Changeset 16

Timestamp:

28.10.2006 13:09:24 (2 years ago)

Author:

decoder

Message:

Added maximum file sizes for all formats (untested code)
Changed config file accordingly

Files:

• trunk/devel/FuzzyOcr.cf (modified) (1 diff)
• trunk/devel/FuzzyOcr.pm (modified) (8 diffs)

trunk/devel/FuzzyOcr.cf

@@ -71 +71 @@
-# Timeout for the plugin, in seconds. (Maximum runtime of the plugin) (Default value: 10)
-#focr_timeout 15
+#
+# Maximum file size for different formats in byte, bigger pictures will not be scanned (Default values: Unlimited)
+#focr_max_size_gif 80000
+#focr_max_size_jpeg 100000
+#focr_max_size_png 80000
+#focr_max_size_bmp 500000
+#focr_max_size_tiff 500000
-#
-# Default detection treshold (see manual) (Default value: 0.3) (Can be changed on a per word basis in the wordlist).

trunk/devel/FuzzyOcr.pm

@@ -15 +15 @@
-
-use String::Approx 'adistr';
-#use Image::Magick;
-use MLDBM qw(DB_File Storable);
-use FileHandle;
@@ -38 +37 @@
-
-our @pgm_opts = qw/personal_wordlist global_wordlist logfile
-
-
-
+
+
+
+
-
-our @paths = qw(/usr/local/netpbm/bin /usr/local/bin /usr/bin);
@@ -50 +50 @@
-$Option{timeout} = 10;
-$Option{logfile} = undef;
+#We should define these max_size values once we have good default values
+$Option{max_size_gif} = undef;
+$Option{max_size_jpeg} = undef;
+$Option{max_size_png} = undef;
+$Option{max_size_tiff} = undef;
+$Option{max_size_bmp} = undef;
-$Option{enable_image_hashing} = 0;
-$Option{hashing_learn_scanned} = 1;
@@ -839 +845 @@
-            debuglog("Found GIF header name=\"$$pic{fname}\"");
-            $ptype = 1;
+            my $s = (stat($file))[7];
+            if (defined($Option{max_size_gif}) and ($s > $Option{max_size_gif})) {
+                debuglog("GIF file size ($s) exceeds maximum file size for this format, skipping...");
+                next;
+            }
+
-            if ( $$pic{ctype} !~ /gif/i ) {
-                wrong_ctype( "GIF", $$pic{ctype} );
@@ -982 +994 @@
-            debuglog("Found JPEG header name=\"$$pic{fname}\"");
-            $ptype = 2;
+            my $s = (stat($file))[7];
+
+            if (defined($Option{max_size_jpeg}) and ($s > $Option{max_size_jpeg})) {
+                debuglog("JPEG file size ($s) exceeds maximum file size for this format, skipping...");
+                next;
+            }
-            if ( $$pic{ctype} !~ /(jpeg|jpg)/i ) {
-                wrong_ctype( "JPEG", $$pic{ctype} );
@@ -1004 +1022 @@
-            debuglog("Found PNG header name=\"$$pic{fname}\"");
-            $ptype = 3;
+            my $s = (stat($file))[7];
+
+            if (defined($Option{max_size_png}) and ($s > $Option{max_size_png})) {
+                debuglog("PNG file size ($s) exceeds maximum file size for this format, skipping...");
+                next;
+            }
-            if ( $$pic{ctype} !~ /png/i ) {
-                wrong_ctype( "PNG", $$pic{ctype} );
@@ -1027 +1051 @@
-            debuglog("Found BMP header name=\"$$pic{fname}\"");
-            $ptype = 4;
+            my $s = (stat($file))[7];
+            if (defined($Option{max_size_bmp}) and ($s > $Option{max_size_bmp})) {
+                debuglog("BMP file size ($s) exceeds maximum file size for this format, skipping...");
+                next;
+            }
-            if ( $$pic{ctype} !~ /bmp/i ) {
-                wrong_ctype( "BMP", $$pic{ctype} );
@@ -1053 +1082 @@
-            debuglog("Found TIFF header name=\"$$pic{fname}\"");
-            $ptype = 5;
+            my $s = (stat($file))[7];
+            if (defined($Option{max_size_tiff}) and ($s > $Option{max_size_tiff})) {
+                debuglog("TIFF file size ($s) exceeds maximum file size for this format, skipping...");
+                next;
+            }
-            if ( $$pic{ctype} !~ /tiff/i ) {
-                wrong_ctype( "TIFF", $$pic{ctype} );