rcm / PlagScan

This tool can be used to detect plagiarisms in software based submissions made to e.g., university courses or contests.

Software Plagiarism Detection Based on Code Property Graphs

Usage

1. Make sure you have Docker installed. This repository does not need to be cloned.

2. Pull the prebuilt PlagScan image: docker pull stefankoerner1/plagscan:latest. Note that this will download about 1.6 GB.

3. cd into the folder with the submissions that should be tested. Every submission must be in a separate subfolder. The structure should look something like this:

   .
   └── submissions
       ├── 08891236_F_A082
       │   └── f.c
       ├── 08534792_F_A026
       │   └── f.c
       ├── 08845662_F_A167
       │   └── f.c
       ├── 02343900_F_A255
       │   └── f.c
       ├── 08234263_F_A056
       │   └── f.c
       ├── 08666956_F_A157
       │   └── f.c
       ├── 08235553_F_A030
       │   └── f.c
       └── 08962935_F_A150
           └── f.c
   

4. Inside this folder run:

   docker run --rm -it -v $PWD:/app/submissions/ stefankoerner1/plagscan:latest

   (Note: $PWD might not work on every OS. See here and here for possible solutions.)

   with the following arguments:

   usage: [-h] -l LANGUAGE [-t THREADS] [-o OUTPUT] [-v VERBOSE]
   
   arguments:
       -h, --help            show this help message and exit
       -l LANGUAGE, --Language LANGUAGE
                               Programming language that is being used. Options: c, python
       -t THREADS, --Threads THREADS
                               Number of threads used (default: 4).
       -o OUTPUT, --Output OUTPUT
                               CSV that contains all suspicious combinations (default: report.csv).
       -v VERBOSE, --Verbose VERBOSE
                               Print [all] results or only [suspicious] (default: suspicious).
   

   Example usage for the language "C" that will use 8 threads:

   docker run --rm -it -v $PWD:/app/submissions/ stefankoerner1/plagscan:latest --Language c --Threads 8

Build locally

To build the image locally run:

docker build --pull --rm -f "Dockerfile" -t plagscan:latest "."

Development

For development, it is recommended to use Visual Studio Code Remote - Containers with the .devcontainer/devcontainer.json file from this repository.