Changeset 108

Timestamp:

11.12.2006 17:38:57 (2 years ago)

Author:

jorge

Message:

Added: Option to not log PMS(msgid,from,to) These are logged in spamd.
Change: Automatically skip images with width < 4 pixels (tall images)
Change: Automatically skip images with height < 4 pixels (flat images)
Fixed: Handling of BMP images (wrongly decoded)
Fixed: Exiting subroutine via next at FuzzyOcr?.pm line 689.

Files:

• trunk/devel/FuzzyOcr.cf (modified) (1 diff)
• trunk/devel/FuzzyOcr.pm (modified) (19 diffs)
• trunk/devel/FuzzyOcr/Config.pm (modified) (1 diff)

trunk/devel/FuzzyOcr.cf

@@ -33 +33 @@
-# Default value: 1
-#focr_verbose 3
+
+# Log Message-Id, From, To
+# Default: 1
+#focr_log_pmsinfo 0
-
-# Send logging output to stderr.

trunk/devel/FuzzyOcr.pm

@@ -125 +125 @@
-    my $main     = $self->{main};
-
-    my $from = $pms->get('From') ? $pms->get('From') : "<no sender>";
-    my $to = $pms->get('To') ? $pms->get('To') : "<no receipients>";
-    my $msgid = $pms->get('Message-Id') ? $pms->get('Message-Id') : "<no messageid>";
-
-    chomp($from, $to, $msgid);
-
-    debuglog("Starting FuzzyOcr...");
-
+
+
+
+
+
+
+
+
+
+
-    foreach my $p (
-        $pms->{msg}->find_parts(qr(^image\b)i),
@@ -152 +155 @@
-        my $w = 0; my $h = 0;
-
-        my $blah = substr($pdata,0,3);
-
-        if ( substr($pdata,0,3) eq "\x47\x49\x46" ) {
-            ## GIF File
-            $imgfiles{$filename}{ftype} = 1; 
-            ($w,$h) = unpack("vv",substr($pdata,6,4));
-
+ ($pdatalen)
-            $imgfiles{$filename}{width}  = $w;
-            $imgfiles{$filename}{height} = $h;
@@ -185 +186 @@
-            } else {
-                ($h,$w) = unpack("nn",substr($pdata,$pos+3,4));
-
+ ($pdatalen)
-                $imgfiles{$filename}{ftype} = 2;
-                $imgfiles{$filename}{height} = $h;
@@ -196 +197 @@
-            $imgfiles{$filename}{width}  = $w;
-            $imgfiles{$filename}{height} = $h;
-
+ ($pdatalen)
-        } elsif ( substr($pdata,0,2) eq "BM" ) {
-            ## BMP File
-NN
+VV
-            $imgfiles{$filename}{ftype}  = 4;
-            $imgfiles{$filename}{width}  = $w;
-            $imgfiles{$filename}{height} = $h;
-
+ ($pdatalen)
-        } elsif (
-            ## TIFF File
@@ -219 +220 @@
-                last if ($h != 0 and $w != 0);
-            }
-
+pdatalen) ($
-            infolog("Cannot determine size of TIFF image, setting to '1x1'") if ($h == 0 and $w == 0);
-            $imgfiles{$filename}{ftype}  = 5;
@@ -231 +232 @@
-            delete $imgfiles{$filename};
-            next;
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-        }
-
@@ -309 +323 @@
-    }
-
-    IMAGE:
-    my $haserr;
-    foreach my $filename (keys %imgfiles) {
@@ -352 +365 @@
-            if ($conf->{focr_skip_gif}) {
-                infolog("Skipping image check");
- IMAGE
+
-            }
-            if (defined($conf->{focr_max_size_gif}) and ($$pic{fsize} > $conf->{focr_max_size_gif})) {
@@ -375 +388 @@
-                unless (defined $conf->{"focr_bin_$a"}) {
-                    errorlog("Cannot exec $a, skipping image");
-
-
-
+
+
+
+
-            my @stderr_data;
-
-            my ($retcode, @stdout_data) = save_execute(
-                "$conf->{focr_bin_giftext} $file",
@@ -504 +517 @@
-            if ($conf->{focr_skip_jpeg}) {
-                infolog("Skipping image check");
- IMAGE
+
-            }
-
@@ -524 +537 @@
-                unless (defined $conf->{"focr_bin_$a"}) {
-                    errorlog("Cannot exec $a, skipping image");
- IMAGE
+
-                }
-            }
@@ -546 +559 @@
-            if ($conf->{focr_skip_png}) {
-                infolog("Skipping image check");
- IMAGE
+
-            }
-            if (defined($conf->{focr_max_size_png}) and ($$pic{fsize} > $conf->{focr__max_size_png})) {
@@ -563 +576 @@
-                unless (defined $conf->{"focr_bin_$a"}) {
-                    errorlog("Cannot exec $a, skipping image");
- IMAGE
+
-                }
-            }
@@ -586 +599 @@
-            if ($conf->{focr_skip_bmp}) {
-                infolog("Skipping image check");
- IMAGE
+
-            }
-            if (defined($conf->{focr_max_size_bmp}) and ($$pic{fsize} > $conf->{focr_max_size_bmp})) {
@@ -603 +616 @@
-                unless (defined $conf->{"focr_bin_$a"}) {
-                    errorlog("Cannot exec $a, skipping image");
- IMAGE
+
-                }
-            }
@@ -625 +638 @@
-            if ($conf->{focr_skip_tiff}) {
-                infolog("Skipping image check");
- IMAGE
+
-            }
-            if (defined($conf->{focr_max_size_tiff}) and ($$pic{fsize} > $conf->{focr_max_size_tiff})) {
@@ -643 +656 @@
-                unless (defined $conf->{"focr_bin_$a"}) {
-                    errorlog("Cannot exec $a, skipping image");
- IMAGE
+
-                }
-            }
@@ -689 +702 @@
-                if ($score > 0) {
-                    infolog("Image in KNOWN_GOOD. Skipping OCR checks...");
- IMAGE
+
-                }
-            }
-            if ($digest eq '') {
-                infolog("Empty Hash, skipping...");
- IMAGE
+
-            }
-        } else {
@@ -833 +846 @@
-                        }
-                    }
-infolog("Resorting scanset list...")
+my $newlist
-                    @$scansets = sort { $b->{hit_counter} <=> $a->{hit_counter} } @$scansets;
+                    foreach my $s (@$scansets) {
+                        $newlist .= ' ' . $s->{label} . '(' . $s->{hit_counter} . ')';
+                    }
+                    infolog("New sorted list:$newlist");
+
-                }
-                last;

trunk/devel/FuzzyOcr/Config.pm

@@ -241 +241 @@
-    push (@cmds, {
-        setting => 'focr_log_stderr',
+        default => 1,
+        type => $Mail::SpamAssassin::Conf::CONF_TYPE_BOOL
+    });
+
+    push (@cmds, {
+        setting => 'focr_log_pmsinfo',
-        default => 1,
-        type => $Mail::SpamAssassin::Conf::CONF_TYPE_BOOL