PDFlib Cookbook

cookbook

pdfa/zugferd_add_xml_to_pdfa

Download PHP Code      Switch to Java Code         Show Output PDF         Show Input Files

<?php
/*
 * $Id: zugferd_add_xml_to_pdfa.php,v 1.2 2017/01/03 13:49:02 rp Exp $
 * 
 * ZUGFeRD Add XML Invoice to PDF:
 * Import the pages of a plain PDF/A invoice with PDI and add a
 * corresponding XML invoice to the output file. Note that the application
 * must take care that the imported PDF invoice and the XML invoice match.
 *
 * Required software: PDFlib+PDI/PPS 9
 * Required data: PDF/A input document, XML invoice file, XMP metadata
 */
/* This is where the data files are. Adjust as necessary. */
$searchpath =  dirname(dirname(dirname(__FILE__)))."/input";

$infile = "ZUGFeRD_invoice_noxml.pdf";
$title = "ZUGFeRD Add XML Invoice to PDF";
$xmpfile = "ZUGFeRD_extension_schema.xmp";
$xmlinvoicefile = "ZUGFeRD-invoice.xml";

$p = new PDFlib();
$exitcode = 0;

try {
    $p = new pdflib();

    $p->set_option("searchpath={" . $searchpath . "}");

    /* This means we must check return values of load_font() etc. */
    $p->set_option("errorpolicy=exception");

    if ($p->begin_document("", "pdfa=PDF/A-3b metadata={filename={"
	. $xmpfile . "}}") == 0)
	throw new Exception("Error: " . $p->get_errmsg());

    /* Open the input PDF */
    $indoc = $p->open_pdi_document($infile, "");
    if ($indoc == 0) {
	throw new Exception("Error: " . $p->get_errmsg());
    }
    
    /*
     * Clone PDF/A or PDF/X output intent
     */
    if ($p->process_pdi($indoc, -1, "action=copyoutputintent") == 0) {
	throw new Exception("Error: " . $p->get_errmsg());
    }

    $p->set_info("Creator", "zugferd_add_xml_to_pdf.java");
    $p->set_info("Title", $title . ' $Revision: 1.2 $');
    
    $endpage = (int) $p->pcos_get_number($indoc, "length:pages");

    /* Loop over all pages of the input document */
    for ($pageno = 1; $pageno <= $endpage; $pageno++) {
	$page = $p->open_pdi_page($indoc, $pageno, "");

	if ($page == 0) {
	    throw new Exception("Error: " . $p->get_errmsg());
	}
	
	/* Dummy page size; will be adjusted later */
	$p->begin_page_ext(10, 10, "");

	/*
	 * Place the imported page on the output page, and adjust the
	 * page size
	 */
	$p->fit_pdi_page($page, 0, 0, "adjustpage");
	$p->close_pdi_page($page);

	$p->end_page_ext("");
    }
    
    /*
     * Clone XMP metadata of input document if present. In order to
     * be able to merge the XMP from the input document with the
     * XMP needed according to ZUGFeRD for identifying the
     * XML invoice, the XMP from the input document is supplied
     * to the "metadata" option of PDF_end_document(), while the
     * ZUGFeRD XMP was supplied to the "metadata" option of
     * PDF_begin_document().
     */
    $pfv_xmpfile = "/xmp/document.xmp";
    if (!$p->pcos_get_string($indoc, "type:/Root/Metadata") == "stream") {
	throw new Exception(
	    "Error: XMP metadata missing in input document");
    }
    
    $xmp = $p->pcos_get_stream($indoc, "", "/Root/Metadata");
    $p->create_pvf($pfv_xmpfile, $xmp, "");
    $xmp_optlist = "metadata={filename=" . $pfv_xmpfile . "}";
    
    $p->close_pdi_document($indoc);

    /*
     * Load the XML file for the invoice, and associate it with the
     * document with relationship "Alternative". Also mark it as an
     * attachment that can be retrieved with Acrobat's normal attachment
     * dialog.
     */
    $xml_invoice = $p->load_asset("Attachment", $xmlinvoicefile,
	"mimetype=text/xml "
	    . "description={Rechnungsdaten im ZUGFeRD-XML-Format} "
	    . "relationship=Alternative documentattachment=true");

    if ($xml_invoice == 0) {
	throw new Exception("Error: " . $p->get_errmsg());
    }
    
    $p->end_document("associatedfiles={" . $xml_invoice . "} "
		    . $xmp_optlist);
    
    $p->delete_pvf($pfv_xmpfile);

    $buf = $p->get_buffer();
    $len = strlen($buf);

    header("Content-type: application/pdf");
    header("Content-Length: $len");
    header("Content-Disposition: inline; filename=" . $title . ".pdf");
    print $buf;
} catch (PDFlibException $e) {
    echo("PDFlib exception occurred:\n".
        "[" . $e->get_errnum() . "] " . $e->get_apiname() .
        ": " . $e->get_errmsg() . "\n");
    exit(1);
} catch (Exception $e) {
    echo($e->getMessage());
    exit(1);
}

$p = 0;