Skip to content
This is the NewsFinder software, designed to automatically crawl the web for news related to artificial intelligence, filter, categorize, and rank the news, and publish to a wiki, mailing list, and RSS feeds.
Pull request Compare This branch is 32 commits behind AAAI:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Full details of this software can be found at the AITopics wiki.

The main script,, performs the training/crawling/publishing process. The script can be called with one of three arguments:

python crawl

This crawls the sources list (database table sources) and stores its results back into the database (table urllist).

python publish

This grabs the articles that have been crawled but not yet processed, filters stories based on relevance, and publishes the resutls as wiki pages, email, and RSS feeds.


NewsFinder is primarily coded in Python and requires the following libraries:

"Installation" of AINews should only involve downloading the code in this repository. Assuming Python includes the AINews code in its path, execution of AINews should be as simple as running python

There are various supplementary files that AINews expects; these involve publishing the news into PmWiki format and generating an email-formatted output. These files can be obtained from another repository: AINewsSupplementary

In our working configuration, we have a directory in our home called NewsFinder. In this directory are the AINewsSupplementary directories (e.g. templates, resource, etc.). The AINews code found in this repository is stored in a code directory in the main NewsFinder folder. In any event, the paths.ini file (in the config directory of this project) can be modified to point to whereever you keep your files.


Look at tables.sql to create tables needed by NewsFinder.

News sources

Sources are specified in the sources table of the database. A source has the following properties: URL, parser identifier, description, status, and relevance. The URL points to either an RSS feed or a search results page that is to be parsed. The parser identifier is composed of two parts separated by :: -- the first part is the publisher name, the second part is the type (e.g. rss); the parser identifer (both parts) is used in to determine how the URL will be processed. If you wish to crawl a news source not yet represented in, you may have to write a new parser (see the file for examples). The description of the source is just helpful text to disambiguate among different search terms used on the same publisher (e.g. all the different Google News searches). The description is not used by the AINews code. Finally, status is a boolean value (1 or 0) indicating whether this source should be crawled, and relevance is a ranking (higher = better) of how relevant or credible is the source. Stories from more relevant sources are more likely to be published.


All configuration is provided in the config/*.ini files:

  • config.ini has miscellaneous configuration options, described in the file
  • db.ini has database connectivity information (host, username, password, etc.)
  • paths.ini stores the paths of various data and output directories used by AINews. This file allows you to store data and output in whatever way makes sense for a particular installation (for example, data files should be outside of a webserver root)

Make a copy of the .ini.sample files (rename to .ini) to build your own configuration.


python train

The trainer collects user ratings and finds the best support vector machines to categorize the articles and select which are relevant to AINews readers. Output is saved to the svm_data path (defined in config/paths.ini).


The most recent version (as of this writing) of NewsFinder was written by Joshua Eckroth. The prior iteration was written by Liang Dong. Before that, NewsFinder was coded by Tom Charytoniuk. The project has been supervised by Bruce Buchanan and Reid Smith.


The NewsFinder software is been documented in the following publications:

L. Dong, R. G. Smith and B.G. Buchanan. NewsFinder: Automating an Artificial Intelligence News Service. Twenty-Third IAAI Conference on Innovative Applications of Artificial Intelligence (IAAI11), July, 2011.

L. Dong, R. G. Smith and B.G. Buchanan. Automating the Selection of Stories for AI in the News. Twenty-fourth International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems (IEA/AIE 2011), June, 2011.


Copyright (c) 2011 by the Association for the Advancement of Artificial Intelligence. This program and parts of it may be used and distributed without charge for non-commercial purposes as long as this notice is included.

Something went wrong with that request. Please try again.