VisualOutputTarget (PDFTextStream API Reference)

Overview

Package

Class

Tree

Deprecated

Index

Help

PREV CLASS NEXT CLASS

FRAMES NO FRAMES

SUMMARY: NESTED | FIELD | CONSTR | METHOD

DETAIL: FIELD | CONSTR | METHOD

com.snowtide.pdf
Class VisualOutputTarget

java.lang.Object
  com.snowtide.pdf.OutputHandler
      com.snowtide.pdf.VisualOutputTarget

public class VisualOutputTarget
extends OutputHandler
extends OutputHandler

This OutputHandler implementation aims to preserve as much of a PDF's text layout as possible so that text extracts yielded by this OutputHandler will retain the visual arrangement of text as present in the original document. This is ideal when the content being extracted is to be used as input into a downstream conversion process. For example, this OutputHandler will maintain the layout of most tabular data (this table is a sample for illustration purposes only):

         Column 1     Column 2      Column 3
 Row 1      $500        $1,000           14B
 Row 2    $1,000        $5,621            8A
 Row 3    $6,009          $121           N/A

whereas the default OutputTarget is more likely to output such tabular data with proper read-ordering, but with no concern for spacing or line breaks between the table's cells and rows:

 Column 1 Column 2 Column 3
 Row 1 $500 $1,000 14B
 Row 2 $1,000 $5,621 8A
 Row 3 $6,009 $121 N/A

Please note the following regarding VisualOutputTarget:

Because VisualOutputTarget attempts to maintain the visual appearance of each page's text, it will not separate columns and other document features that might be very important to a semantically-sensitive text analysis process (such as search indexing).
VisualOutputTarget will yield very poor output when used to format rotated text; in such a case, the results are essentially undefined. You may optionally suppress the inclusion of rotated characters from VisualOutputTarget's output using setIncludingRotatedChars(boolean).
Using VisualOutputTarget is likely to impose a slight performance penalty compared to using the default OutputTarget. This penalty should be no more than 5%, and is necessary because of the processing needed to normalize the extracted text to appear as it does on the page.
The current implementation of VisualOutputTarget performs best when working with text rendered using a monospace font. Proportional fonts and (especially) justified text complicates the process of normalizing the spacing of the text formatted by this class. Improvements will likely be made in future PDFTextStream releases to make this class more capable when handling justified text or text rendered using proportional fonts.

Since:: v2.0
Version:: ©2004-2012 Snowtide Informatics Systems, Inc.

Constructor Summary
`VisualOutputTarget(java.lang.Appendable sb)`
`VisualOutputTarget(java.io.Writer w)`

Method Summary
`void`	`endLine(Line line)` Invoked when PDFTextStream has finished processing a Line.
`void`	`endPage(Page page)` Invoked when PDFTextStream has finished processing a page
`float`	`getSpacingScale()` Returns the spacing scale currently in effect for this `VisualOutputTarget`.
`boolean`	`isIncludingRotatedChars()` Return true if this `VisualOutputTarget` will include rotated `TextUnit`s in its output (true by default).
`void`	`linebreaks(int linebreakCnt)` Invoked when PDFTextStream determines that a series of line breaks should be outputted between the previous entity (page, block, line, etc) and the next entity (page, block, line, etc).
`void`	`setIncludingRotatedChars(boolean includingRotatedChars)` Used to set whether or not this `VisualOutputTarget` will include rotated `TextUnit`s in its output (true by default).
`void`	`setSpacingScale(float scale)` Modifies the spacing scale that is used when outputting content laid out using this `VisualOutputTarget`.
`void`	`spaces(int spaceCnt)` Invoked when PDFTextStream determines that a series of spaces should be outputted between the previous entity (block, line, text unit, etc) and the next entity (block, line, text unit, etc).
`void`	`startBlock(Block b)` Invoked when a Block is about to be processed.
`void`	`startLine(Line line)` Invoked when a Line is about to be processed.
`void`	`startPage(Page page)` Invoked when a page is about to be processed.
`void`	`textUnit(TextUnit tu)` Invoked when a run of characters is to be outputted, as represented by the given `TextUnit` instance.

Methods inherited from class com.snowtide.pdf.OutputHandler
`endBlock, endPDF, startPDF`

Methods inherited from class java.lang.Object
`clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait`

Constructor Detail

VisualOutputTarget

public VisualOutputTarget(java.io.Writer w)

VisualOutputTarget

public VisualOutputTarget(java.lang.Appendable sb)

Method Detail

setSpacingScale

public void setSpacingScale(float scale)

Modifies the spacing scale that is used when outputting content laid out using this VisualOutputTarget. The default is 1; using a value of 2 will (approximately) double the number of spaces that are outputted between recognized words, while a value of .5 will (approximately) halve that number. This is useful in circumstances where:

Columnar data elements should be separated by greater distances, perhaps to aid data collection
Words are being outputted without any spaces between them at all, perhaps because of the use of (pathologically?) small fonts.

getSpacingScale

public float getSpacingScale()

Returns the spacing scale currently in effect for this VisualOutputTarget.

See Also:: setSpacingScale(float)

isIncludingRotatedChars

public boolean isIncludingRotatedChars()

Return true if this VisualOutputTarget will include rotated TextUnits in its output (true by default).

setIncludingRotatedChars

public void setIncludingRotatedChars(boolean includingRotatedChars)

Used to set whether or not this VisualOutputTarget will include rotated TextUnits in its output (true by default).