seconding request for more info on customization

Post Reply
User avatar
Thunderstone
Site Admin
Posts: 2504
Joined: Wed Jun 07, 2000 6:20 pm

seconding request for more info on customization

Post by Thunderstone »



Well, I really have tried to RTFM by browsing the
Thunderstone pages and searching for info. I'm still
not sure I understand what I need to do to customize
my search page.

First, before I start propagating links to our Webinator
search form, I'm wondering if we can buffer ourselves against
future local changes by defining a virtual host address
that would map right to the search form. We are running
Apache on our server and we do this frequently for our
users who want to have a customized address, defining or mapping
in Apache where it points in the information tree on the server.
But we've always done that with ordinary HTML files, not
with a cgi file. So, if we defined the virtual host to be
http://www.webinator.unh.edu/
are you aware of any reasons why we could not (or should
not) map that to our current Webinator URL of
http://www.unh.edu/cgi-bin/texis/webinator/search

Even if I've misstated some step here, I hope you can
see what I'm intending to do.

Second, to see if I understand the "mechanics".
We have an uncustomized installation with licensed Webinator,
but not a license for Texis. When I use the URL
http://www.unh.edu/cgi-bin/texis/webinator/search
is that displaying the basic search form by running
the "search" cgi-script which knows to look in my
webinator home directory and run the standard file
"search.vtx"?

I see that I have, from the basic installation, both
search (ascii) and search.vtx (compiled, evidently).
If I want to customize my search page, is it the
search file that I edit? If I do that, what causes
it to be recompiled? Am I required to have my search
script named "search"? If I wanted it to be "foobar",
how would I then get it to run? Would I do that by
linking to it from within _any_ HTML file on my server
where Webinator is installed and Webinator/Vortex
would be smart enough to compile and run the script?
I feel tantalizingly close to understanding this, but
confused.

A different question.
Now that I've walked my server and can do retrievals,
I'm not sure how many total pages are in the database.
Is there an SQL command I can issue of the form
gw -s "put-SQL-statement-here"

And finally.
Most commercial search engines display the count of
the total number of matches before they start listing
the first 10 or whatever matches. Is there a demo
that shows the commands that would be needed to
customize the search script to do that? I assume
that once I understand the "mechanics" asked about in
my questions above, that this would be straight-forward
to do. Similarly in a Q/A on 1998-01-28, someone asked
about query logging and the answer showed a sample
<sql> statement to use and where to place it -- again
presumably a simple customization once the mechanics
are understood?

Jim Cerny, Computing & Information Services, Univ.NH



User avatar
Thunderstone
Site Admin
Posts: 2504
Joined: Wed Jun 07, 2000 6:20 pm

seconding request for more info on customization

Post by Thunderstone »




I think

gw -s "select count(*) from html"

should do the trick.

Adam



User avatar
Thunderstone
Site Admin
Posts: 2504
Joined: Wed Jun 07, 2000 6:20 pm

seconding request for more info on customization

Post by Thunderstone »




Edit the file ~htdocs/webinator/search .

Here's a Quick explanation:

Anytime you see /cgi-bin/texis/xyz/script

The actual location of the script is in

the "xyz" directory under your webserver's document root.
The source file in this case is called "script" , and
the Texis executable is called "script.vtx" .



You can do anything you wish, but the string "/texis" must
occur in the URL in order for the script to run.


Webinator is nothing more than an application program
written in Texis. If you have a Commercial Webinator,
you have a limited use license to use Texis within the
context of the Webinator only.

See: http://www.thunderstone.com/webinator/commlic.html

As to the path discussion, see the section of the Texis Webscript
manual which describes "URL Syntax".
(http://www.thunderstone.com/vortexman/node15.html)



Texis will auto re-compile the source code if you change
it. We suggest you make a copy of the ascii search script
into another file eg: mysearch. Then test your edits by
looking at the URL http://myserver/cgi-bin/texis/webinator/mysearch .




The SQL for this is: "select count(Url) from html"



The Webscript variable $indexcount contains the
the number of items in a text index that match the
query. The variable's contents are valid inside
the first iteration of a <SQL> loop containing a text
search.

With respect to a query log. We'll defer to your creativity
and newly discovered Vortex abilities. Anything you
would probably ever want to do is available by writing
a few lines of Vortex code.




Post Reply