Changeset 2996

Timestamp:

2002-02-26T09:50:41+13:00 (22 years ago)

Author:

sjboddie

Message:

* empty log message *

File:

• trunk/gsdl/perllib/plugins/W3ImgPlug.pm (modified) (6 diffs)

trunk/gsdl/perllib/plugins/W3ImgPlug.pm

@@ -1 +1 @@
 ###########################################################################
 #
-# W3ImgPlug.pm -- Web image indexing plugin
+# W3ImgPlug.pm -- Context-based image indexing plugin for HTML documents
 #
 # A component of the Greenstone digital library software
@@ -25 +25 @@
 ###########################################################################
 
-#  Subclass of HTMLPlug, designed to
-#  extract images from HTML pagesimages 
-#  with meta-data and associated text
-#  allowing indexing of images
+# DESCRIPTION:
+#
+#  Extracts images and associated text and metadata from
+#  web pages as individual documents for indexing. Thumbnails
+#  are created from each image for browsing purposes.
+#
+#  Options are available for configuring the aggressiveness of the
+#  associated text extraction mechanisms. A higher level of
+#  aggressiveness will extract more text and consequently
+#  may mean lower accuracy (precision); however, it may also 
+#  retrieve more of the relevant images from the collection (recall). 
+#  Lower levels of aggressiveness maybe result in slightly faster
+#  collection builds at the import stage.
+#
+#  W3ImgPlug is a subclass of HTMLPlug (i.e. it will index pages also 
+#  if required). It can be used in place of HTMLPlug to index both
+#  pages and their images.
+#
+# REQUIREMENTS:
 #   
-#  can be used in place of HTMLPlug with -cache_pages
-#  to index the images to complement the page index
-
-# TODO: have a gsdl/bin/imgconvert prog that is different for Win vs Linux
-#       and calls 'convert' with parameters?
-# TODO: windows doesn't have `cksum` utility. subst with one in bin/windows?
-#       (currently cludges filename/ext instead)
-
+#  The ImageMagick image manipulation is used to create
+#  thumbnails and extract some image metadata. (Available 
+#  from http://www.imagemagick.org/)
+#
+#  Unix:
+#    Many Linux distributions contain ImageMagick.
+#
+#  Windows:
+#    ImageMagick can be downloaded from the website above.
+#    Make sure the system path includes the ImageMagick binaries
+#    before using W3ImgPlug.
+#
+#    NOTE: NT/2000/XP contain a filesystem utility 'convert.exe' 
+#    with the same name as the image conversion utility. The
+#    ImageMagick FAQ recommends renaming the filesystem
+#    utility (e.g. to 'fsconvert.exe') to avoid this clash.
+#
+# USAGE:  
+#
+#  An image document consists of metadata elements:
+#
+#   OriginalFilename, FilePath, Filename, FileExt, FileSize,
+#   Width, Height, URL, PageURL, ThumbURL, CacheURL, CachePageURL
+#   ImageText, PageTitle
+#
+#  Most of these are only useful in format strings (e.g. ThumbURL, 
+#  Filename, URL, PageURL, CachePageURL). 
+#
+#  ImageText, as the name suggests contains the indexable text.
+#  (unless using the -document_text plugin option)
+#
+#  Since image documents are made up of metadata elements 
+#  alone, format strings are needed to display them properly. 
+#  NOTE: The receptionist will only display results (e.g. thumbnails)
+#  in 4 columns if the format string begins with "<td><table>".
+#
+#  The example below takes the user to the image within the
+#  source HTML document rather than using a format string
+#  on DocumentText to display the image document itself.
+#
+#  Example collect.cfg:
+#
+#   ...
+#
+#   indexes document:ImageText document:text
+#   defaultindex document:ImageText
+#  
+#   collectionmeta .document:ImageText "images"
+#   collectionmeta .document:text "documents"
+#
+#   ...
+#
+#   plugin W3ImgPlug -index_pages -aggressiveness 6
+#
+#   ...
+#  
+#   format SearchVList '<td>{If}{[Title],[link][icon]&nbsp;[Title][[/link],
+#    <table><tr><td align="center"><a href="[CachePageURL]">
+#    <img src="[ThumbURL]"></a></td></tr><tr><td align="center">
+#    <a href="[CachePageURL]"><font size="-1">[OriginalFilename]</font></a>
+#    <br>[Width]x[Height]</td></tr></table>}</td>'
+#
+#   ...
+#
+ 
 package W3ImgPlug;
 
@@ -151 +223 @@
     }
     # else use URL for referencing
-#    if ( $file =~ /(.*)[\/\\]/ ) { $self->{'htpath'} = $1; } else { $self->{'htpath'} = $file; }
+    #if ( $file =~ /(.*)[\/\\]/ ) { $self->{'htpath'} = $1; } else { $self->{'htpath'} = $file; }
 
     $self->{'htpath'} = $base_dir if (-d $base_dir);
@@ -173 +245 @@
 
 # get complex configuration options from configuration files
-# -- $GSDLCOLLECTION/etc/W3ImgPlug.cfg
-# -- $GSDLHOME/etc/packages/phind/stopword/en/brown.sw
-# (only done once, at import start)
+# -- $GSDLCOLLECTION/etc/W3ImgPlug.cfg (tag sets for aggr 2+)
+# -- $GSDLHOME/etc/packages/phind/stopword/en/brown.sw (stopwords for aggr 5+)
 
 # If there's no W3ImgPlug.cfg file we'll use the following default values
@@ -350 +421 @@
     $orig_fp =~ tr/+/ /;
     $orig_fp =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C", hex($1))/eg; # translate %2E to space, etc
+    $orig_fp =~ s/\\/\//g;
     $filepath = "$self->{'htpath'}/$filepath";
     ($onlyfn) = $filename =~ /([^\\\/]*)$/;
@@ -944 +1016 @@
     my ($front, $link, $back, $base_dir, 
     $file, $doc_obj, $section) = @_;
+    $link =~ s/\"//g;
     my ($a_name) = $link;
     $a_name =~ s/[\/\\\:\&]/_/g;