Convivea

Bit Che => Scripts & Development => Topic started by: CuF on January 19, 2013, 04:03:18 pm

Title: How much filtering should Bit Che have to do of results?
Post by: CuF on January 19, 2013, 04:03:18 pm

Some results tend to have junk between results that gives results that have to be filtered later, some websites have no ability to filter dead torrents from results, etc.

Should script writers shrug this off or is it preferred that if... and loop be used to test and remove known issues with a site.

Zero seeder results CAN be filtered by the script, but I suppose some people may want those results so they can find them in the 'removed' window.

Just thinking out loud mostly.  Opinions?

Title: Re: How much filtering should Bit Che have to do of results?
Post by: Bovski on January 19, 2013, 10:52:14 pm

What sort of junk ?

Don't filter out zero seeder results Bit Che handles this there are options in Preferences so the end user has a choice though if the website has the option to do this its always better to select it as you will get more results with seeders , doing it in a script will be a lot slower than letting bit che take care of it.

Loops should be avoided if possible a change to the site could result in a infinite loop.

Title: Re: How much filtering should Bit Che have to do of results?
Post by: chip! on January 22, 2013, 01:00:43 am

Couple thoughts:

The less functions in the start/aloop section the better, but sometimes another function or find/extract is needed to ensure withstanding minor site changes.

I like to aim towards as much pre-processing (prior to 'start') that can be done to manipulate the data so that once its split, there are less (or none) false results in the array.

using IF and a 'loop' will not create an infinite loop since all its doing is increasing the result array index and going back to 'start'. So it's OK to use that for detecting an anomaly and making the script work.

Let the Bit Che handle any removing of results related to dead/seeders, unless you have really good reason to do so from the script.

Above all, if it works and its not visibly slow, and you have tried to optimize it, then it works :)

Check out the script for TPB where it can use .torrent if available, http or HTTPS depending on what the user has selected for the mirror or also fall back to the magnet links. Its not pretty, but it works.

Oh lastly, problems with login sites might be with the failed failed2 where the first should be if you are not logged in and the second is if you are logged in but need to relogin. I usually request the final search page from another browser that wasn't logged in to see that scenerio. Could also be something else like the site is checking a referrer header tag which might need some tweaking.