NLP & LSS 2022 project aimed to create conditional clickbait titles based on the article content.
Check the Huggingface Spaces app at : https://huggingface.co/spaces/Dusan/clickbaitonator

ClickbaitClassifier.ipynb - notebook for training clickbait classifier.
Fudge_running.ipynb - Fudge based implementation for generating controllable clickbait titles. It contains setup for generating titles, rating them according to our clickbait classifier and saving in respective dataframes. Our Huggingface Spaces implementation uses this model due to realtime response. At the end of this notebook there is also a Gradio section where you can launch in-notebook Gradio application.
PPLM_running.ipynb - PPLM based implementation for generating controllable clickbait titles. It contains setup for generating titles, rating them according to our clickbait classifier and saving in respective dataframes.
DataSet_sentence_collector.ipynb - sentence collector on Fudge and PPLM dataframes generated by Fudge_running.ipynb and PPLM_running.ipynb. It was used for sampling best titles that were generated.
DatasetAnalysis.ipynb - notebook for running wordcloud and LDA topic modelling analysis on ground truth titles.

DataProcessing - used for generating all relevant tables in a report and presentation. Contains fudge_csv, PPLM_csv and pegasus_clickbait_summarizer_csv folders for .csv files that were used for the report and visualizations of respective models.
fudge - Fudge adapted to our clickbait summarizer.
PPLM - PPLM adapted to our clickbait summarizer.