The NuzeBot

The NuzeBot is the robot that finds the news you want to see.

Download NuzeBot 4.1
(20kB zip file)

v4.1 is only about 50% faster in generating news from headlines!

You like reading some news articles, but you always have to wade through a lot of stories you don't like and visit a bunch of different news websites that have slow, seemingly poorly designed front pages, all of which is just annoying and might even deter you from reading much news.

You need an assistant to go and surf the web to find the headlines that interest you. That's the goal of the NuzeBot. I think it does a fairly good job, though it helps if you're willing to do frequent tweaking of the configuration files to fine tune the output to match your needs.

See some example output on the news page on this website.

Version 4 of the NuzeBot is the first release in a year and a half and features many changes:

Demo

The demo shows how the bot can produce useful results. The demo runs the bot for science news and unexplained news, generating science.htm and unexplained.htm pages. The unexplained news piggybacks on the science news data-wise so the bot doesn't need to repeat network activity to look at the same sites for news again. The science news config tells the bot to save headlines to a news.dat file, and the unexplained news config tells the bot to load headlines from that file instead of loading them from the internet.

Just run the menu.sh file.

How It Works

Headlines are loaded, scored according to your rules, and sorted by score. Old headlines are penalized. Headlines that are too similar to better-scoring headlines are eliminated, the worst scoring headlines are eliminated, and the results are output to a page. For more details, see the source code.

V4.1

I've been playing with profiling with GCC, something I hadn't done in many years. It gave me an idea that I could speed it up by using local versions of a couple of simple functions. In fact, this speeds up the bot considerably when generating news from stored or already gathered headlines, especially important when running for multiple news pages.

Changelog

v4 - First release in 1.5 years. Many changes. See above.
v4.1 - About 50% faster generating news from headlines that are stored or already gathered.

Future Development

I'm considering moving back to a modular design as I was using about ten years ago, before I had released the bot as open source software. This would allow the operations of the bot to be divided among a number of smaller, simpler programs, instead of requiring a single big monolithic monster. It would also allow easier testing of individual parts of the bot.

See Also

News Page - Some actual outputs of the NuzeBot, currently updated every four hours.

IRC Bot - A bot to help with managing your chat room on IRC.




Page generated in 0.149206 seconds.

© 2024 RockyMount.US