Dependency Smells in JavaScript Projects

This is the replication package for our paper on dependency smells.

Here is a short description of what is contained in this package:

Code

This folder contains the code used for extracting, parsing, and analyzing the smells in the dataset, along with statistical analyses. The "parser.py" parses the project information (such as package.json) and prepares them in the databases. The "analyzer.py" file is responsible for the majority of the empirical analyses.

Datasets

This folder contains the intermediate datasets created and used in our analyses. The "smelldataset.db" file contains all smelly and clean dependencies for the latest snapshot. The "smell_counts.csv" file contains smells statistics for the projects in our dataset. The "changehistory.db" file contains the historical smell statistics for a period of 5 years. The code also requires the GhTorrent Dataset available at: https://ghtorrent.org/downloads.html.

Survey Questionnaires and Responses

These two folders contain the full set of questions that we asked the developers in our surveys along with the responses. The responses have been anonymised to hide personal information.

Tool

This is the published tool which is also available at: https://github.com/abbasjavan/DependencySniffer

Visualization Scripts

This folder contains the scripts used to create the figures for the paper.