Skip to main content

MarkShark — the open-source bubble hunter (Optical Mark Recognition toolkit)

Project description

MarkShark

An easy, fast, accurate, customizable, test bubble sheet scorer

MarkShark Logo

MarkSharkOMR is a versatile and fast tool to grade, and analyze your own bubble-sheet exams.

License: AGPL v3 PyPI version


MarkShark exists because grading paper exams should not require permission, fees, or delays.

I built MarkShark so I could scan and score exams in my own office, release results the same day, and use bubble sheets that actually fit my tests. I wanted to work on the computer I already use, without buying software or relying on a central scanning service.

For many instructors, automated grading still means giving up control. Exams must fit a single Scantron form. Requests are submitted in advance. Tests are processed on someone else’s schedule. Small changes or rescoring take days. Students wait, email, and worry.

MarkShark is designed so instructors can grade their own assessments, on their own timeline, using forms they designed themselves or selected from ready-made templates. It works with ordinary scanners or photocopiers, runs on Mac, Windows, and Linux, and produces results you can inspect and reuse anywhere.

If you can print an exam and scan it, you should be able to grade it quickly and move on.


Prerequisites

  • Python 3.9–3.12

Installation

pip install markshark

Dependencies (installs automatically with pip)

  • typer>=0.12
  • rich>=13.7
  • opencv-contrib-python>=4.9
  • numpy>=1.26,<2.0
  • pymupdf>=1.24
  • pdf2image>=1.17
  • pandas>=2.1,<2.3
  • matplotlib>=3.8
  • streamlit>=1.35
  • pyyaml
  • openpyxl>=3.0
  • rapidfuzz>=3.0
  • streamlit-antd-components>=0.3.2

About MarkShark

MarkShark is in development. It currently works well after being tested in a real-world scenario of a class of 270 students with two midterms and a final (a total of over 700 scans). You can go from 500 scanned student bubblesheets to a final report in less than five minutes.

We provide a manual as a pdf in the main github directory. Help is also available from the command line. For the time being documentation is being kept current in the pdf manual and not on this markdown page.

MarkShark works with a variety of bubblesheet formats (you can easily generate your own custom bubblesheets) and it rapidly generates reports about student performance and the quality of the test questions. We provide template bubblesheets and necessary mapping files for each that you can customize for your own use.

It accepts multiple versions of the same test (you provide a single key file that has the correct answsers for all versions).

If you provide MarkShark with your class roster it can tell you who was missing from the test (absent) and flag 'orphan' scans (where the student didn't fill in their information properly. MarkShark also flags issues like unfilled bubbles (unanswered question), rows where more than one bubble was filled in, for you to review and correct if necessary. It provides student scores in a spreadsheet format that is easily pasted into spreadsheets and into LMSs like Blackboard and Canvas.

For testing prior to use MarkShark can generate fake student data ncluding fake filled in bubblesheets that you can print out, rescan, and test for yourself before committing yourself to using MarkShark.

What you provide

If you want to use a bubblesheet you've been using in the past you will need to generate a map (bubblemap.yaml) file that tells MarkShark where the bubbles are and what the bubbles represent. You need to provide MarkShark with a pdf of the blank bubblesheet (the master template) and its corresponding bubblemap one time. You can save these templates into MarkShark and reuse them again and again with a single click.

Then for each test you simply upload your scanned student sheets as a pdf. If you have folder of jpg or png images we have a utitlity that quickly connverts them to a pdf. A key is optional but required for getting student scores (percent correct, etc). If you also provide the class roster, MarkShark identify students who were absent and give you back scores in a format you can easily upload back to your learning management software (LMS like Blackboard, Canvas, Moodle, Sakai, Open edX).

License

MarkShark — the open-source bubble hunter
Copyright © 2026 William Navarre, University of Toronto
Licensed under the GNU Affero General Public License v3 (AGPL-3.0).

You may use, modify, and redistribute this software for academic, research, and open-source purposes, provided derivative works remain open-source under the same license.

Commercial or institutional use (e.g., SaaS platforms, proprietary educational tools, or for-profit distribution) requires a separate license.

For licensing inquiries, contact william.navarre@utoronto.ca.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

markshark-1.1.0.tar.gz (9.4 MB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

markshark-1.1.0-py3-none-any.whl (9.5 MB view details)

Uploaded Python 3

File details

Details for the file markshark-1.1.0.tar.gz.

File metadata

  • Download URL: markshark-1.1.0.tar.gz
  • Upload date:
  • Size: 9.4 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.11.14

File hashes

Hashes for markshark-1.1.0.tar.gz
Algorithm Hash digest
SHA256 91a5465ff474b29a2fbee149c373bbfb8c881c21c3cbb3b15d7cdd48bcbcd1e7
MD5 f23dc4215cee65224f0565aabdb8cb97
BLAKE2b-256 02e5477c4c94d21db7399c27104f82c74755bcca0c4467cafd9701e8dca148ad

See more details on using hashes here.

File details

Details for the file markshark-1.1.0-py3-none-any.whl.

File metadata

  • Download URL: markshark-1.1.0-py3-none-any.whl
  • Upload date:
  • Size: 9.5 MB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.11.14

File hashes

Hashes for markshark-1.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 a8e6aca3b98053bc76555f554a31ba5f695fb1d4cd0fffb05c6971e22ea31d27
MD5 2b63a67d917ea249996b13354ecc0d2f
BLAKE2b-256 54dae2c0280be62260adad7bef0095da13becfca6cb15e3cc5141e179b7b0546

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page