Examples of it.unimi.dsi.mg4j.document.DocumentSequence.factory()

Class it.unimi.dsi.mg4j.document.DocumentSequence

Examples of it.unimi.dsi.mg4j.document.DocumentSequence.factory()

it.unimi.dsi.mg4j.document.DocumentSequence.factory()
Returns the factory used by this sequence.
Every document sequence is based on a document factory that transforms raw bytes into a sequence of characters. The factory contains useful information such as the number of fields. @return the factory used by this sequence.

    if ( ( jsapResult.userSpecified( "builderClass" ) || jsapResult.userSpecified( "exact" ) ) && ! jsapResult.userSpecified( "buildCollection" ) )  throw new IllegalArgumentException( "To specify options about the collection building process, you must specify a basename first." );
    if ( jsapResult.userSpecified( "sequence" ) && jsapResult.userSpecified( "objectSequence" ) ) throw new IllegalArgumentException( "You cannot specify both a serialised and an parseable-object sequence" );
    
    final DocumentSequence documentSequence = jsapResult.userSpecified( "objectSequence" ) ? (DocumentSequence)jsapResult.getObject( "objectSequence" ) : Scan.getSequence( jsapResult.getString( "sequence" ), jsapResult.getClass( "factory" ), jsapResult.getStringArray( "property" ), jsapResult.getInt( "delimiter" ), LOGGER );


    final DocumentFactory factory = documentSequence.factory();
    final int[] indexedField = parseFieldNames( jsapResult.getStringArray( "indexedField" ), factory, jsapResult.getBoolean( "allFields" ) );
    final int batchSize = jsapResult.getInt( "batchSize" );
    final VirtualDocumentResolver[] virtualDocumentResolver = parseVirtualDocumentResolver( jsapResult.getStringArray( "virtualDocumentResolver" ), indexedField, factory );
    final int[] virtualDocumentGap = parseVirtualDocumentGap( jsapResult.getStringArray( "virtualDocumentGap" ), indexedField, factory );

View Full Code Here

    DocumentCollectionBuilder builder = null;
    if ( jsapResult.userSpecified( "buildCollection" ) ) {
      final Class<? extends DocumentCollectionBuilder> builderClass = jsapResult.getClass( "builderClass" );
      builder = builderClass != null ? builderClass.getConstructor( String.class, DocumentFactory.class, boolean.class ).newInstance( 
          jsapResult.getString( "buildCollection" ), 
          documentSequence.factory().numberOfFields() == indexedField.length ? documentSequence.factory().copy() : new SubDocumentFactory( documentSequence.factory().copy(), indexedField ), 
          Boolean.valueOf( jsapResult.getBoolean( "exact" ) ) ) : null;
    }


    run( jsapResult.getString( "basename" ), documentSequence, Completeness.valueOf( jsapResult.getString( "completeness" ) ), jsapResult.getBoolean( "downcase" ) ? DowncaseTermProcessor.getInstance() : ObjectParser.fromSpec( jsapResult
        .getString( "termProcessor" ), TermProcessor.class, MG4JClassParser.PACKAGE, new String[] { "getInstance" } ), builder, jsapResult

View Full Code Here

    DocumentCollectionBuilder builder = null;
    if ( jsapResult.userSpecified( "buildCollection" ) ) {
      final Class<? extends DocumentCollectionBuilder> builderClass = jsapResult.getClass( "builderClass" );
      builder = builderClass != null ? builderClass.getConstructor( String.class, DocumentFactory.class, boolean.class ).newInstance( 
          jsapResult.getString( "buildCollection" ), 
          documentSequence.factory().numberOfFields() == indexedField.length ? documentSequence.factory().copy() : new SubDocumentFactory( documentSequence.factory().copy(), indexedField ), 
          Boolean.valueOf( jsapResult.getBoolean( "exact" ) ) ) : null;
    }


    run( jsapResult.getString( "basename" ), documentSequence, Completeness.valueOf( jsapResult.getString( "completeness" ) ), jsapResult.getBoolean( "downcase" ) ? DowncaseTermProcessor.getInstance() : ObjectParser.fromSpec( jsapResult
        .getString( "termProcessor" ), TermProcessor.class, MG4JClassParser.PACKAGE, new String[] { "getInstance" } ), builder, jsapResult

View Full Code Here

    DocumentCollectionBuilder builder = null;
    if ( jsapResult.userSpecified( "buildCollection" ) ) {
      final Class<? extends DocumentCollectionBuilder> builderClass = jsapResult.getClass( "builderClass" );
      builder = builderClass != null ? builderClass.getConstructor( String.class, DocumentFactory.class, boolean.class ).newInstance( 
          jsapResult.getString( "buildCollection" ), 
          documentSequence.factory().numberOfFields() == indexedField.length ? documentSequence.factory().copy() : new SubDocumentFactory( documentSequence.factory().copy(), indexedField ), 
          Boolean.valueOf( jsapResult.getBoolean( "exact" ) ) ) : null;
    }


    run( jsapResult.getString( "basename" ), documentSequence, Completeness.valueOf( jsapResult.getString( "completeness" ) ), jsapResult.getBoolean( "downcase" ) ? DowncaseTermProcessor.getInstance() : ObjectParser.fromSpec( jsapResult
        .getString( "termProcessor" ), TermProcessor.class, MG4JClassParser.PACKAGE, new String[] { "getInstance" } ), builder, jsapResult

View Full Code Here


    if ( ( jsapResult.userSpecified( "builderClass" ) || jsapResult.userSpecified( "exact" ) ) && ! jsapResult.userSpecified( "buildCollection" ) )  throw new IllegalArgumentException( "To specify options about the collection building process, you must specify a basename first." );
    if ( jsapResult.userSpecified( "sequence" ) && jsapResult.userSpecified( "objectSequence" ) ) throw new IllegalArgumentException( "You cannot specify both a serialised and an parseable-object sequence" );
    
    final DocumentSequence documentSequence = jsapResult.userSpecified( "objectSequence" ) ? (DocumentSequence)jsapResult.getObject( "objectSequence" ) : Scan.getSequence( jsapResult.getString( "sequence" ), jsapResult.getClass( "factory" ), jsapResult.getStringArray( "property" ), jsapResult.getInt( "delimiter" ), LOGGER );
    final DocumentFactory factory = documentSequence.factory();


    final int[] indexedField = Scan.parseFieldNames( jsapResult.getStringArray( "indexedField" ), factory, jsapResult.getBoolean( "allFields" ) );
    final VirtualDocumentResolver[] virtualDocumentResolver = Scan.parseVirtualDocumentResolver( jsapResult.getStringArray( "virtualDocumentResolver" ), indexedField, factory );
    final int[] virtualDocumentGap = Scan.parseVirtualDocumentGap( jsapResult.getStringArray( "virtualDocumentGap" ), indexedField, factory );

View Full Code Here

    DocumentCollectionBuilder builder = null;
    if ( jsapResult.userSpecified( "buildCollection" ) ) {
      final Class<? extends DocumentCollectionBuilder> builderClass = jsapResult.getClass( "builderClass" );
      builder = builderClass != null ? builderClass.getConstructor( String.class, DocumentFactory.class, boolean.class ).newInstance( 
          jsapResult.getString( "buildCollection" ), 
          documentSequence.factory().numberOfFields() == indexedField.length ? documentSequence.factory().copy() : new SubDocumentFactory( documentSequence.factory().copy(), indexedField ), 
          Boolean.valueOf( jsapResult.getBoolean( "exact" ) ) ) : null;
    }


    final IndexBuilder indexBuilder = new IndexBuilder( jsapResult.getString( "basename" ), documentSequence )
    .termProcessor( termProcessor )

View Full Code Here

    DocumentCollectionBuilder builder = null;
    if ( jsapResult.userSpecified( "buildCollection" ) ) {
      final Class<? extends DocumentCollectionBuilder> builderClass = jsapResult.getClass( "builderClass" );
      builder = builderClass != null ? builderClass.getConstructor( String.class, DocumentFactory.class, boolean.class ).newInstance( 
          jsapResult.getString( "buildCollection" ), 
          documentSequence.factory().numberOfFields() == indexedField.length ? documentSequence.factory().copy() : new SubDocumentFactory( documentSequence.factory().copy(), indexedField ), 
          Boolean.valueOf( jsapResult.getBoolean( "exact" ) ) ) : null;
    }


    final IndexBuilder indexBuilder = new IndexBuilder( jsapResult.getString( "basename" ), documentSequence )
    .termProcessor( termProcessor )

View Full Code Here

    DocumentCollectionBuilder builder = null;
    if ( jsapResult.userSpecified( "buildCollection" ) ) {
      final Class<? extends DocumentCollectionBuilder> builderClass = jsapResult.getClass( "builderClass" );
      builder = builderClass != null ? builderClass.getConstructor( String.class, DocumentFactory.class, boolean.class ).newInstance( 
          jsapResult.getString( "buildCollection" ), 
          documentSequence.factory().numberOfFields() == indexedField.length ? documentSequence.factory().copy() : new SubDocumentFactory( documentSequence.factory().copy(), indexedField ), 
          Boolean.valueOf( jsapResult.getBoolean( "exact" ) ) ) : null;
    }


    final IndexBuilder indexBuilder = new IndexBuilder( jsapResult.getString( "basename" ), documentSequence )
    .termProcessor( termProcessor )

View Full Code Here

    JSAPResult jsapResult = jsap.parse( arg );
    if ( jsap.messagePrinted() ) return;
    
    DocumentSequence documentSequence = it.unimi.dsi.mg4j.tool.Scan.getSequence( jsapResult.getString( "sequence" ), jsapResult.getClass( "factory" ), jsapResult.getStringArray( "property" ), jsapResult.getInt( "delimiter" ), LOGGER );
    
    final DocumentFactory factory = documentSequence.factory();
    final boolean stem = jsapResult.getBoolean( "stem" );
    final boolean termLists = jsapResult.getBoolean( "termLists" );
    final int[] indexedField = it.unimi.dsi.mg4j.tool.Scan.parseFieldNames( jsapResult.getStringArray( "indexedField" ), factory, jsapResult.getBoolean( "allFields" )  );
    
    LOGGER.debug( "Parsed indexed field: " + IntArrayList.wrap( indexedField ) );

View Full Code Here

TOP

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.