Changeset 126 for trunk/devel

Timestamp:

08.01.2007 01:01:47 (4 years ago)

Author:

decoder

Message:

Added focr_max_height/width

Fixed bug with stripped spaces matching vs. non stripped spaces matching counting words twice

Location:

trunk/devel

Files:

• FuzzyOcr.cf (modified) (1 diff)
• FuzzyOcr.pm (modified) (2 diffs)
• FuzzyOcr/Config.pm (modified) (1 diff)

trunk/devel/FuzzyOcr.cf

@@ -179 +179 @@
 # ones specified here will be skipped:
 # Default: Height:4 Width:4
+#
 #focr_min_height 4
 #focr_min_width 4
+
+# Maximum image size to scan. Images with dimensions bigger than the
+# ones specified here will be skipped:
+# Default: Height:800 Width:800
+#
+#focr_max_height 800
+#focr_max_width 800
+
 
 # Maximum file size for different formats in byte, bigger pictures 

trunk/devel/FuzzyOcr.pm

@@ -249 +249 @@
         }
 
+        #Skip too big images, screenshots etc
+        if ($imgfiles{$filename}{height} > $conf->{focr_max_height}) {
+            infolog("Skipping image: height > $conf->{focr_max_height}");
+            delete $imgfiles{$filename};
+            next;
+        }
+
+        #Skip too big images, screenshots etc
+        if ($imgfiles{$filename}{width} > $conf->{focr_max_width}) {
+            infolog("Skipping image: width > $conf->{focr_max_width}");
+            delete $imgfiles{$filename};
+            next;
+        }
+
         #Found Image!! Get a temporary dir to save image
         $imgdir = Mail::SpamAssassin::Util::secure_tmpdir();
@@ -782 +796 @@
             debuglog("ocrdata=>>".join("",@result)."<<=end");
             foreach $modus (0 .. 1) {
+                $cmcnt = 0;
+                @cfound = ();
                 foreach my $ww (keys %words) {
                     my $w = lc $ww;

trunk/devel/FuzzyOcr/Config.pm

@@ -204 +204 @@
             setting => 'focr_min_'.$t,
             default => 4,
+            type => $Mail::SpamAssassin::Conf::CONF_TYPE_NUMERIC
+        });
+        push (@cmds, {
+            setting => 'focr_max_'.$t,
+            default => 800,
             type => $Mail::SpamAssassin::Conf::CONF_TYPE_NUMERIC
         });