Paper: Automatically Selecting Follow-up Questions for Deficient Bug Reports

In the following, we briefly describe the different components that are included in this project and the softwares required to run the experiments.

Project Structure

The project includes the following files and folders:

• /data: A folder that contains inputs that are used for the experiments
• dataset.csv: CSV file that contains 25k bug reports with follow-up questions and answers
• github_issue_titles.csv: Titles of those 25k bug reports
• github_issue_labels.csv: Labels of those 25k bug reports
• github_repo_labels.csv: Repository labels of those 25k bug reports
• post_data.tsv: 25k bug reports processed by Lucene
• qa_data.tsv: 25k bug reports with 10 candidate follow-up questions selected by Lucene
• test_ids.txt: Test dataset ids
• train_ids.txt: Train dataset ids
• /embeddings: A folder that contains the embeddings we have used
• /script: Contains the scripts for running the experiments
  • /models: Contains the scripts of the model
    • run_main.sh: The entry point of the experiment
• /survey: A folder that contains surevey related files

Software Requirements

We used the following software to run our experiments

• python3.6
• torch
• torchvision
• spacy>=2.2.4
• numpy pandas gensim jsonschema
• conda

Setup

Conda

conda install numpy pandas gensim jsonschema
conda install pytorch torchvision cudatoolkit=10.1 -c pytorch
python -m spacy download en_core_web_sm

Conda currently does not provide required version of spacy, so we use pip to install spacy:

conda install pip
pip install spacy==2.2.4

Running Experiments

Step 1: Install software requirements mentioned above.

Step 2: Update the filepaths and parameters in script/models/run_main.sh

Step 3: ./script/models/run_main.sh