This is my contribution to https://github.com/dariusk/NaNoGenMo
Thus far, it consists of:
- nell2pl.sh - a ZSH script that converts the ontology format used by CMU NELL ( http://rtw.ml.cmu.edu/rtw/ ) into valid prolog fact sets
- questions.pl - a set of prolog rules that generate question and answer pairs about subjects covered in the ontology
- fav.sh - a script that generates a list of question and answer pairs about favorite things found in the ontology
The dialog is of the formula:
What is your favorite $X?
$Y, because it is also $Z.
For instance:
What is your favorite actor?
Tom cruise, because it is also scientologist.
Sample output is here: https://github.com/enkiv2/NaNoGenMo2013/blob/master/example_output/favorite_generated_clean.txt