thank you for the feedback, felix, so, for a start, i will just check if there's a exact match for title/name fields. if not found, try a more broad regex LIKE '%$conditions%' for title/name , if not found, search in other fields
$conditions = $this->params['url']['s']; $results['exact']['filme'] = $this->Filme->findAll(array('title'=>$conditions)); if not found $results['fuzzy']['filme'] = $this->Filme->findAll("title LIKE '%".$conditions."%'"); if not found $results['full']['filme'] = $this->Filme->findAll("slug LIKE '%".$conditions."%' OR titulo_original LIKE '%".$conditions."%' OR titulo_ingles LIKE '%".$conditions."%' OR sinopse LIKE '%".$conditions."%' OR description LIKE '%".$conditions."%'"); and use that array on the view looks simple and effective, but i'll post when i can make something better thank you spark On 7/19/06, Felix Geisendörfer <[EMAIL PROTECTED]> wrote: > > Hi spark, > > unfortunately the code I've written was for searching local files and > doesn't make usage of several advantages you've > got when working with databases, so it would not be worth for you trying to > apply it to your own solution. Other then that > I've written some regex that will parse the queries, so "connected > keywords" will be recognized and stupid user mistakes > like doubling keywords will be ignored. But if you are good with regex > yourself that shouldn't take you that long. > > Regarding ranking, there is almost no limit on the complexity of the > algorithm you may use to determine one items > popularity. I think user feedback should be weighted pretty high, but only > if the input can be trusted. And to determine > trusted feedback from spam & manipulation can be pretty tricky. So this > isn't something I would completely rely on, when > implementing search in a site where advantages can be taken from having > certain items high up in the serps (search result > pages). > > A better solution is to put efforts into understanding the relationships of > the items you are trying to compare. Now, when searching > your own database, you are in the lucky situation that you can track things > like item popularity, creation / updated time stamps and > tons of other useful information, just like Amazon does with "Other people > who liked this product also bought ..." that you wouldn't be > able to track, when searching through foreign databases (or web sites). Of > course, this is sort of suffering from similar problems > like the idea of taking user feedback. > > So again, I think it would be really nice to have a generic search solution > for cakephp that's based on a central Search Model that > get's it's informations from special Crawler (Components?) written by the > community. This would be quite challenging, but as a result, > we could have a common data interface for the applications we write with, > that we could make available as a web service. I'm sure > there are tons of sweet things we could do with that. > > Alright, but before I start to sound like an employee of this big company > that tries to index the entire world, I'll stop here and leave it > to you to figure out what will be the best solution for the site you are > creating ; ). > > > Best Regards, > Felix Geisendörfer > > -------------------------- > http://www.thinkingphp.org > http://www.fg-webdesign.de > > DJ Spark schrieb: > quite elegant solution, gwoo ! thanx! > > i'm also working on a tiny search for a project in cakephp, Felix. > Anyone have some references on search enhancements? ( like that in > youtube, for example, where a string is sent and it says it comes from > a search result, a ranks it better on the next search for the same > term). > Any article related would be a good reading for me at this point ;) > > spark > > > > On 7/19/06, Felix Geisendörfer <[EMAIL PROTECTED]> wrote: > > > Well you could refactor the Search functionality into an own model, which > would make > sense to me. But you don't necessarily have to do that. You can put this in > any controller > to say that no Model is being used: > > var $uses = array(); > > But really, if you plan to implement a quality search into a high usage web > site, you should > seriously consider maintaining an own search index as well as to do some > regex on the > queries (so users can use "two words" and such). Once you get into this > topic you'll find out, > that doing a good job on search (even for your own db), is not all that > easy. I mean you can > literally spend 50 hours (and more ...) to write the code required for an > intelligent search tool. > I've worked on a file search tool for my personal use in CakePHP, and I was > really surprised > about the complexity that comes with it. If I find some time this sommer I > might try to create a > Generic Search Model for CakePHP, that will work together with different > types of crawlers you > can create by yourself (DB Tables, RSS feeds, Html, XML, Image Meta data, > ...). I think it's quite > an interesting topic and if anybody else would be interested in it it might > well be worth a little > CakeForge project. > > Best Regards, > Felix Geisendörfer > > -------------------------- > http://www.thinkingphp.org > http://www.fg-webdesign.de > > carlosrg schrieb: > But, we need a model for this Controller, it's ok? What should be the > model? > > > > > > > > > > > > > > > > > > -- [web] http://synapsisdi.com.br [livesets] http://djspark.com.br --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Cake PHP" group. To post to this group, send email to cake-php@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/cake-php -~----------~----~----~----~------~----~------~--~---