This folder contains the files to reproduce the results reported in our paper.

Each of the files contain certain TODOs that are required to be completed before executing them
These TODOs consist of assigning proper path to data or path of trained checkpoint etc.
Dataset is also provided in the data supplementary section of the submission.
The splits will be automatically created. To have the same splits that we used please keep the random seed value at 42.

File guide:
Proverb prediction: proverb_prediction.py
Joint proverb and span prediction: joint_proverb_span_prediction
Narrative generator models : bart_generator_model.py and t5_generator_model.py

The remaining files were used for dataset analysis.