RE: [docbook-apps] search engine with docbook

["David Cramer" <dcramer@motive.com>]

The Javascript Offline Search library looks interesting. I hope you submit your changes adding utf-8 support back to the project. I would be interested in using it as part of a "webhelp for docbook" thing I've started putting together. I've used a indexer from the htmlsearch plugin from the DITA project which sounds similar to the JS Offline Search library, but is less formally supported as a project (i.e. there's no project page...it's just a zip file available from the files page of the DITA yahoogroup).

 

I've posted the webhelp for docbook thing here: http://www.thingbag.net/docbook/docbook-webhelp/index.html This should provide the original poster with enough to get started for generating webhelp-like output from DocBook and could be used pretty much as-is.

 

As I mention in that link, the support for Asian languages is limited due to the fact that the indexer depends on spaces and punctuation to identify words. Since Japanese lacks spaces between words and has little punctuation, I've customized the xsls to add indexterms to the list of keywords. So if you use copious keywords and indexterms, at lest some searchs will return at least results for Japanese content. Perhaps this is good enough for some people.

 

David

---

From: Skopik Pavel [mailto:Pavel.Skopik@aipsafe.cz]
Sent: Monday, December 15, 2008 2:13 AM
To: BG; docbook-apps@lists.oasis-open.org
Subject: RE: [docbook-apps] search engine with docbook

Hello,

 

in or webhelp we have been using Javascript Offline Search library for some time (http://sourceforge.net/projects/jsofflinesearch/) with success. You can have a look here how it works: http://www.aipsafe.cz/files/tutorial/index.html.

 

It is a java library which indexes set of html files and builds a javascript index. It has some nice features – you can use boolean operators or search for phrase. More importantly it also implementes basic stemming.  The only issue is that it does not support other characters then basic ascii, so we had to alter the code to be able to index character with diacritics and ouput index files in UTF-8.

 

No packages have been released yet, but you can donwload it from svn and compile it yourself. It is then easy to automate the indexing process for example using ant. The only thing you have to do is to generate a xml file with list of all resulting html files to index which is then passed to the java library. This can be done in your customization layer during docbook source files processing.

 

If you are interested I can prepare a package (basicaly xslt template for generating files to index)  to share with the DocBook community and some thorough usage info.

 

Greetings,

 

Pavel

 

 

-----Original Message-----
From: BG [mailto:llo-143e5@myamail.com]
Sent: Friday, December 12, 2008 3:49 PM
To: docbook-apps@lists.oasis-open.org
Subject: [docbook-apps] search engine with docbook

 

Hi,
I realize an online help (html) with docbook and I would like to integrate a search engine.
Are there solutions to integrate a search engine with docbook?

Moreover I am a little amazed because having a search engine in an online help is very common, nevertheless I don't find many solutions on the subject.

Thank you,
BG

--------------------------------------------------------------------- To unsubscribe, e-mail: docbook-apps-unsubscribe@lists.oasis-open.org For additional commands, e-mail: docbook-apps-help@lists.oasis-open.org