dolos

plagiarism detector for programming assignments

TLDR

Run an analysis on a set of source files

$ dolos run [path/to/files/*]

Open results in a local web UI

$ dolos run -f web [path/to/files/*]

Set the language explicitly

$ dolos run -l [python] [path/to/files/*]

Ignore template/boilerplate code

$ dolos run -i [path/to/template] [path/to/files/*]

Write a CSV report

$ dolos run -f csv [path/to/files/*]

Serve a previously generated report

$ dolos serve [path/to/report_directory]

SYNOPSIS

dolos command [options] [paths...]

dolos (Dodona, Ghent University) finds similar fragments across student programming submissions. It tokenizes source, compares pairs, and highlights copied regions. Typical use is a directory of assignment files plus optional template code to ignore.The web format launches a local UI for browsing pairs and overlap. CSV is for further processing. Language coverage includes common teaching languages; pass -l when extensions are mixed or missing.

PARAMETERS

-f, --output-format web|csv|...

How to present results. web starts a local viewer.

-l, --language name

Programming language of the files (otherwise inferred from extensions).

-i, --ignore file

Template or boilerplate to exclude from matching.

-h, --help

Show help.

-v, --version

Print version.

COMMANDS

run [options] files...

Tokenize submissions, compare them, and write a report.

serve reportdir_

Serve an existing analysis directory in the browser.

CAVEATS

High similarity is not proof of cheating; shared starter code, generated code, and small assignments inflate scores. Always ignore official templates. Serving reports binds a local HTTP port; do not expose student work on a public interface.

HISTORY

Developed for the Dodona learning platform at Ghent University.

RESOURCES

Homepage · Documentation · Source code