[pmwiki-users] Search

Kathryn Andersen kat_lists at katspace.homelinux.org
Fri Sep 21 03:28:49 CDT 2007


On Thu, Sep 20, 2007 at 06:38:48PM -0400, Sameer Kumar wrote:
>> What I've done is to set up an external search engine (I'm using
>> swish-e(1)) which indexes the whole server, and then alter the 
>> $SearchBoxFmt so that it creates a form which points to the external
>> search rather than the PmWiki search.
> 
> How and where are the results of the search displayed? In the browser?
> In the wiki? 

The idea is that one bypasses the Wiki search altogether, and uses an
external search engine, which one will have to set up and configure
separately.  That's the only way you can get searches on things that ARE
on the web server, but are NOT on the wiki.

If you want to access things that are not available through the web
server at all, you've got problems.

My setup is that
(a) I have a wiki under /wiki/ on the intranet website (that is, a
webserver serves the site, but only to machines on our local intranet;
it is behind a firewall)
(b) I have other documents under other locations on the intranet
website.

I have then set up swish-e as an external search engine and put the cgi
part of it under "/search/" on the intranet website.  The setup of
swish-e is a completely separate exersize, it has nothing to do with the
wiki.  But the wiki can be indexed by it.  I can't remember if I had to
make some configuration changes to the wiki (robots following stuff)
to enable the external search engine to index the wiki.

What happens is that the external search engine indexes both the wiki
and non-wiki parts of the site, because as far as it is concerned, the
whole thing is just one website.  Which is presumably what you want.
 
> Do they include the wiki page results?

Wiki pages would be treated as any old website pages.

> Would the user then need to
> 'download' the non-web format documents to view them?

If the documents can't be viewed in the browser, naturally they would
have to be downloaded in order to be viewed by something else that can
view them.

> Thanks for the tip. How exactly does one define the $SearchBoxFmt ?

In my case, I've set it to the following:

$SearchBoxFmt = '<form class="wikisearch" method="get"
action="/search/">
<input class="inputbox searchbox" size="40" maxlength="200" value=""
type="text" name="query"/>
<input class="inputbutton searchbutton" value="Search" type="submit"
name="submit"/><br>
</form>';

Note that the form's action is "/search/" -- that is where I put the
swish-e search cgi (as /search/index.cgi).  So now, when (:searchbox:)
is used in the wiki, it goes straight to the external search script, and
completely bypasses PmWiki's own search.
 
> Also, will swish-e be able to index documents on network drives other
> than the server itself? Since my wiki is on a public IP but the rest of
> the library is on the intranet, should I prepare myself for security
> related issues? You may have guessed, I am quite a newbie at this, with
> very little programming / networking experience.

It depends on how everything is set up.  You could well have security
issues, or you might not.  What is the purpose of this webserver?  Is it
supposed to be accessable to the public, or only the intranet?
And as for network drives, I think that would depend on how your network
is set up, and in any case, the webserver should only be able to serve
files from directories that it has been explicitly set up to serve files
from.

Kathryn Andersen
-- 
 _--_|\     | Kathryn Andersen	<http://www.katspace.com>
/      \    | 
\_.--.*/    | GenFicCrit mailing list <http://www.katspace.com/gen_fic_crit/>
      v     | 
------------| Melbourne -> Victoria -> Australia -> Southern Hemisphere
Maranatha!  |	-> Earth -> Sol -> Milky Way Galaxy -> Universe



More information about the pmwiki-users mailing list