[opencms-dev] Search engine

Thomas Weckert t.weckert at alkacon.com
Tue Jan 28 21:32:02 CET 2003


Hello Alexis,

I already ran into the same problem, tough I cannot remember the proper 
solution at the moment. If you receive the error message "htsearch 
returned more lines in it's output than expected!", then please check 
whether the last character of the last line in ${common_dir}/long.html 
is a newline character or not, and if so, please remove it!

Regards,
/Thomas

Thomas Weckert 
Alkacon Software - The OpenCms Experts 
http://www.alkacon.com 


Quoting agrandemange at amadeus.net:

> 
> Hi Thomas,
> 
> thank you for the explanation. I'm now a step further. Still the same error
> but not for the same reason:
> [28.01.2003 17:22:34] <opencms_info> Updated JSP file
> "/WEB-INF/jsp/offline/ThtdigsearchTindex.html.-1611620409.jsp" for resource
> "/htdigsearc
> h/index.html"
> [28.01.2003 17:22:35] <opencms_info> Updated JSP file
> "/WEB-
INF/jsp/offline/TsystemTmodulesTcom.alkacon.htdigTelementsTsearch_functions.jsp.
13
> 75346473.jsp" for resource
> "/system/modules/com.alkacon.htdig/elements/search_functions.jsp"
> [28.01.2003 17:22:38] <opencms_info> [com.alkacon.htdig.HtdigBean]
> java.lang.ArrayIndexOutOfBoundsException
> [28.01.2003 17:22:38] <opencms_info> [com.alkacon.htdig.HtdigBean] htsearch
> returned more lines in it's output than expected!
> Please check the format of ${common_dir}/long.html
> htsearch URL: http://ncegrid0/cgi-bin/htsearch
> htsearch parameters: format=long&words=oracle&matchesperpage=10&restrict
> =&config=opencms&page=1
> current line: 0
> 
> Though I use the long.html listed in the module doc it fails. I looked at
> the stream returned by htdig. Quite often I have lines like this:
> >>
> http://ncegrid1/oracle/doc/a85344/a85344/ch5.htm
> Configuring Net8
> <b><tt>... </tt></b> Intel: * Core Net8 Products and Features * Net8
> Protocol Support * The BEQ Protocol * The IPC Protocol * The RAW Protocol *
> The TCP/IP Protocol * <strong><a href
> ="http://ncegrid1/oracle/doc/a85344/a85344/ch5.htm#16264">Oracle</a></strong>
> 
> Enterprise Manager * Configuring <strong>Oracle</strong> Intelligent Agent
> for <strong>Oracle</strong> SNMP * <strong>Oracle</strong> Advanced
> Security * Supplementary Documentation Core Net8 Products<b><tt>
> ...</tt></b>
> 44792
> 38
> 01/17/2001
> >>
> It is quite correct except for the excerpt which is made of two lines (It
> might be displayed on more lines with your mailing tool).
> But the code assumes that the excerpt contains exactly one line:
> // the URL of the search result
> searchResultUrl = htsearchOutput[currentLine];
> // the title
> String searchResultTitle = htsearchOutput[currentLine + 1];
> // the excerpt
> String searchResultExcerpt = htsearchOutput[currentLine + 2];
> // the size (in kBytes)
> searchResultSize = Integer.parseInt(htsearchOutput[currentLine + 3]);
> // the hit rate (in percent)
> searchResultHitRate = Integer.parseInt(htsearchOutput[currentLine + 4]);
> // the last modified date string
> String searchResultLastModifiedTimestamp = htsearchOutput[currentLine + 5];
> // the type of the URL resource (might be a PDF!)
> int resourceType = HtdigBean.C_RESULT_TYPE_DEFAULT;
> 
> Actually with the long.html setting in the doc the excerpt is one or more
> lines long.
> Is it something I can set to circumvent the problem? Otherwise we need to
> put something special in the long.html after $(EXCERPT) and adds lines to
> the excerpt up to this special thing. Do you see another solution?
> 
> Best regards,
> Alexis
> 
> 
> 
> From:  Thomas Weckert <t.weckert at alkacon.com>@www.opencms.org  on
>        27/01/2003 22:58
> 
> Please respond to opencms-dev at www.opencms.org
> 
> Sent by:    owner-opencms-dev at www.opencms.org
> 
> 
>                                                                           
>  To:    opencms-dev at www.opencms.org                                       
>         agrandemange at amadeus.net                                          
>                                                                           
>                                                                           
>  cc:                                                                      
>                                                                           
>                                                                           
>                                                                           
>                                                                           
>                                                                           
>                                                                           
>  Subjec RE: [opencms-dev] Search engine                                   
>  t:                                                                       
>                                                                           
>                                                                           
>                                                                           
> 
> 
> 
> 
> 
> Hi Alexis,
> 
> it seems to me as if there is a typo in your htsearch URL. The htdig
> documentation for the config input parameter of the htsearch command
> says: "The name here is the name without the path and without the
> .conf at the end.". So config=htdig instead of config=htdig.conf
> would do the job. Use setHtsearchConfig() in HtdigBean to set the
> name of the config file before you invoke search().
> 
> Regards,
> /Thomas
> 
> Thomas Weckert
> Alkacon Software - The OpenCms Experts
> http://www.alkacon.com
> 
> 
> Quoting agrandemange at amadeus.net:
> 
> >
> > Alex,
> >
> > My opencms was in a poor condition between beta1 and RC1. Now it still
> > doesn't work but in a much cleaner way.
> > I get a "Our search database is currently maintenanced, please try again
> > later." message.
> > I looked at the query string in the code,
> > In my case it generates
> > http://localhost/cgi-bin/htsearch?
> format=long&words=oracle&matchesperpage=10&restrict
> >
> > &config=htdig.conf&page=1.
> > I tried on the browser and htdig 3.16 returns "Unable to read
> configuration
> > file" though my htdig config file is htdig.conf.
> > If I query
> > http://localhost/cgi-bin/htsearch?
> format=long&words=oracle&matchesperpage=10&restrict
> >
> > &page=1 (remove the config parameter) it works.
> > Do you have an idea?
> > I could remove the config parameter from the code but I find this
> parameter
> > handy. The OpenCms long.html setting is not suitable for browser queries.
> > It would really help having two or more configs.
> >
> > Regards,
> > Alexis
> >
> >
> >
> > From:  "Alexander Kandzior" <alex at opencms.org>@www.opencms.org  on
> >        10/01/2003 18:43
> >
> > Please respond to opencms-dev at www.opencms.org
> >
> > Sent by:    owner-opencms-dev at www.opencms.org
> >
> >
> >
> 
> >  To:    <opencms-dev at www.opencms.org>
> 
> >
> 
> >
> 
> >  cc:
> 
> >
> 
> >
> 
> >
> 
> >
> 
> >
> 
> >
> 
> >  Subjec RE: [opencms-dev] Search engine
> 
> >  t:
> 
> >
> 
> >
> 
> >
> 
> >
> >
> >
> >
> >
> > Alexis,
> >
> > you can download the pre-release version of the JSP/htdig search module
> > here:
> >
> > http://www.opencms.info/download/com.alkacon.htdig_4.zip (OpenCms
> > module)
> > http://www.opencms.info/download/com.alkacon.htdig_4_src.zip (Source
> > code)
> >
> > To run the search engine integration you just need the module, it
> > contains the compiled java classes. Download the file and import it in
> > OpenCms using the "Module Management" feature in the "Administration"
> > view. You must be in the "Online" project to do this.
> >
> > You must restart the OpenCms/Tomcat server after installing the module.
> >
> > After installing the module you have a directory "/htdigsearch/" which
> > contains a demo search page and some configuration hints. You still will
> > need to read the ht:dig documentation on how to install ht:dig first.
> >
> > This will work only with OpenCms 5.0rc1. You need to have the
> > "org.opencms.welcome" module installed (this is the default).
> >
> > Any feedback about this module is appreciated.
> >
> > Best Regards,
> > Alex.
> >
> > Alexander Kandzior
> > Alkacon Software
> > http://www.alkacon.com
> >
> >
> > > -----Original Message-----
> > > From: owner-opencms-dev at www.opencms.org
> > > [mailto:owner-opencms-dev at www.opencms.org] On Behalf Of
> > > agrandemange at amadeus.net
> > > Sent: Friday, January 10, 2003 2:14 PM
> > > To: opencms-dev at www.opencms.org
> > > Subject: RE: [opencms-dev] Search engine
> > >
> > >
> > >
> > > Alexander,
> > >
> > > yes please send me the htdig module. As I run Linux htdig
> > > should be fine.
> > >
> > > Best regards,
> > > Alexis
> > >
> > >
> > >
> > > From:  "Alexander Kandzior" <alex at opencms.org>@www.opencms.org  on
> > >        10/01/2003 11:21
> > >
> > > Please respond to opencms-dev at www.opencms.org
> > >
> > > Sent by:    owner-opencms-dev at www.opencms.org
> > >
> > >
> > >
> > >
> > >  To:    <opencms-dev at www.opencms.org>
> > >
> > >
> > >
> > >
> > >
> > >  cc:
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >  Subjec RE: [opencms-dev] Search engine
> > >
> > >  t:
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > > Alexis,
> > >
> > > I know of no plans to update the Lucene based search modules
> > > for OpenCms 5. We have a new JSP based module for htdig
> > > integration, but we will be publish this in a couple of weeks
> > > as we are still working on the documentation. If you want I
> > > can mail this directly to you now. You might also be able to
> > > integrate Lucene based on this.
> > >
> > > Best Regards,
> > > Alex.
> > >
> > > Alexander Kandzior
> > > OpenCms Group / Alkacon Software
> > >
> > > > -----Original Message-----
> > > > From: owner-opencms-dev at www.opencms.org
> > > > [mailto:owner-opencms-dev at www.opencms.org] On Behalf Of
> > > > agrandemange at amadeus.net
> > > > Sent: Thursday, January 09, 2003 6:32 PM
> > > > To: opencms-dev at opencms.com
> > > > Subject: [opencms-dev] Search engine
> > > >
> > > >
> > > > Hi,
> > > >
> > > > we use OpenCms 5 RC 1 and we need to add a search facility
> > > (preferably
> > > > Lucene-based.)
> > > > Which one should we use? The one at
> > > > http://www.opencms.com/opencms/opencms/communi>
> > > ty/lucene.html or the
> > > > one at http://www.opencms.com/opencms/opencms/service/modules.html?
> > > > If none of them is known to work with OpenCms 5, is it a plan
> > > > for a new version of this module? If needed I can contribute
> > > > but I would like to have the best starting point.
> > > >
> > > > Regards,
> > > > Alexis
> > > >
> > > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
> >
> >
> >
> >
> 
> 
> --
> Thomas Weckert
> 
> Alkacon Software GbR
> Im Meisengrund 4a
> 50996 Koeln, DE
> 
> Tel: +49 (0)2236 963491
> Fax: +49 (0)2236 963492
> Email: thomas at alkacon.com
> 
> http://www.alkacon.com
> 
> 
> 
> 
> 
> 
> 


-- 
Thomas Weckert

Alkacon Software GbR 
Im Meisengrund 4a 
50996 Koeln, DE 

Tel: +49 (0)2236 963491 
Fax: +49 (0)2236 963492 
Email: thomas at alkacon.com 

http://www.alkacon.com 




More information about the opencms-dev mailing list