NAME

CAM::PDF::Content - PDF page layout parser

LICENSE

See CAM::PDF.

SYNOPSIS

use CAM::PDF;
my $pdf = CAM::PDF->new($filename);

my $contentTree = $pdf->getPageContentTree(4);
$contentTree->validate() || die 'Syntax error';
print $contentTree->render('CAM::PDF::Render::Text');
$pdf->setPageContent(5, $contentTree->toString());

DESCRIPTION

This class is used to manipulate the layout commands for a single page of PDF. The page content is passed as a scalar and parsed according to Adobe's PDF Reference 3rd edition (for PDF v1.4). All of the commands from Appendix A of that document are parsed and understood.

Much of the content object's functionality is wrapped up in renderers that can be applied to it. See the canonical renderer, CAM::PDF::GS, and the render() method below for more details.

FUNCTIONS

new CONTENT

new CONTENT, DATA

new CONTENT, DATA, VERBOSE

Parse a scalar CONTENT containing PDF page layout content. Returns a parsed, but unvalidated, data structure.

The DATA argument is a hash reference of contextual data that may be needed to work with content. This is only needed for toString() method (which needs doc => CAM::PDF object to work with images) and the render methods, to which the DATA reference is passed verbatim. See the individual renderer modules for details about required elements.

The VERBOSE boolean indicates whether the parser should Carp when it encounters problems. The default is false.

parse CONTENTREF

This is intended to be called by the new() method. The argument should be a reference to the content scalar. It's passed by reference so it is never copied.

validate

Returns a boolean if the parsed content tree conforms to the PDF specification.

render RENDERERCLASS

Traverse the content tree using the specified rendering class. See CAM::PDF::GS or CAM::PDF::Renderer::Text for renderer examples. Renderers should typically derive from CAM::PDF::GS, but it's not essential. Typically returns an instance of the renderer class.

The rendering class is loaded via require if not already in memory.

computeGS

computeGS SKIPTEXT

Traverses the content tree and computes the coordinates of each graphic point along the way. If the SKIPTEXT boolean is true (default: false) then text blocks are ignored to save time, since they do not change the global graphic state.

This is a thin wrapper around render() with CAM::PDF::GS or CAM::PDF::GS::NoText selected as the rendering class.

findImages

Traverse the content tree, accumulating embedded images and image references, according to the CAM::PDF::Renderer::Images renderer.

traverse RENDERERCLASS

This recursive method is typically called only by wrapper methods, like render(). It instantiates renderers as needed and calls methods on them.

toString

Flattens a content tree back into a scalar, ready to be inserted back into a PDF document. Since whitespace is discarded by the parser, the resulting scalar will not be identical to the original.

AUTHOR

Clotho Advanced Media Inc., cpan@clotho.com