Welcome! Log In Create A New Profile

Advanced

Indexing Bugs

Posted by simondoyle 
Indexing Bugs
April 11, 2007 06:15PM
I have set up Sphider on a new site I'm working on, and it seems great. But there are a couple of bugs which I will have to squash. Just wondering if anyone else has found solutions for them before I start...

1. An author on the site has the surname "Schlepper-Connolly", but Sphider is indexing this as "Schlepper-Connolli". When I click through, the correct spelling is shown but not highlighted... I have tried to fix this in the DB, but then the search reference stops working. Weird...
2. As has been mentioned by other users, umlauts and acutes are being indexed correctly, but aren't searching or suggesting correctly. Any solution for this yet?
3. Finally, <br />s don't seem to get recognised, and so words separated by a <br /> get read together as a keyword.



Edited 2 time(s). Last edit at 04/11/2007 06:30PM by simondoyle.
Re: Indexing Bugs
April 11, 2007 11:07PM
Okay - I have now noticed that every name terminating in Y is indexed as terminating in I... Weirdsville. Can't guess what the problem is...

There is a new fix for the special characters issue, but unfortunately this doesn't help me, as I am DB'ing them in htmlencoded fashion. So I will need to encode and unencode at relevant places... Not a big deal.

Third issue should be easy to fix - just add <br> as a linebreak designation to the indexing script.

I will find fixes to all these problems over the next couple of days, and share the solutions with you all. (Unless anyone cares to share a solution in advance!)
Anonymous User
Re: Indexing Bugs
April 12, 2007 06:22PM
In this site you can search "query"

http://www.sphider.eu/forum/search.php?0,search=query,page=1,match_type=ALL,match_dates=30,match_forum=ALL

but you don't find queri
http://www.sphider.eu/forum/search.php?2,search=queri,page=1,match_type=ALL,match_dates=30,match_forum=ALL

you find "Schlepper-Connolly"
http://www.sphider.eu/forum/search.php?2,search=Schlepper-Connolly,page=1,match_type=ALL,match_dates=30,match_forum=ALL

and also "Schlepper-Connolli"
http://www.sphider.eu/forum/search.php?2,search=Schlepper-Connolli,page=1,match_type=ALL,match_dates=30,match_forum=ALL

So, in two similar situations (words terminated in "y"winking smiley, you have different results

It just enlarges the problem.
Anonymous User
Re: Indexing Bugs
April 12, 2007 06:26PM
Sorry. It finds "query"
Anonymous User
Re: Indexing Bugs
April 12, 2007 07:01PM
There is someone who found the solution (it is a site a site powered by Sphider)
http://pauloquerido.net/dre-pesquisa/

He can't find "Terry", because it doen't exist

http://pauloquerido.net/dre-pesquisa/?query=terri&type=and&results=20&search=1


But he can find "Terry"
http://pauloquerido.net/dre-pesquisa/?query=terry&type=and&results=20&search=1
Anonymous User
Re: Indexing Bugs
April 12, 2007 07:11PM
Other searches where you have no problems with Y (you find the word "Terry", but you do not find the word terri, that doesn' exist

http://memoriavirtual.net/dre-pesquisa/?query=terri&type=and&results=20&search=1

http://memoriavirtual.net/dre-pesquisa/?query=terri&type=and&results=20&search=1


My personal site has the "y" problem. I shall wait for your solution.
Re: Indexing Bugs
April 16, 2007 08:20PM
The Y problem is the result of the word-stem option being turned on. Some words get incorrectly turned into stems - among them words ending in Y which get turned into words ending in I.
Anonymous User
Re: Indexing Bugs
April 16, 2007 09:53PM
simondoyle.

How do I turn on/off the word-stem option with sphider?

Thanks
Re: Indexing Bugs
April 17, 2007 05:54PM
Go into the admin page, then click on "Settings", inside "Spider settings", you will see
"Use word stemming"
remove the check mark and it will be off.

Diego Medina
[url=http://www.fmpwizard.com]Web Developer[/url]
Anonymous User
Re: Indexing Bugs
April 17, 2007 08:27PM
Diego Medina

Thank you. It works.
Sorry, only registered users may post in this forum.

Click here to login