Wikipedia Game Mapper

The Wikipedia Game Mapper is a tool that crawls through a random wikipedia page (and all pages that it links to) until it can find the goal page that is provided as input. The mapper performs a breadth-first search, only going n levels deep, serving as a brute force approach to solving The Wiki Game.

This started as a project that I planned to finish in a couple of hours. Needles to say, it ended up taking me a lot longer to finish simply because I kept seeing potential to make it better and better.

Visualization

This generated 56,000 paths from the random article page to the Adolf Hitler page with over 82,000 UNIQUE nodes (which is an interesting observation in itself). The javascript branch of this project contains the code used to visualize this set of data in the browser, and is adapted from allnpmviz3d with a couple of modifications, namely color.

In this visualization, the color represents the click level, so the red sphere is the final page (Adolf Hitler), the yellow color is the third page, the green is the second page, and the blue is the random page.

This interactive visualization can be found at https://lharri73.github.io/wikipediaProject/.

Building From Source

This project was created using the C++98 standard, but with the addition of multithreading, the libraries added in C++11 made this significantly easier.

NOTE: This project requires the use of Google's Gumbo parser.

Install dependencies for the gumbo parser

    sudo apt install libtool automake

Install mysql header files

    sudo apt install libmysqlcppconn-dev

Clone and set up this repository. The gumbo parser is added to this repository as a submodule.

    git clone https://github.com/lharri73/wikipediaProject.git
    cd wikipediaProject
    git submodule init
    git submodule update

Install the Gumbo parser. Detailed instructions can be found on the Gumbo repository, but I'll leave the spark notes version here

    cd lib/gumbo
    ./autogen.sh
    ./configure
    make
    sudo make install

Build the wikipedia project.

    cd ../../ ## back to the wikipediaProject directory (or wherever you cloned the repo to)
    make

Usage

    ./bin/wikipedia "goal_page_name" max_search_depth

Flags

Flag	Description	Default
`--single`	Disable multithreading	False
`--debug`	Run a single iteration on random page	False
`--sql_ip #.#.#.#`	Set the ip of the MySQL server	127.0.0.1
`--sql_user XXX`	Set the user of the MySQL server	root
`--sql_pass XXX`	Set the password of the MySQL server	""
`--ram #.#`	Set the amount of ram to use (in GB)	8.0

NOTE: --ram Does not set a hard limit. The program will not create threads until memory usage is below 75% of this number.

Contributions

Feel free to do your own experimentation and create pull requests as necessary!

If you find some interesting results, or generate a significant number of entries that can be added to my table of "Adolf Hitler" paths, send me an email!

License

GPLv2

lharri73 / wikipediaProject