pdf_and_ocr.pl

#! env perl
use 5.38.0;
use Data::Printer;
use DateTime::Format::Text;
use FileHandle; 
STDOUT->autoflush(); # So Data::Printer output is inline with our other debugging
# ./pdf_and_ocr.pl 2>&1 | head -60

# Read from sorted.txt. For each issue:
# * squash clustered TIFFs together
# * convert that TIFF to PDF
# * OCR that PDF

my $in_dir = "/Users/jhannah/Dropbox/1989-1991-WHAMO";
my $out_dir = "$in_dir/experiments";

my $dft = DateTime::Format::Text->new();

open my $in, "<", "sorted.txt";
my %issues;
my $this_issue = [];
my ($tiff, $name, $subject, $comments, $authors);
while (<$in>) {
  chomp;
  # say;
  next if (/^#/);  # Ignore comments
  unless (/\w/) {
    # We've hit a blank line, which means a split between issues
    if (@$this_issue) {
      my $date = convert_date($subject);
      next unless $date;  # uhh... end of file?
      my $out_filename;
      if ($name =~ /WHAMO/) {
        $out_filename = "WHAMO-$date";
      } elsif ($name =~ /Observer/) {
        $out_filename = "NebraskaObserver-$date";
      } elsif ($name eq "") {
        # What is this? Skip it
        next;
      } else {
        $name =~ s/ /-/g;
        $out_filename = "$name-$date";
      }
      # Only run on specific files:
      if (
        $out_filename eq "WHAMO-1990-04-01"
#        $out_filename eq "NebraskaObserver-1990-08-02" ||
#        $out_filename eq "NebraskaObserver-1990-05-02"
#        $out_filename eq "NebraskaObserver-1989-04-26"
      ) {
        # p $this_issue, as => "Issue complete! adding!";
        $issues{$out_filename} = $this_issue;
      }
    }
    $this_issue = [];
    next;
  }
  ($tiff, $name, $subject, $comments, $authors) = split m/\|/;
  # say "  Adding $tiff to \$this_issue";
  push @$this_issue, "$in_dir/$tiff";
  # say $tiff;
}

foreach my $i (sort keys %issues) {
  my @tiffs = @{$issues{$i}};
  printf("%s %s\n", $i, ("." x scalar(@tiffs)));

  # tiffcp ../Image_20221212_0098-002.tiff ../Image_20221212_0099-001.tiff ../Image_20221212_0099-002.tiff ../Image_20221212_0098-001.tiff output.tiff
  my $cmd = "tiffcp " . (join " ", @tiffs) . " $out_dir/$i.tiff";
  say "  $cmd";
  system($cmd);

  # tiff2pdf output.tiff > output.pdf
  $cmd = "tiff2pdf $out_dir/$i.tiff > $out_dir/${i}_no_ocr.pdf";
  say "  $cmd";
  system($cmd);

  # ocrmypdf output.pdf output_ocr.pdf
  $cmd = "ocrmypdf $out_dir/${i}_no_ocr.pdf $out_dir/$i.pdf";
  say "  $cmd";
  system($cmd);

  print "\n";
}

sub convert_date {
  my $in = shift;
  return undef unless $in;
  if ($in =~ /^\w+, \d+/) {
    # e.g. "March, 1988". We'll pretend it was the 1st of the month
    $in =~ s/^(\w+), (\d+)/$1 1, $2/;
  }
  # say "sending $in";
  my $dt = $dft->parse($in);
  return $dt->ymd;
}


__END__
Image_20221212_0098-002.tiff|World-Herald Attitude Monitoring Operation (WHAMO)|February 28, 1990|Vol. VI, #3, p.1|Frances Mendenhall, Publisher
Image_20221212_0099-001.tiff|World-Herald Attitude Monitoring Operation (WHAMO)|February 28, 1990|Vol. VI, #3, p.2|Frances Mendenhall, Publisher
Image_20221212_0099-002.tiff|World-Herald Attitude Monitoring Operation (WHAMO)|February 28, 1990|Vol. VI, #3, p.3|Frances Mendenhall, Publisher
Image_20221212_0098-001.tiff|World-Herald Attitude Monitoring Operation (WHAMO)|February 28, 1990|Vol. VI, #3, p.4|Frances Mendenhall, Publisher

Image_20221212_0040-002.tiff|The Nebraska Observer|May 24, 1989|Vol. IV, #5, p.1|Frances Mendenhall, Publisher
Image_20221212_0041-001.tiff|The Nebraska Observer|May 24, 1989|Vol. IV, #5, p.2|Frances Mendenhall, Publisher
Image_20221212_0041-002.tiff|The Nebraska Observer|May 24, 1989|Vol. IV, #5, p.2|Frances Mendenhall, Publisher
Image_20221212_0042-002.tiff|The Nebraska Observer|May 24, 1989|Vol. IV, #5, p.3|Frances Mendenhall, Publisher
Image_20221212_0043-001.tiff|The Nebraska Observer|May 24, 1989|Vol. IV, #5, p.4|Frances Mendenhall, Publisher
Image_20221212_0043-002.tiff|The Nebraska Observer|May 24, 1989|Vol. IV, #5, p.5|Frances Mendenhall, Publisher
Image_20221212_0042-001.tiff|The Nebraska Observer|May 24, 1989|Vol. IV, #5, p.6|Frances Mendenhall, Publisher