Shedding Light on Software Engineering-specific Metaphors and Idioms

This repository contains the data and code to reproduce the experiments from our paper titled "Shedding Light on Software Engineering-specific Metaphors and Idioms".

Repository Structure

• /annotation: A folder contains resources related to dataset annotation and the annotated CSV file.

• /contrastive_learning: Implementation of contrastive learning method. Before run RQ2 and RQ3, you must run the script from this folder and save the model weights which can be loaded later.

  • contrastive_learning.py: the script of contrastive learning.

  • Fig_Lan_Annotation.csv: contains the dataset of contrastive learning.

  • ReadMe.md: contains the readMe on how to run contrastive_learning.py.

• /RQ1: Folder that contains experiments for RQ1 on LLM figurative language interpretation using cosine similarity.

  • run.py: the script of cosine similarity and other metrics.

  • annotated-dataset.csv: the dataset of this experiment.

  • ReadMe.md: contains the readMe on how to run run.py.

• /RQ2: Folder that contains experiments for RQ2 on improving affect analysis via fine-tuning.

  • /github_emotion: Folder that contains experiments for emotion classifcation.

    • emotion_classification.py: the script of emotion classification.

  • github-train.csv: the train dataset of this experiment.

  • github-test.csv: the test dataset of this experiment.

  • ReadMe.md: contains the readMe on how to run emotion_classification.py.

  • /github_incivility: Folder that contains experiments for incivility classifcation.

    • incivility_classification.py: the script of incivility classification.

  • incivility-train.csv: the train dataset of this experiment.

  • incivility-test.csv: the test dataset of this experiment.

  • ReadMe.md: contains the readMe on how to run incivility_classification.py.

• /RQ3: Folder that contains experiments for RQ3 on bug report prioritization.

  • bug_priority_classification.py: the script of bug priority classification.

  • priority-small-train.csv: the train dataset of this experiment.

  • priority-small-test.csv: the test dataset of this experiment.

  • ReadMe.md: contains the readMe on how to run bug_priority_classification.py.

• requirements.txt: Python package dependencies.

Each folder contains separate ReadME showing how to run each separate scripts.