aashishyadavally / ns-slicer

Artifact for "A Learning-Based Approach to Static Program Slicing": Paper accepted at OOPSLA'24

Artifact for "A Learning-Based Approach to Static Program Slicing"

NS-Slicer is a learning-based static program slicing tool, which extends such an analysis to partial Java programs. The source code, data, and model artifacts are publicly available on GitHub (https://github.com/aashishyadavally/ns-slicer), and Zenodo (https://zenodo.org/records/10463878).

Table of Contents

• Getting Started
  • Setup
    • Hardware Requirements
    • Project Environment
  • Directory Structure
  • Usage Guide
• Contributing Guidelines
• License

Getting Started

This section describes the preqrequisites, and contains instructions, to get the project up and running.

Setup

Hardware Requirements

NS-Slicer requires a GPU to run fast and produce the results. On machines without a GPU, note that it can be notoriously slow.

Project Environment

Currently, NS-Slicer works well on Ubuntu OS, and can be set up easily with all the prerequisite packages by following these instructions (if conda is already installed, update to the latest version with conda update conda, and skip steps 1 - 3):

1. Download the latest, appropriate version of conda for your machine (tested with conda 23.11.0).

2. Install it by running the conda_install.sh file, with the command:

   $ bash conda_install.sh

3. Add conda to bash profile:

   $ source ~/.bashrc

4. Navigate to ns-slicer (top-level directory) and create a conda virtual environment with the included environment.yml file using the following command:

   $ conda env create -f environment.yml

   To test successful installation, make sure autoslicer appears in the list of conda environments returned with conda env list.

5. Activate the virtual environment with the following command:

   $ conda activate autoslicer

Directory Structure

1. Data Artifacts

Navigate to ns-slicer/data/ to find:

• the dataset files ({train|val|test}-examples.json) -- use these files to benchmark learning-based static slicing approaches, or replicate intrinsic evaluation results in the paper (Sections 6.1 - 6.3).
• aliasing dataset files (aliasing-{examples|dataloader}.pkl) -- use these files to replicate variable aliasing experiment in the paper (Section 6.4).
• vulnerability detection dataset file (filtered-methods.json) -- use this file to replicate extrinsic evaluation experiment in the paper (Section 6.5).

2. Model Artifacts

Navigate to ns-slicer/models/ to find the trained model weights with CodeBERT and GraphCodeBERT pre-trained language models -- use these files to replicate results from the paper, or to produce static program slices for custom Java programs.

3. Code

Navigate to ns-slicer/src/ to find the source code for running experiments/using NS-Slicer to predict backward and forward static slices for a Java program.

4. Preliminary Study

Navigate to ns-slicer/empirical-study/ to find the details from the preliminary empirical study (see Section 3) in the paper.

Usage Guide

See link for details about replicating results in the paper, as well as using NS-Slicer to predict static program slices for Java programs. Here's an executive summary of the same:

Experiment	Table # in Paper	Data Artifact(s)	Run Command(s)	Model Artifact(s) for Direct Inference
(RQ1) Intrinsic Evaluation on Complete Code	1	data/{train|val|test}-examples.json	click here	CodeBERT, rows 7-9
				GraphCodeBERT, rows 10-12
(RQ2) Intrinsic Evaluation on Partial Code	2	data/{train|val|test}-examples.json	click here	GraphCodeBERT
(RQ3) Ablation Study	3	data/{train|val|test}-examples.json	click here	-
(RQ4) Variable Aliasing	4	data/aliasing-{examples|dataloader}.pkl	click here	CodeBERT, rows 1-2
				GraphCodeBERT, rows 3-4
(RQ5) Extrinsic Evaluation	5	data/filtered-methods.json	click here	GraphCodeBERT, row 2

Contributing Guidelines

There are no specific guidelines for contributing, apart from a few general guidelines we tried to follow, such as:

• Code should follow PEP8 standards as closely as possible
• Code should carry appropriate comments, wherever necessary, and follow the docstring convention in the repository.

If you see something that could be improved, send a pull request! We are always happy to look at improvements, to ensure that ns-slicer, as a project, is the best version of itself.

If you think something should be done differently (or is just-plain-broken), please create an issue.

License

See the LICENSE file for more details.