ocr-d--page-to-alto.pl

#!/usr/bin/perl
use strict;
use utf8;
binmode(STDOUT, ":utf8");
binmode(STDERR, ":utf8");
use XML::LibXML;
use File::Slurp;
use XML::Quote;
use List::Util qw(min max);

# QUICK AND DIRTY. NOT FOR PRODUCTION USE.
# for output of OCR-D.

my $prog="$0";

for my $xml_fn (@ARGV) {
  if($xml_fn=~/\.alto\.xml$/) { next; }
  my $xml_str=read_file($xml_fn);
  my $dom;
  eval {
    $dom=XML::LibXML->load_xml(string=>$xml_str);
  };
  if($@) { warn "$prog: $xml_fn: $@\n"; next; }
  
  my $out='<?xml version="1.0" encoding="UTF-8" standalone="yes"?>';
  $out.='<!-- generated by '.$prog.' -->';
  $out.='<alto xmlns="http://www.loc.gov/standards/alto/ns-v2#">';
  $out.='<Layout>';

	my $xpc = XML::LibXML::XPathContext->new($dom);
	$xpc->registerNs('PAGE',  'http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15');
	my $root=$dom->documentElement;

  my $pageNr=0;
	for my $Page ($xpc->findnodes('/PAGE:PcGts/PAGE:Page', $root)) { 
    $pageNr++;
    my $w=$Page->getAttribute("imageWidth");
    my $h=$Page->getAttribute("imageHeight");
    $out.=qq{<Page ID="Page1" PHYSICAL_IMG_NR="$pageNr" HEIGHT="$h" WIDTH="$w"><PrintSpace>};
    for my $RegionRefIndexed ($xpc->findnodes('//PAGE:RegionRefIndexed', $Page)) {
      my $id=$RegionRefIndexed->getAttribute("regionRef");
      for my $region ($xpc->findnodes('//PAGE:*[@id="'.$id.'"]')) {
        # if($region->nodeName ne "TextRegion") { next; } # perhaps Table with TextLine?
        my $outLine="";
        for my $TextLine ($xpc->findnodes(".//PAGE:TextLine", $region)) {
          my $xmin=1e308; my $xmax=-1e308;
          my $ymin=1e308; my $ymax=-1e308;
          for my $coords ($xpc->findnodes(".//PAGE:Coords", $TextLine)) {
            for my $xy (split /\s+/, $coords->getAttribute("points")) {
              my($x,$y)=split /,/, $xy, 2;
              $xmin=min($xmin, $x);
              $xmax=max($xmax, $x);
              $ymin=min($ymin, $y);
              $ymax=max($ymax, $y);
            }
          }
          my $w=$xmax-$xmin;
          my $h=$ymax-$ymin;
          for my $TextEquiv ($xpc->findnodes(".//PAGE:TextEquiv/PAGE:Unicode", $TextLine)) {
            my $txt=$TextEquiv->textContent; $txt=~s/[\r\n]//g; $txt=~s/[\x00-\x08\x0a-\x1f]/ /g;
            $outLine.=qq{<TextLine HEIGHT="$h" WIDTH="$w" VPOS="$ymin" HPOS="$xmin"><String CONTENT="}.xml_quote($txt).qq{"/></TextLine>};
          }
        }
        if(length($outLine)) {
          $out.='<TextBlock>'.$outLine.'</TextBlock>';
        }
      }
    }
    $out.='</PrintSpace></Page>';
	}
  $out.='</Layout></alto>';
  my $ofn=$xml_fn; $ofn=~s/\.xml$/.alto.xml/;
  open my $f, ">:utf8", $ofn or die;
    print $f $out;
  close $f;
}