.Logger	logger
static String	NAMESPACE_DOCUMENTS The default namespace for all parsers to be loaded from.
protected boolean	Recurse Whether directories should be recursed into by this class
protected String	thisFilename The filename of the current file we are processing.
protected Tokeniser	tokeniser

Constructor Summary

Constructors

Constructor and Description
SimpleFileCollection() A default constructor that uses the files to be processed by this collection, as specified by the property collection.spec
SimpleFileCollection(List<String> filelist, boolean recurse) Constructs an instance of the class with the given list of files.
SimpleFileCollection(String addressCollectionFilename) Creates an instance of the class.

Method Summary

Methods

Modifier and Type	Method and Description
protected void	addDirectoryListing() Called when thisFile is identified as a directory, this adds the entire contents of the directory onto the list to be processed.
void	close()
protected void	createExtensionDocumentMapping() Parses the properties indexing.simplefilecollection.extensionsparsers and indexing.simplefilecollection.defaultparser and attempts to load all the mentioned classes, in a hashtable mapping filename extension to their respective parsers.
boolean	endOfCollection() Checks whether there are more documents in the colection.
String	getDocid() Returns the current document's identifier string.
Document	getDocument() Return the current document in the collection.
List<String>	getFileList() Returns the ist of indexed files in the order they were indexed in.
boolean	hasNext() Check whether there is a next document in the collection to be processed
static void	main(String[] args) Simple test case.
protected Document	makeDocument(String Filename, InputStream in) Given the opened document in, of Filename and File f, work out which parser to try, and instantiate it.
Document	next() Move onto the next document in the collection to be processed.
boolean	nextDocument() Move onto the next document in the collection to be processed.
void	remove() This is unsupported by this Collection implementation, and any calls will throw UnsupportedOperationException Throws UnsupportedOperationException on all invocations
void	reset() Starts again from the beginning of the collection.

Methods inherited from class java.lang.Object

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

Field Detail

logger

protected static final org.slf4j.Logger logger

NAMESPACE_DOCUMENTS

public static final String NAMESPACE_DOCUMENTS

The default namespace for all parsers to be loaded from. Only used if the class name specified does not contain any periods ('.')

See Also:

Constant Field Values

FileList

protected LinkedList<String> FileList

The list of files to index.

firstList

protected List<String> firstList

Contains the list of files first handed to the SimpleFileCollection, allowing the SimpleFileCollection instance to be simply reset.

indexedFiles

protected List<String> indexedFiles

This is filled during traversal, so document IDs can be matched with filenames

Docid

protected int Docid

The identifier of a document in the collection.

Recurse

protected boolean Recurse

Whether directories should be recursed into by this class

extension_DocumentClass

protected Map<String,Class<? extends Document>> extension_DocumentClass

Maps filename extensions to Document classes. The entry |DEFAULT| maps to the default document parser, specified by indexing.simplefilecollection.defaultparser

thisFilename

protected String thisFilename

The filename of the current file we are processing.

currentStream

protected InputStream currentStream

The InputStream of the most recently opened document. This is used to ensure that files are closed once they have been finished reading.

tokeniser

protected Tokeniser tokeniser

Constructor Detail

SimpleFileCollection

public SimpleFileCollection(List<String> filelist,
                    boolean recurse)

Constructs an instance of the class with the given list of files.

Parameters:

filelist - ArrayList the files to be processed by this collection.

SimpleFileCollection

public SimpleFileCollection()

A default constructor that uses the files to be processed by this collection, as specified by the property collection.spec

SimpleFileCollection

public SimpleFileCollection(String addressCollectionFilename)

Creates an instance of the class. The files to be processed are specified in the file with the given name.

Parameters:

addressCollectionFilename - String the name of the file that contains the list of files to be processed by this collecion.

Method Detail

createExtensionDocumentMapping

protected void createExtensionDocumentMapping()

Parses the properties indexing.simplefilecollection.extensionsparsers and indexing.simplefilecollection.defaultparser and attempts to load all the mentioned classes, in a hashtable mapping filename extension to their respective parsers. If indexing.simplefilecollection.defaultparser is set, then that class will be used to attempt to parse documents that no explicit parser is set.

hasNext

public boolean hasNext()

Check whether there is a next document in the collection to be processed

Returns:

has next

next

public Document next()

Move onto the next document in the collection to be processed.

Returns:

next document

remove

public void remove()

This is unsupported by this Collection implementation, and any calls will throw UnsupportedOperationException Throws UnsupportedOperationException on all invocations

nextDocument

public boolean nextDocument()

Move onto the next document in the collection to be processed.

Specified by:

nextDocument in interface Collection

Returns:

boolean true if there are more documents in the collection, otherwise return false.

getDocument

public Document getDocument()

Return the current document in the collection.

Specified by:

getDocument in interface Collection

Returns:

Document the next document object from the collection.

makeDocument

protected Document makeDocument(String Filename,
                    InputStream in)

Given the opened document in, of Filename and File f, work out which parser to try, and instantiate it. If you wish to use a different constructor for opening documents, then you need to subclass this method.

Parameters:

Filename - the filename of the currently open document

in - The stream of the currently open document

Returns:

Document object to parse the document, or null if no suitable parser exists.

endOfCollection

public boolean endOfCollection()

Checks whether there are more documents in the colection.

Specified by:

endOfCollection in interface Collection

Returns:

boolean true if there are no more documents in the collection, otherwise it returns false.

reset

public void reset()

Starts again from the beginning of the collection.

Specified by:

reset in interface Collection

getDocid

public String getDocid()

Returns the current document's identifier string.

Returns:

String the identifier of the current document.

close

public void close()

Specified by:

close in interface Closeable

Specified by:

close in interface AutoCloseable

getFileList

public List<String> getFileList()

Returns the ist of indexed files in the order they were indexed in.

addDirectoryListing

protected void addDirectoryListing()

Called when thisFile is identified as a directory, this adds the entire contents of the directory onto the list to be processed.

main

public static void main(String[] args)

Simple test case. Pass the filename of a file that lists files to be processed to this test case.