Transforming Proprietary to High-Level Trigger-Action Programming Rules Dataset

This dataset contains the results of performing Natural Language Processing (NLP) techniques on the  May~2017 dataset of Mi et al. (https://www-users.cse.umn.edu/~fengqian/ifttt_measurement/download/201705.tar.gz) using 3 different approaches.
It comprises folders for each of the approaches (allennlp, spacy and combined) as well as folders containing the raw dataset divided into triggers and actions. The allennlp and spacy datasets are also divided into (filtered) triggers and actions.
The allennlp folders comprise 1 JSON file which shows

The proprietary trigger or action name - ifttt_name
The EUPont translation - eupont_hypothesis
A value for which the ifttt_name and the eupont_hypothesis may be similar - allen_nlp_entailment
A value for which the ifttt_name and the eupont_hypothesis may be contradictory - allen_nlp_contradiction
A value for which the ifttt_name and the eupont_hypothesis may be neutral -  allen_nlp_neutral
The spacy folders comprise 1 JSON file for each trigger and action which shows

The EUPont translation
The proprietary trigger or action name
A value for which the ifttt_name and the eupont_hypothesis may be similar
The combined folder comprises folders for triggers and actions which comprise 1 JSON file for each trigger and action which shows

The proprietary trigger or action name - ifttt_name
The EUPont translation - eupont_hypothesis
A value for the spacy similarity - spacy_similarity
A value for which the ifttt_name and the eupont_hypothesis may be similar -  allen_nlp_entailment
A value for which the ifttt_name and the eupont_hypothesis may be contradictory -  allen_nlp_contradiction
A value for which the ifttt_name and the eupont_hypothesis may be neutral -  allen_nlp_neutral
A value for the combined similarity - combined_similarity