Heh. I've written a similar python program, for the purpose of "scraping" information from websites, specifically, myspace, friendster, gmail, aol, yahoo, & hotmail. Urlgrabber and BeautifulSoup are excellent libraries for this purpose. I also made extensive use of the 're' module. Unfortunately, that work was paid for so I can't publish it under an opensource license.
I also wrote a scraper for ebay, similar to
bidwatcher. That *is* opensource but unfinished/unmaintained and thus unpublished.
One major difficulty you'll probably encounter maintaining such programs is that they tend to be pretty fragile (even if you use BeautifulSoup) and break easily if the site(s) in question make even somewhat minor changes.
Nice program you have there anyway. Not really into that sort of thing, though. Might I suggest, however, for organizational purposes you could put all your projects under a single domain instead of registering one for each?
Also, if you are concerned about testing, you should write a regression suite. I wrote one for the abovementioned contracted project, and it was quite easy. Though for the GUI side of things, not so much.