A Python framework for generating quizzes based off of lists in Wikipedia articles. Uses scikit-learn to categorize answers into different categories so that they can be presented in a multiple-choice list.

Installing:

• Clone this repo
• Install BeautifulSoup4, sklearn, nltk via pip3. (There are a few more dependencies that I haven't listed here - you'll have to install them as you attempt to run the scripts, sorry!)
• Use the tools below

Tools:

train.py

Used to train and generate the model used by crawler.py. Makes use of the data in data to generate a model. Data in data is classified according to the directory it is in.

Provides details about the model test results when run and saves the new model into trained_model*.pickle files.

Usage: python3 train.py

crawler.py

Targets and crawl a section of a Wikipedia page, turning the bullet points in that section into quiz questions and answers. It targets from one <h2> element on the page to the next, as that is a relatively common way to break Wikipedia pages into separate sections.

Usage: python3 crawler.py <page name> <h2 ID of section> Example: python3 crawler.py "2004 in music" "Popular_songs"

quiz.py

Provides a way to test the question and answer functionality in the terminal. A specific question set can be provided for one-off testing, otherwise all quiz data available in quiz_content is used to provide 5 questions to the user.

Usage: python3 quiz.py Example: python3 quiz.py "2004 in music" "Popular_songs"

websocket_server.py

Used to serve and keep track of game state for the wikipedia-quiz-client.

Usage: python3 websocket_server.py