Changeset 10280

Timestamp:

2005-07-25T14:19:14+12:00 (19 years ago)

Author:

chi

Message:

Some major changes to allow secondary plugin setting.

File:

• trunk/gsdl/perllib/plugins/BasPlug.pm (modified) (15 diffs)

trunk/gsdl/perllib/plugins/BasPlug.pm

@@ -213 +213 @@
 
 
-
-
 sub get_arguments
 {
@@ -386 +384 @@
     die "\n";
     }
+
 
     # else parsing was successful.
@@ -431 +430 @@
         $self->{'places'} = $places_ref;
     }
-    }   
+    }
     return bless $self, $class;
     
@@ -673 +672 @@
 
 
-# The BasPlug read() function. This function does all the right things
-# to make general options work for a given plugin. It calls the process()
-# function which does all the work specific to a plugin (like the old
-# read functions used to do). Most plugins should define their own 
-# process() function and let this read() function keep control.
-#
-# recursive plugins (e.g. RecPlug) and specialized plugins like those
-# capable of processing many documents within a single file (e.g.
-# GMLPlug) should normally implement their own version of read()
-#
-# Return number of files processed, undef if can't recognise, -1 if can't
-# process
-# Note that $base_dir might be "" and that $file might 
-# include directories
-
-sub read {
+sub read_block {
     my $self = shift (@_);  
   
     my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
 
-    if ($self->is_recursive()) {
-    gsprintf(STDERR, "{BasPlug.read_must_be_implemented}") && die "\n";
-    }
-
-    my $outhandle   = $self->{'outhandle'};
-    my $smart_block = $self->{'smart_block'};
-    my $smart_block_BN = $self->{'smart_block_BN'};
 
     my $filename = $file;
@@ -707 +684 @@
     # a form of smart block
     $self->{'num_blocked'} ++;
-    return 0; # blocked
-    }
+    return (0,undef); # blocked
+    }
+
+    my $smart_block = $self->{'smart_block'};
+    my $smart_block_BN = $self->{'smart_block_BN'};
 
     if ($smart_block || $smart_block_BN) {
@@ -714 +694 @@
     if (defined $self->{'file_blocks'}->{$filename} && $self->{'file_blocks'}->{$filename} == 1){
         $self->{'num_blocked'} ++;
-        return 0; # blocked
+        return (0,undef); # blocked
     }
     } elsif ($self->{'block_exp'} ne "" && $filename =~ /$self->{'block_exp'}/) {
     $self->{'num_blocked'} ++;
-    return 0; # blocked
+    return (0,undef); # blocked
     }
     
     if ($filename !~ /$self->{'process_exp'}/ || !-f $filename) {
-    return undef; # can't recognise
-    }
+    return (undef,undef); # can't recognise
+    }
+    
+    return (1,$filename);
+}
+
+sub read_tidy_file {
+
+    my $self = shift (@_);  
+  
+    my ($file) = @_;
+
     $file =~ s/^[\/\\]+//; # $file often begins with / so we'll tidy it up
+
+    return $file;
+}
+
+
+
+# The BasPlug read_into_doc_obj() function. This function does all the
+# right things to make general options work for a given plugin.  It reads in
+# a file and sets up a slew of metadata all saved in doc_obj, which
+# it then returns as part of a tuple (process_status,doc_obj)
+#
+# Much of this functionality used to reside in read, but it was broken
+# down into a supporting routine to make the code more flexible.  
+#
+# recursive plugins (e.g. RecPlug) and specialized plugins like those
+# capable of processing many documents within a single file (e.g.
+# GMLPlug) will normally want to implement their own version of
+# read_into_doc_obj()
+#
+# Note that $base_dir might be "" and that $file might 
+# include directories
+sub read_into_doc_obj {
+    my $self = shift (@_);  
+    my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
+
+    if ($self->is_recursive()) {
+    gsprintf(STDERR, "{BasPlug.read_must_be_implemented}") && die "\n";
+    }
+
+    my $outhandle   = $self->{'outhandle'};
+
+    my ($block_status,$filename) = $self->read_block(@_);    
+    return $block_status if ((!defined $block_status) || ($block_status==0));
+    $file = $self->read_tidy_file($file);
 
     # Do encoding stuff
@@ -760 +784 @@
     $self->{'num_not_processed'} ++;
 
-    return 0; # what should we return here?? error but don't want to pass it on
+    return (0,undef); # what should we return here?? error but don't want to pass it on
     }
    
@@ -773 +797 @@
     undef $text;
     print STDERR "<ProcessingError n='$file'>\n" if ($gli);
-    return -1;
+    return (-1,undef);
     }
     $text='';
@@ -790 +814 @@
     $doc_obj->set_OID();
     }
-
-    # process the document
-    $processor->process($doc_obj);
-
-    if(defined($self->{'places_filename'})){
-    &util::rm($self->{'places_filename'});
-    $self->{'places_filename'} = undef;
-    }
-
-    $self->{'num_processed'} ++;
-    undef $doc_obj;
-    return 1; # processed the file
+    
+    return (1,$doc_obj);
+}
+
+
+# The BasPlug read() function. This function calls read_into_doc_obj()
+# to ensure all the right things to make general options work for a
+# given plugin are done. It then calls the process() function which
+# does all the work specific to a plugin (like the old read functions
+# used to do). Most plugins should define their own process() function
+# and let this read() function keep control.  
+#
+# recursive plugins (e.g. RecPlug) and specialized plugins like those
+# capable of processing many documents within a single file (e.g.
+# GMLPlug) might want to implement their own version of read(), but
+# more likely need to implement their own version of read_into_doc_obj()
+#
+# Return number of files processed, undef if can't recognise, -1 if can't
+# process
+
+sub read {
+    my $self = shift (@_);  
+    my ($pluginfo, $base_dir, $file, $metadata, $processor, $maxdocs, $total_count, $gli) = @_;
+
+    my ($process_status,$doc_obj) = $self->read_into_doc_obj(@_);
+    
+    if ((defined $process_status) && ($process_status == 1)) {
+    # process the document
+    $processor->process($doc_obj);
+    
+    if(defined($self->{'places_filename'})){
+        &util::rm($self->{'places_filename'});
+        $self->{'places_filename'} = undef;
+    }
+    
+    $self->{'num_processed'} ++;
+    undef $doc_obj; 
+    }
+
+    # if process_status == 1, then the file has been processed.
+    return $process_status;
+
 }
 
@@ -830 +884 @@
     return;
     }
-
     $$textref = "";
-
     if (!open (FILE, $filename)) {
     gsprintf(STDERR, "BasPlug::read_file {BasPlug.could_not_open_for_reading} ($!)\n", $filename);
-     die "\n";
-     }
-
+    die "\n";
+    }
+     
     if ($encoding eq "ascii") {
     undef $/;
@@ -847 +899 @@
     $reader->set_encoding ($encoding);
     $reader->read_file ($textref);
-
-    #Now segments chinese if the separate_cjk option is set
+        #Now segments chinese if the separate_cjk option is set
     if ($self->{'separate_cjk'}) {
         # segment the Chinese words
@@ -854 +905 @@
     }
     }
-
     close FILE;
 }
+
+# write_file -- used by ConvertToPlug, for example in post processing
+#
+sub utf8_write_file {
+    my $self = shift (@_);
+    my ($textref, $filename) = @_;
+    
+    if (!open (FILE, ">$filename")) {
+    gsprintf(STDERR, "ConvertToPlug::write_file {ConvertToPlug.could_not_open_for_writing} ($!)\n", $filename);
+     die "\n";
+     }
+    print FILE $$textref;
+    
+    close FILE;
+}
+
 
 sub filename_based_title
@@ -887 +953 @@
     my ($filename) = @_;
 
+    
     my ($language, $encoding, $extracted_encoding);
     if ($self->{'input_encoding'} eq "auto") {
@@ -910 +977 @@
         $encoding = $self->{'input_encoding'};
     }
+
     return ($language, $encoding);
 }
@@ -942 +1010 @@
     }
     }
-
+   
 
     # remove <title>stuff</title> -- as titles tend often to be in English