IST456 / tmoss

Code for the TMOSS software.

Temporal Measure of Software Similarity (TMOSS)

This repository contains a version of TMOSS, which is a method for analyzing intermediate student code. While this tool has been tested with MOSS as a backend software similarity detection software, any backend should be compatible.

This version of TMOSS has been tested to work on an Ubuntu machine.

Setup

1. Connecting the backend: After you clone your repository, rename your moss backend directory to the bin directory. For example:

   code/
     bin/
       LICENSE
       moss        # the moss binary file
   

   You must download the MOSS binary from the MOSS website. We have provided a useful HTML interface (pymoss) for running and parsing the MOSS results, but this does not include the binary.

2. Linking your data:

   Your student data should be located in the data directory. Each course offering should be in its own sub-directory under data. Online submissions are expected to be in their own online directory. If you do not have online or starter code directories, the script will still run.

   data/
       fall_2012/
         student_1/
           .git/
         student_2/
           .git/
         ...
       fall_2013/
         ...
       fall_2014/
         ...
   online/
       online_solution1/
           solutionFile.ext
       ...
   starter/
       starter_code1.ext
       ...
   

3. Install dependencies:

   sudo apt-get install python-matplotlib git
   

Running the program

TMOSS is split into two steps:

1. Compute the top similarity score of all snapshots in all student repositories.

   The results of running TMOSS will be located in the out directory. The directory structure will be similar to data.

2. Calculate the top match of each student.

   The top match results are located in out/<course_dir>/top_matches.csv and follow the following format:

   student_name,matched_student,snapshot_time_and_name,similarity_score
   

To run TMOSS:

python code/run.py

If you also want to plot a gumbel fit to your data afterwards, run:

python code/gumbel.py

The Gumbel script assumes your HEC and non-HEC results adhere to the following guidelines:

• Follow the format of top_matches.csv, the unparsed results from running TMOSS.

• Have been looked through by a human to separate into the HEC and non-HEC groups.

• Are located in the out directory as follows:

  out/
    fall_2012/
      hec.csv
      non_hec.csv
      top_matches.csv       # the unparsed results from tmoss
    fall_2013/
      ...
    fall_2014/
      ...
  

Citing this work

If you use TMOSS in your research, we ask that you please cite the following:

L. Yan, N. McKeown, M. Sahami, and C. Piech, “TMOSS: Using Intermediate Assignment Work to Understand Excessive Collaboration in Large Classes,” in Proceedings of The 49th ACM Technical Symposium on Computer Science Education (SIGCSE), February 2018.

ACM library link

Our paper is also available at this link.