images ocr

bin/autodoc_process.pl

 
 $Data::Dumper::Sortkeys = 1;
 
+print "Loading configuration\n";
 my $conf = load_conf("../etc/autodoc.json");
+
+print "Connecting to database\n";
 my $dbh  = sqlconnect($conf->{sql});
 
 my $queuedir    = $conf->{path}{global}.'/'.$conf->{path}{queue};
 my $originaldir = $conf->{path}{global}.'/'.$conf->{path}{original};
 
-my %langid;
 my %primary;
+
+print "Loading languages\n";
+my %langid;
 my $q = sqlquery($dbh, "SELECT id,short FROM lang");
 while(my($id,$short)=$q->fetchrow_array()) { $langid{$short}=$id; }
 
+print "Opening queue folder $queuedir\n";
 opendir(Q,$queuedir);
 foreach my $file ( readdir(Q) ) {
+	print "processing file $file\n";
 	if ( $file =~ /^([0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12})_([0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12})\.([a-z]+)$/ ) {
 		my $docid = $1;
 		my $ext = $3;
 
-		print "Found document id $docid of type $ext\n";
+		print "\tdocument id $docid of type $ext\n";
 
 		if ( $ext eq 'pdf' ) {
 			my @pages;
 			for(my $page=0;; $page++) {
 				my $txt = '';
-				print "texting page $page\n";
 				open(TXT,sprintf("pdftotext -f %s -l %s %s/%s - 2>/dev/null |", $page+1, $page+1, $queuedir, $file)) || last;
 				while(<TXT>) {
 					chomp;
 				# end of PDF
 				last if $?;
 
+				print "\t\textracted text from PDF for page $page\n";
+
+
 				my ($lang,$words) = detect_lang($txt);
-				print "language is $lang\n";
-				#spell_check($txt,$lang);
+				print "\t\tdetected language is $lang\n";
+
 				my $pageid = get_new_page($docid);
-				print "new page id $pageid\n";
+				print "\t\tcreated page id $pageid\n";
+				print "\t\tupdating page status to 'inprogress'\n";
 				update_page_status($pageid, 'inprogress');
 
 				if ( !exists $primary{$docid} ) {
+					print "\t\tsetting document for default primary thumbnail\n";
 					$primary{$docid}=undef;
 					sqlquery($dbh, "CALL set_primary_page(?)",$pageid);
 				}
 
 
-				print "create original page jpeg $pageid.jpeg";
+				print "\t\tcreating original page jpeg $pageid.jpeg";
 				system(sprintf("pdftoppm -f %s -l %s -r 300 -jpeg -singlefile %s/%s %s/%s", $page+1, $page+1, $queuedir, $file, $originaldir, $pageid));
 				system(sprintf("mv %s/%s.jpg %s/%s.jpeg",
 					$originaldir, $pageid,
 					$originaldir, $pageid ));
 
+				print "\t\tloading extracted words into database\n";
 				create_page_words($pageid, $lang, $words);
 
+				print "\t\tupdating page status to 'ok'\n";
 				update_page_status($pageid, 'ok');
 
 			}
 		}
+		elsif ( $ext =~ /^(jpeg|png)$/ ) {
+			print "\tdetecting image rotation\n";
+			for(my $rot=0; $rot<=360; $rot+=90) {
+				print "\t\ttrying $rot degrees rotation\n";
+				my $tempfile = "/tmp/autodoc.$$.jpeg";
+				system(sprintf("convert %s/%s -rotate %s %s",
+					$queuedir, $file, $rot, $tempfile));
+
+				my($lang,$words) = detect_lang(ocr_file($tempfile));
+
+				print Dumper($lang, $words);
+
+				unlink($tempfile);
+			}
+			print "\trunning OCR on page\n";
+		}
 		else {
+			print "\terror: don't know how to process files of $ext type";
 		}
 	}
+	else {
+		print "\terror: file doesn't contain manadatory UUIDs in its name\n";
+	}
 }
 closedir(Q);
 
+print "done\n";
+
 # open queue
 	# decompose PDF
 	# normalise all files as jpegs
 	# ocr / lang detect
 	# update db
 
+sub ocr_file {
+	my($file) = @_;
+	my $txt = '';
+
+	open(OCR,"tesseract -l eng+deu+fra+ita %s - |", $file));
+	while(<OCR>) {
+		$txt .= $_;		
+	}
+	close(OCR);
+	
+	return $txt;
+}
 sub create_page_words {
 	my($pageid, $lang, $words) = @_;