clone_transcompiler

Code for paper, "Pathways to Leverage Transcompiler based Data Augmentation for Cross-Language Clone Detection", ICPC 2023

Installation

To set up the project, follow these steps:

1. Install dependencies by executing the following scripts:
   • install_server.sh
   • srcml_dep.sh
   • requirements.txt

Model Training and Testing

To train or test the models, it is recommended to use a virtual environment. Follow the specific requirements outlined in the requirements.txt file. For additional model-specific instructions, refer to the repository of the target model.

Setting up ANTLR and Transcoder

Detailed instructions for setting up ANTLR and Transcoder can be found in the following files:

• setup_antlr.txt
• setup_transcoder.txt

Clone Pairs and Dataset Generation

To generate clone pairs and datasets, follow these steps:

• Run the provided notebooks sequentially, ensuring dependencies are met
• Feature extraction using ANTLR (find in CLCDSA repo)
• Utilize the clone pairs generation method provided in the CLCDSA repo (requires Java)

For more information, refer to the respective repositories and documentation.

Required Models and Other Repositories

• Pre-trained Model

  • Transcoder: Unsupervised Translation of Programming Languages

• Cross-Language Clone Detection Models

  • CLCDSA: Cross-Language Code Clone Detection using Syntactical Features and API Documentation
  • C4: Contrastive Cross-Language Code Clone Detection

• Graph Matching Network for Single-Language Clones

  • GMN: Detecting Code Clones with Graph Neural Network and Flow-Augmented Abstract Syntax Tree

• Parsers

  • Javalang
  • srcML

Contact

Subroto Nag Pinku, subroto.npi@usask.ca