Plucene Search Engine Add-On

TWiki original search engine is a simple yet powerful tool. However, it can not search within attached documents. That has been discused in many topics in the Codev web:

• TWiki:Codev/ImprovedSearchByKeywordIndex
• TWiki:Codev/SearchAttachmentContent
• TWiki:Codev/SearchAttachments
• and many others, just look at TWiki:Codev/SearchEnhancements which lists more than 100 topics about search issues

Time ago I found Plucene, which is a Perl port of the java library Lucene. So this plugin/addon intends to be a topic/attachment search engine, with Plucene as its backend.

I would like to thank TWiki:Main.SopanShewale for his many suggestions and contributions.

Note that this plugin have a release for each TWiki major version, namely Cairo and Dakar.

Usage

Indexing with plucindex

The plucindex script indexes all the public webs, and it uses some TWiki::Func code to retrieve the list of available webs and to retrieve their topic list. For each topic, the meta data is inspected and indexed, as the text body. Also, if the topic has attachments, those are indexed (see below for more details).

By now, you should run this script manually after installation to create the index files used by plucsearch. If you want, you can also schedule a weekly or monthly crontab job to create the index files again, or maybe execute it manually when you take down your server for maintenance tasks. To prevent browser access, it has been placed out of the public bin folder.

Please, to suggest indexing improvements read/post to TWiki:Plugins/SearchEnginePluceneAddOnDev

Updating with plucupdate

The plucupdate script uses the web's .changes files to know about topic modifications, in a way such old mailnotify worked. Also, a .plucupdate file is used on each web directory storing the last timestamp the script was run on it. So when this script is executed, first checks if there are any topic updates since last execution. The most recent topic updates are removed from the index and then reindexed again (the same goes for attachments).

This script should be executed by an hourly crontab. As before, this script has been placed out of the public bin folder.

Please, to suggest indexing improvements read/post to TWiki:Plugins/SearchEnginePluceneAddOnDev

Attachment file types to be indexed

All the PDF, HTML and text attachments are also indexed by default. If you want to override this setting you can use a TWiki preference PLUCENEINDEXEXTENSIONS. The DOT before the extension type is required. You can copy & paste the next lines in your TWiki.TWikiPreferences topic

   * Plucene settings
      * Set PLUCENEINDEXEXTENSIONS = .pdf, .html, .txt, .doc

or whatever extensions you want. By default, Plucene comes with PDF, HTML and TXT file support. However, PDF needs additional software to be installed (see intall instructions).

You may need additional CPAN:Plucene::SearchEngine::Index libraries and install additional third party tools such as antiword or xlhtml which provide required text extracting capabilities. You can find/post additional CPAN:Plucene::SearchEngine::Index libraries for many file types at TWiki:Plugins/SearchEnginePluceneAddOnDev. Thanks again to TWiki:Main/SopanShewale for his contributions.

Searching with plucsearch

The plucsearch script uses a template plucsearch.tmpl (that can be adapted to your site skin easily) or the plucsearch.pattern.tmpl (if you use the pattern skin). There is also a PluceneSearch topic with a form ready to use with the plucsearch script.

The query syntax has been improved

• you can use + for and and - for and not
• you can limit to the topic body or attachment body, using the prefix text: or just type the search string
• if you want to search using some meta data, you should use the prefix field: where field is the meta data name (like author)
• if you want to search using some form field, you should use the prefix field: where field is the form's field name
• plucene adds the type field for the indexed attachments, so you can use it to filter your results (like type:pdf)
• attachments also have a special field, attachment:yes, which is used in the PluceneSearch topic to search again only displaying attachments

Query examples (just type it in your PluceneSearch site topic)

• text:plucene searches for plucene in topic/attachment text
• plucene as above
• author:JoanMVigo searches for topics/attachments authored by this author
• TopicClassification:ItemToDo searches for topics with a form field named TopicClassification with value ItemToDo
• +perl -type:pdf +attachment:yes searches for attachments only with perl as text, excluding PDF files

Please, to suggest searching improvements read/post to TWiki:Plugins/SearchEnginePluceneAddOnDev

Other features

This new version provides some extra functionality:

• skip unuseful webs from the index (with a new preference PLUCENEINDEXSKIPWEBS)
  • all other webs are indexed, however if a web has Set NOSEARCHALL = on in its WebPreferences, then no topic from that web is shown when displaying results
• skip annoying or unindexable attachments from the index (with a new preference PLUCENEINDEXSKIPATTACHMENTS)
• index variables for web (with a new preference PLUCENEINDEXVARIABLES). For example, if set to CONTACTINFO, a search for CONTACTINFO:JohnSmith will provide the WebHome topics of the webs which have Set CONTACTINFO = JohnSmith in its WebPreferences.
• displaying the search results, show an option for diaplaying only attachments if PLUCENESEARCHATTACHMENTSONLY enabled. You can set PLUCENESEARCHATTACHMENTSONLYLABEL to a text or an image.

Please, to request further features read/post to TWiki:Plugins/SearchEnginePluceneAddOnDev

Search form

The following form submits text to the plucsearch script. The installation instructions are detailed below.

| Help

Add-On Installation Instructions

Note: You do not need to install anything on the browser to use this add-on. The following instructions are for the administrator who installs the add-on on the server where TWiki is running.

• You can install Plucene and its dependencies running:
  • perl -MCPAN -e "install Plucene"
  • perl -MCPAN -e "install Plucene::SearchEngine"
• Install third party text extracting tools, like xpdf which provides pdftotext. OPTIONAL You may wish to install additional CPAN:Plucene::SearchEngine::Index libraries so that this add on can index such file types. More information at TWiki:Plugins/SearchEnginePluceneAddOnDev#ExtraBackendParsers
• Download the ZIP file from the Add-on Home (see below)
• Unzip SearchEnginePluceneAddOn.zip in your twiki installation directory. Content:

  File:	Description:
  bin/plucsearch	script that searches the index files
  data/TWiki/PluceneSearch.txt	Plucene search topic
  data/TWiki/PluceneSearch.txt,v	Plucene search topic repository
  data/TWiki/SearchEnginePluceneAddOn.txt	Add-on topic
  data/TWiki/SearchEnginePluceneAddOn.txt,v	Add-on topic repository
  templates/plucsearch.pattern.tmpl	template used by new search script for the pattern skin
  plucene/bin/LocalLib.cfg	this file should is required and should be modified according to the twiki/lib absolute path of your installation
  plucene/bin/plucindex	script that indexes all topics and PDF/HTML/TXT attachments
  plucene/bin/plucupdate	script that uses web's .changes files to update the index
  plucene/index/	directory for index files to be stored
  plucene/logs/	the index and update logs will be written here - admin should monitor this folder

• ATTENTION! This add on now uses several preferences which should be set at TWiki.TWikiPreferences

   * Plucene settings
      * Set PLUCENEINDEXEXTENSIONS = .pdf, .htm, .html, .txt, .doc
      * Set PLUCENEINDEXPATH = /srv/www/twiki/plucene/index
      * Set PLUCENEATTACHMENTSPATH = /srv/www/twiki/pub
      * Set PLUCENESEARCHATTACHMENTSONLY = 1
      * Set PLUCENESEARCHATTACHMENTSONLYLABEL = Display only attachments
      * Set PLUCENEINDEXVARIABLES = CONTACTINFO, JUSTANOTHERONE
      * Set PLUCENEINDEXSKIPWEBS = Trash, Sandbox
      * Set PLUCENEINDEXSKIPATTACHMENTS = Web.SomeTopic.AnAttachment.txt, Web.OtherTopic.OtherAttachment.pdf
      * Set PLUCENEDEBUG = 1

• ATTENTION! Remember to edit the file plucene/bin/LocalLib.cfg and modify twikiLibPath accordingly to your configuration
• Test if the installation was successful:
  • change the working directory to the plucene/bin twiki installation directory
  • run ./plucindex
  • once finished, open a browser window and point it to the TWiki/PluceneSearch topic
  • just type a query and check the results
• Just create a new hourly crontab entry for the plucene/bin/plucupdate script.

Add-On Info

Add-on Author:	TWiki:Main/SopanShewale, TWiki:Main/JoanMVigo
Add-on Version:	27 Jun 2006 (v2.200 for Dakar, v1.500 for Cairo)
Change History:
27 Jun 2006:	TWikiDakar (v2.200) - Searching issue solved when using template authentication, update index bug solved
27 Jun 2006:	TWikiCairo (v1.500) - Update index bug solved
21 Mar 2006:	TWikiDakar (v2.100) & TWikiCairo (v1.400) - Update index issue solved
03 Mar 2006:	TWikiDakar (v2.000) & TWikiCairo (v1.300)
15 Dec 2004:	Use of TWiki preferences for indexing path & attachment extensions (v1.210)
26 Nov 2004:	TWikiCairo release compatible version (v1.200)
23 Nov 2004:	Incremental version (v1.100)
18 Nov 2004:	Initial version (v1.000)
CPAN Dependencies:	CPAN:Bit::Vector::Minimal, CPAN:IO::Scalar, CPAN:Lingua::GL::Stemmer, CPAN:Lingua::PT::Stemmer, CPAN:Lingua::Stem::Fr, CPAN:Lingua::Stem::It, CPAN:Lingua::Stem::Ru, CPAN:Lingua::Stem::Snowball::Da, CPAN:Lingua::Stem::Snowball::No, CPAN:Lingua::Stem::Snowball::Se, CPAN:Text::German, CPAN:Lingua::Stem::En, CPAN:Tie::Array::Sorted, CPAN:Time::Piece, CPAN:Plucene, CPAN:Plucene::SearchEngine
Other Dependencies:	xpdf (pdftotext) and additional 3rd party tools for text extracting
Perl Version:	Tested with 5.8.0
License:	GPL
Add-on Home:	http://TWiki.org/cgi-bin/view/Plugins/SearchEnginePluceneAddOn
Feedback:	http://TWiki.org/cgi-bin/view/Plugins/SearchEnginePluceneAddOnDev
Appraisal:	http://TWiki.org/cgi-bin/view/Plugins/SearchEnginePluceneAddOnAppraisal

-- TWiki:Main/JoanMVigo - 27 Jun 2006