Simple crawler (just script)

I need a simple crawler that will crawl about 30 torrent sites and store in an sql database: - Title (of the downloadable file) - date crawled - Short description of file (if possible) - url This database needs to be refreshed every 3-5 days. And links crawled longer than 3-5 days ago should be removed. The crawler should be implemented as a webscript so it can run on a webserver with cron. I need just the backend script, I will write the front end myself. I need a web based crawler that can run efficiently, without overloading my server for instance with sql queries (simple greengeeks account)! This is part of a larger project, so I have a quite limited budget for this.