The main haiku generation code is in experiment.py
Content generation using Word Association Networks (WAN) is in wan.py
Content generation using WordNet is in word_net_util.py
Scraping/haiku collection code is in crawl.py
Other files contain utility functions used in the main modules or some experiments we did (like using bigrams for everything) before we found the optimal method to generate haiku.
To resolve dependencies, please run:
sudo pip install -r requirements.txt
If you get an issue with a python import related to nltk, please run sudo python -m nltk.downloader -d /usr/share/nltk_data all