A simple cheat detector script which can provide some insight into cases for direct copying of assignments.
This program will compare all given singular input files in the specified directory with one another and compare them.
Based on a given closeness factor, the program will output an HTML diff of files which appear to
be similar.
Finally, output is provided which can be loaded into Excel to see the similarities between all inputs. The lowest, average, and highest comparison values will be provided as well.
To run, use Python 2:
python cd.py config.txt
Note: This can take a long time to process. It's best to leave it running unattended for a while.
STARTPATH = r'example'
The starting directory to look for input files. Will recursively search subdirectories. You can
specify a ..\ at the start to navigate up parents folders first.
SKIPPATHS = [] # actually subtrees
Paths to skip when searching.
EXTS = ['java']
List of extensions to take as input.
RESULTFILE = r'result.tsv'
This file is a tab separated file which can be loaded in Excel. The diagonal will be marked NA as
entries aren't compared to themselves. You can highlight the inner matrix and select Conditional Formatting -> Color Scales -> Red - White to get a clear picture of likely copies.
CLOSENESSFACTOR = 0.93
This value represents the threshold for two files to be similar in order to output a diff of the two files. These diffs will be output to the same location the script is running. They can be loaded in the browser to see the differences in the similar files.
AUTOJUNK = False
Parameter for special 'junk' heuristic. See Python Docs.